[ { "key": "6EGTBUAC", "version": 1, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/6EGTBUAC", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/6EGTBUAC", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "World Federation of Neurological Surgeons Committee", "parsedDate": "1988-06", "numChildren": 0 }, "data": { "key": "6EGTBUAC", "version": 1, "itemType": "journalArticle", "title": "Report of World Federation of Neurological Surgeons Committee on a Universal Subarachnoid Hemorrhage Grading Scale", "creators": [ { "creatorType": "author", "name": "World Federation of Neurological Surgeons Committee" } ], "abstractNote": "", "publicationTitle": "Journal of Neurosurgery", "publisher": "", "place": "", "date": "Jun 1988", "volume": "68", "issue": "6", "section": "", "partNumber": "", "partTitle": "", "pages": "985-6", "series": "", "seriesTitle": "", "seriesText": "", "journalAbbreviation": "J. Neurosurg.", "DOI": "", "citationKey": "", "url": "http://www.ncbi.nlm.nih.gov/pubmed/3131498", "accessDate": "2009-03-06T16:05:35Z", "PMID": "", "PMCID": "", "ISSN": "0022-3085", "archive": "", "archiveLocation": "", "shortTitle": "", "language": "", "libraryCatalog": "NCBI PubMed", "callNumber": "", "rights": "", "extra": "PMID: 3131498", "tags": [], "collections": [ "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T10:25:04Z" } }, { "key": "KMESWP23", "version": 66, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/KMESWP23", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/KMESWP23", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "Reynolds et al.", "parsedDate": "2003-10-01", "numChildren": 0 }, "data": { "key": "KMESWP23", "version": 66, "itemType": "journalArticle", "title": "Early Biomarkers of Stroke", "creators": [ { "creatorType": "author", "firstName": "Mark A.", "lastName": "Reynolds" }, { "creatorType": "author", "firstName": "Howard J.", "lastName": "Kirchick" }, { "creatorType": "author", "firstName": "Jeffrey R.", "lastName": "Dahlen" }, { "creatorType": "author", "firstName": "Joseph M.", "lastName": "Anderberg" }, { "creatorType": "author", "firstName": "Paul H.", "lastName": "McPherson" }, { "creatorType": "author", "firstName": "Kevin K.", "lastName": "Nakamura" }, { "creatorType": "author", "firstName": "Daniel T.", "lastName": "Laskowitz" }, { "creatorType": "author", "firstName": "Gunars E.", "lastName": "Valkirs" }, { "creatorType": "author", "firstName": "Kenneth F.", "lastName": "Buechler" } ], "abstractNote": "Background: The diagnosis and management of acute ischemic stroke are limited by the lack of rapid diagnostic assays for use in an emergency setting. Computed tomography (CT) scanning is used to diagnose hemorrhagic stroke but is relatively ineffective (<33% sensitive) in detecting ischemic stroke. The ability to correlate blood-borne protein biomarkers with stroke phenotypes would aid in the development of such rapid tests. Methods: ELISAs for >50 protein biomarkers were developed for use on a high-throughput robotic workstation. These assays were used to screen plasma samples from 214 healthy donors and 223 patients diagnosed with stroke, including 82 patients diagnosed with acute ischemic stroke. Marker assay values were first compared by univariate analysis, and then the top markers were subjected to multivariate analysis to derive a marker panel algorithm for the prediction of stroke. Results: The top markers from this analysis were S-100b (a marker of astrocytic activation), B-type neurotrophic growth factor, von Willebrand factor, matrix metalloproteinase-9, and monocyte chemotactic protein-1. In a panel algorithm in which three or more marker values above their respective cutoffs were scored as positive, these five markers provided a sensitivity of 92% at 93% specificity for ischemic stroke samples taken within 6 h from symptom onset. Conclusion: A marker panel approach to the diagnosis of stroke may provide a useful adjunct to CT scanning in the emergency setting.", "publicationTitle": "Clinical Chemistry", "publisher": "", "place": "", "date": "Octobre 1, 2003", "volume": "49", "issue": "10", "section": "", "partNumber": "", "partTitle": "", "pages": "1733-1739", "series": "", "seriesTitle": "", "seriesText": "", "journalAbbreviation": "Clin. Chem.", "DOI": "10.1373/49.10.1733", "citationKey": "", "url": "http://www.clinchem.org/cgi/content/abstract/49/10/1733", "accessDate": "2007-09-24T12:10:11Z", "PMID": "", "PMCID": "", "ISSN": "", "archive": "", "archiveLocation": "", "shortTitle": "", "language": "", "libraryCatalog": "HighWire", "callNumber": "", "rights": "", "extra": "", "tags": [ { "tag": "logistic regression" }, { "tag": "panel of biomarkers" }, { "tag": "proteomics" }, { "tag": "stroke" } ], "collections": [ "D4WF8BBU", "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T09:52:36Z" } }, { "key": "FFH4NQMT", "version": 66, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/FFH4NQMT", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/FFH4NQMT", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "Hastie et al.", "parsedDate": "2003-07", "numChildren": 0 }, "data": { "key": "FFH4NQMT", "version": 66, "itemType": "book", "title": "Elements of Statistical Learning: data mining, inference, and prediction", "creators": [ { "creatorType": "author", "firstName": "Trevor", "lastName": "Hastie" }, { "creatorType": "author", "firstName": "Robert", "lastName": "Tibshirani" }, { "creatorType": "author", "firstName": "Jerome", "lastName": "Friedman" } ], "abstractNote": "During the past decade there has been an explosion in computation and information technology. With it have come vast amounts of data in a variety of fields such as medicine, biology, finance, and marketing. The challenge of understanding these data has led to the development of new tools in the field of statistics, and spawned new areas such as data mining, machine learning, and bioinformatics. Many of these tools have common underpinnings but are often expressed with different terminology. This book describes the important ideas in these areas in a common conceptual framework. While the approach is statistical, the emphasis is on concepts rather than mathematics. Many examples are given, with a liberal use of color graphics. It should be a valuable resource for statisticians and anyone interested in data mining in science or industry. The book's coverage is broad, from supervised learning (prediction) to unsupervised learning. The many topics include neural networks, support vector machines, classification trees and boosting---the first comprehensive treatment of this topic in any book.\n\nThis major new edition features many topics not covered in the original, including graphical models, random forests, ensemble methods, least angle regression & path algorithms for the lasso, non-negative matrix factorization, and spectral clustering. There is also a chapter on methods for ``wide'' data (p bigger than n), including multiple testing and false discovery rates.\n\nTrevor Hastie, Robert Tibshirani, and Jerome Friedman are professors of statistics at Stanford University. They are prominent researchers in this area: Hastie and Tibshirani developed generalized additive models and wrote a popular book of that title. Hastie co-developed much of the statistical modeling software and environment in R/S-PLUS and invented principal curves and surfaces. Tibshirani proposed the lasso and is co-author of the very successful An Introduction to the Bootstrap. Friedman is the co-inventor of many data-mining tools including CART, MARS, projection pursuit and gradient boosting.", "series": "Springer Series in Statistics", "seriesNumber": "", "volume": "", "numberOfVolumes": "", "edition": "Springer-Verlag", "date": "2003-07", "publisher": "", "place": "New York", "originalDate": "", "originalPublisher": "", "originalPlace": "", "format": "", "numPages": "552", "ISBN": "978-0-387-95284-0", "DOI": "", "citationKey": "", "url": "http://www-stat.stanford.edu/~tibs/ElemStatLearn/", "accessDate": "2008-08-29T09:23:26Z", "ISSN": "", "archive": "", "archiveLocation": "", "shortTitle": "", "language": "", "libraryCatalog": "", "callNumber": "", "rights": "", "extra": "", "tags": [ { "tag": "machine learning" } ], "collections": [ "D4WF8BBU", "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T09:52:36Z" } }, { "key": "9AMGAZUC", "version": 66, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/9AMGAZUC", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/9AMGAZUC", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "Breiman", "parsedDate": "2001-10-01", "numChildren": 0 }, "data": { "key": "9AMGAZUC", "version": 66, "itemType": "journalArticle", "title": "Random Forests", "creators": [ { "creatorType": "author", "firstName": "Leo", "lastName": "Breiman" } ], "abstractNote": "Random forests are a combination of tree predictors such that each tree depends on the values of a random vector sampled independently and with the same distribution for all trees in the forest. The generalization error for forests converges a.s. to a limit as the number of trees in the forest becomes large. The generalization error of a forest of tree classifiers depends on the strength of the individual trees in the forest and the correlation between them. Using a random selection of features to split each node yields error rates that compare favorably to Adaboost (Y. Freund & R. Schapire, Machine Learning: Proceedings of the Thirteenth International conference, ***, 148–156), but are more robust with respect to noise. Internal estimates monitor error, strength, and correlation and these are used to show the response to increasing the number of features used in the splitting. Internal estimates are also used to measure variable importance. These ideas are also applicable to regression.", "publicationTitle": "Machine Learning", "publisher": "", "place": "", "date": "October 01, 2001", "volume": "45", "issue": "1", "section": "", "partNumber": "", "partTitle": "", "pages": "5-32", "series": "", "seriesTitle": "", "seriesText": "", "journalAbbreviation": "", "DOI": "10.1023/A:1010933404324", "citationKey": "", "url": "http://dx.doi.org/10.1023/A:1010933404324", "accessDate": "2009-10-07T07:01:29Z", "PMID": "", "PMCID": "", "ISSN": "", "archive": "", "archiveLocation": "", "shortTitle": "", "language": "", "libraryCatalog": "SpringerLink", "callNumber": "", "rights": "", "extra": "", "tags": [ { "tag": "machine learning" } ], "collections": [ "D4WF8BBU", "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T09:52:36Z" } }, { "key": "K2M29MZT", "version": 66, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/K2M29MZT", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/K2M29MZT", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "Wicki et al.", "parsedDate": "2001-01-08", "numChildren": 0 }, "data": { "key": "K2M29MZT", "version": 66, "itemType": "journalArticle", "title": "Assessing Clinical Probability of Pulmonary Embolism in the Emergency Ward: A Simple Score", "creators": [ { "creatorType": "author", "firstName": "Jacques", "lastName": "Wicki" }, { "creatorType": "author", "firstName": "Thomas V.", "lastName": "Perneger" }, { "creatorType": "author", "firstName": "Alain F.", "lastName": "Junod" }, { "creatorType": "author", "firstName": "Henri", "lastName": "Bounameaux" }, { "creatorType": "author", "firstName": "Arnaud", "lastName": "Perrier" } ], "abstractNote": "Objective To develop a simple standardized clinical score to stratify emergency ward patients with clinically suspected pulmonary embolism (PE) into groups with a high, intermediate, or low probability of PE to improve and simplify the diagnostic approach. Methods Analysis of a database of 1090 consecutive patients admitted to the emergency ward for suspected PE in whom diagnosis of PE was ruled in or out by a standard diagnostic algorithm. Logistic regression was used to predict clinical parameters associated with PE. Results A total of 296 (27%) of 1090 patients were found to have PE. The optimal estimate of clinical probability was based on 8 variables: recent surgery, previous thromboembolic event, older age, hypocapnia, hypoxemia, tachycardia, band atelectasis, or elevation of a hemidiaphragm on chest x-ray film. A probability score was calculated by adding points assigned to these variables. A cutoff score of 4 best identified patients with low probability of PE. A total of 486 patients (49%) had a low clinical probability of PE (score [<=]4), of which 50 (10.3%) had a proven PE. The prevalence of PE was 38% in the 437 patients with an intermediate probability (score of 5-8; n = 437) and 81% in the 63 patients with a high probability (score [>=]9). Conclusions This clinical score, based on easily available and objective variables, provides a standardized assessment of the clinical probability of PE. Applying this score to emergency ward patients suspected of having PE could allow a more effective diagnostic process.", "publicationTitle": "Archives of Internal Medicine", "publisher": "", "place": "", "date": "January 8, 2001", "volume": "161", "issue": "1", "section": "", "partNumber": "", "partTitle": "", "pages": "92-97", "series": "", "seriesTitle": "", "seriesText": "", "journalAbbreviation": "Arch. Intern. Med.", "DOI": "10.1001/archinte.161.1.92", "citationKey": "", "url": "http://archinte.ama-assn.org/cgi/content/abstract/161/1/92", "accessDate": "2009-03-16T16:19:39Z", "PMID": "", "PMCID": "", "ISSN": "", "archive": "", "archiveLocation": "", "shortTitle": "Assessing Clinical Probability of Pulmonary Embolism in the Emergency Ward", "language": "", "libraryCatalog": "HighWire", "callNumber": "", "rights": "", "extra": "", "tags": [ { "tag": "panel of biomarkers" } ], "collections": [ "D4WF8BBU", "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T09:52:36Z" } }, { "key": "I7M2TSKZ", "version": 66, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/I7M2TSKZ", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/I7M2TSKZ", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "Hoffer and Osmond", "parsedDate": "1961-08", "numChildren": 0 }, "data": { "key": "I7M2TSKZ", "version": 66, "itemType": "journalArticle", "title": "A card sorting test helpful in making psychiatric diagnosis", "creators": [ { "creatorType": "author", "firstName": "A", "lastName": "Hoffer" }, { "creatorType": "author", "firstName": "H", "lastName": "Osmond" } ], "abstractNote": "", "publicationTitle": "Journal of Neuropsychiatry", "publisher": "", "place": "", "date": "Aug 1961", "volume": "2", "issue": "", "section": "", "partNumber": "", "partTitle": "", "pages": "306-330", "series": "", "seriesTitle": "", "seriesText": "", "journalAbbreviation": "J Neuropsychiatr.", "DOI": "", "citationKey": "", "url": "http://www.ncbi.nlm.nih.gov/pubmed/13714996", "accessDate": "2009-03-19T09:36:24Z", "PMID": "", "PMCID": "", "ISSN": "", "archive": "", "archiveLocation": "", "shortTitle": "", "language": "", "libraryCatalog": "NCBI PubMed", "callNumber": "", "rights": "", "extra": "PMID: 13714996", "tags": [], "collections": [ "D4WF8BBU", "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T09:52:36Z" } }, { "key": "8DWXMMBB", "version": 1, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/8DWXMMBB", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/8DWXMMBB", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "Zheng et al.", "parsedDate": "2007", "numChildren": 0 }, "data": { "key": "8DWXMMBB", "version": 1, "itemType": "journalArticle", "title": "A Multiparametric Panel for Ovarian Cancer Diagnosis, Prognosis, and Response to Chemotherapy", "creators": [ { "creatorType": "author", "firstName": "Yingye", "lastName": "Zheng" }, { "creatorType": "author", "firstName": "Dionyssios", "lastName": "Katsaros" }, { "creatorType": "author", "firstName": "Shannon J.C.", "lastName": "Shan" }, { "creatorType": "author", "firstName": "Irene Rigault", "lastName": "de la Longrais" }, { "creatorType": "author", "firstName": "Mauro", "lastName": "Porpiglia" }, { "creatorType": "author", "firstName": "Andreas", "lastName": "Scorilas" }, { "creatorType": "author", "firstName": "Nam W.", "lastName": "Kim" }, { "creatorType": "author", "firstName": "Robert L.", "lastName": "Wolfert" }, { "creatorType": "author", "firstName": "Iris", "lastName": "Simon" }, { "creatorType": "author", "firstName": "Lin", "lastName": "Li" }, { "creatorType": "author", "firstName": "Ziding", "lastName": "Feng" }, { "creatorType": "author", "firstName": "Eleftherios P.", "lastName": "Diamandis" } ], "abstractNote": "Purpose: Our goal was to examine a panel of 11 biochemical variables, measured in cytosolic extracts of ovarian tissues (normal, benign, and malignant) by quantitative ELISAs for their ability to diagnose, prognose, and predict response to chemotherapy of ovarian cancer patients. Experimental Design: Eleven proteins were measured (9 kallikreins, B7-H4, and CA125) in cytosolic extracts of 259 ovarian tumor tissues, 50 tissues from benign conditions, 35 normal tissues, and 44 tissues from nonovarian tumors that metastasized to the ovary. Odds ratios and hazard ratios and their 95% confidence interval were calculated. Time-dependent receiver operating characteristic curves for censored survival data were used to evaluate the performance of the biomarkers. Resampling was used to validate the performance. Results: Most biomarkers effectively separated cancer from noncancer groups. A composite marker provided an area under the curve of 0.97 (95% confidence interval, 0.95-0.99) for discriminating normal and cancer groups. Univariately, hK5 and hK6 were positively associated with progression. After adjusting for clinical variables in multivariate analysis, both hK10 and hK11 significantly predicted time to progression. Increasing levels of hK13 were associated with chemotherapy response, and the predictive power of hK13 to chemotherapy response was improved by a panel of five biomarkers. Conclusions: The evidence shows that a group of kallikreins and multiparametric combinations with other biomarkers and clinical variables can significantly assist with ovarian cancer classification, prognosis, and response to platinum-based chemotherapy. In particular, we developed a multiparametric strategy for predicting ovarian cancer response to chemotherapy, comprising several biomarkers and clinical features.", "publicationTitle": "Clinical Cancer Research", "publisher": "", "place": "", "date": "Décembre 1, 2007", "volume": "13", "issue": "23", "section": "", "partNumber": "", "partTitle": "", "pages": "6984-6992", "series": "", "seriesTitle": "", "seriesText": "", "journalAbbreviation": "Clin. Cancer. Res.", "DOI": "10.1158/1078-0432.CCR-07-1409", "citationKey": "", "url": "http://clincancerres.aacrjournals.org/cgi/content/abstract/13/23/6984", "accessDate": "2008-01-11T08:26:34Z", "PMID": "", "PMCID": "", "ISSN": "", "archive": "", "archiveLocation": "", "shortTitle": "", "language": "", "libraryCatalog": "HighWire", "callNumber": "", "rights": "", "extra": "", "tags": [ { "tag": "Cox proportional hazards" }, { "tag": "bootstrap" }, { "tag": "cancer" }, { "tag": "logistic regression" }, { "tag": "panel of biomarkers" } ], "collections": [ "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T09:52:36Z" } }, { "key": "K52XCI4J", "version": 1, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/K52XCI4J", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/K52XCI4J", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "Zervakis et al.", "parsedDate": "2009", "numChildren": 0 }, "data": { "key": "K52XCI4J", "version": 1, "itemType": "journalArticle", "title": "Outcome prediction based on microarray analysis: a critical perspective on methods", "creators": [ { "creatorType": "author", "firstName": "Michalis", "lastName": "Zervakis" }, { "creatorType": "author", "firstName": "Michalis E", "lastName": "Blazadonakis" }, { "creatorType": "author", "firstName": "Georgia", "lastName": "Tsiliki" }, { "creatorType": "author", "firstName": "Vasiliki", "lastName": "Danilatou" }, { "creatorType": "author", "firstName": "Manolis", "lastName": "Tsiknakis" }, { "creatorType": "author", "firstName": "Dimitris", "lastName": "Kafetzopoulos" } ], "abstractNote": "BACKGROUND:Information extraction from microarrays has not yet been widely used in diagnostic or prognostic decision-support systems, due to the diversity of results produced by the available techniques, their instability on different data sets and the inability to relate statistical significance with biological relevance. Thus, there is an urgent need to address the statistical framework of microarray analysis and identify its drawbacks and limitations, which will enable us to thoroughly compare methodologies under the same experimental set-up and associate results with confidence intervals meaningful to clinicians. In this study we consider gene-selection algorithms with the aim to reveal inefficiencies in performance evaluation and address aspects that can reduce uncertainty in algorithmic validation.RESULTS:A computational study is performed related to the performance of several gene selection methodologies on publicly available microarray data. Three basic types of experimental scenarios are evaluated, i.e. the independent test-set and the 10-fold cross-validation (CV) using maximum and average performance measures. Feature selection methods behave differently under different validation strategies. The performance results from CV do not mach well those from the independent test-set, except for the support vector machines (SVM) and the least squares SVM methods. However, these wrapper methods achieve variable (often low) performance, whereas the hybrid methods attain consistently higher accuracies. The use of an independent test-set within CV is important for the evaluation of the predictive power of algorithms. The optimal size of the selected gene-set also appears to be dependent on the evaluation scheme. The consistency of selected genes over variation of the training-set is another aspect important in reducing uncertainty in the evaluation of the derived gene signature. In all cases the presence of outlier samples can seriously affect algorithmic performance.CONCLUSIONS:Multiple parameters can influence the selection of a gene-signature and its predictive power, thus possible biases in validation methods must always be accounted for. This paper illustrates that independent test-set evaluation reduces the bias of CV, and case-specific measures reveal stability characteristics of the gene-signature over changes of the training set. Moreover, frequency measures on gene selection address the algorithmic consistency in selecting the same gene signature under different training conditions. These issues contribute to the development of an objective evaluation framework and aid the derivation of statistically consistent gene signatures that could eventually be correlated with biological relevance. The benefits of the proposed framework are supported by the evaluation results and methodological comparisons performed for several gene-selection algorithms on three publicly available datasets.", "publicationTitle": "BMC Bioinformatics", "publisher": "", "place": "", "date": "2009", "volume": "10", "issue": "", "section": "", "partNumber": "", "partTitle": "", "pages": "53", "series": "", "seriesTitle": "", "seriesText": "", "journalAbbreviation": "", "DOI": "10.1186/1471-2105-10-53", "citationKey": "", "url": "http://www.biomedcentral.com/1471-2105/10/53", "accessDate": "2009-03-13T12:44:34Z", "PMID": "", "PMCID": "", "ISSN": "1471-2105", "archive": "", "archiveLocation": "", "shortTitle": "Outcome prediction based on microarray analysis", "language": "", "libraryCatalog": "BioMed Central and More", "callNumber": "", "rights": "", "extra": "", "tags": [ { "tag": "SVM" }, { "tag": "panel of biomarkers" } ], "collections": [ "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T09:52:36Z" } }, { "key": "WRJCXPUW", "version": 1, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/WRJCXPUW", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/WRJCXPUW", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "Yeo and Johnson", "parsedDate": "2000", "numChildren": 0 }, "data": { "key": "WRJCXPUW", "version": 1, "itemType": "journalArticle", "title": "A new family of power transformations to improve normality or symmetry", "creators": [ { "creatorType": "author", "firstName": "In-Kwon", "lastName": "Yeo" }, { "creatorType": "author", "firstName": "Richard A.", "lastName": "Johnson" } ], "abstractNote": "We introduce a new power transformation family which is well defined on the whole real line and which is appropriate for reducing skewness and to approximate normality. It has properties similar to those of the Box-Cox transformation for positive variables. The large-sample properties of the transformation are investigated in the contect of a single random sample.", "publicationTitle": "Biometrika", "publisher": "", "place": "", "date": "Décembre 1, 2000", "volume": "87", "issue": "4", "section": "", "partNumber": "", "partTitle": "", "pages": "954-959", "series": "", "seriesTitle": "", "seriesText": "", "journalAbbreviation": "", "DOI": "10.1093/biomet/87.4.954", "citationKey": "", "url": "http://biomet.oxfordjournals.org/cgi/content/abstract/87/4/954", "accessDate": "2008-10-10T08:40:42Z", "PMID": "", "PMCID": "", "ISSN": "", "archive": "", "archiveLocation": "", "shortTitle": "", "language": "", "libraryCatalog": "HighWire", "callNumber": "", "rights": "", "extra": "", "tags": [ { "tag": "box-cox transformation" }, { "tag": "normalization" } ], "collections": [ "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T09:52:36Z" } }, { "key": "JC256BWM", "version": 1, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/JC256BWM", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/JC256BWM", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "Wu et al.", "parsedDate": "2003-09-01", "numChildren": 0 }, "data": { "key": "JC256BWM", "version": 1, "itemType": "journalArticle", "title": "Comparison of statistical methods for classification of ovarian cancer using mass spectrometry data", "creators": [ { "creatorType": "author", "firstName": "Baolin", "lastName": "Wu" }, { "creatorType": "author", "firstName": "Tom", "lastName": "Abbott" }, { "creatorType": "author", "firstName": "David", "lastName": "Fishman" }, { "creatorType": "author", "firstName": "Walter", "lastName": "McMurray" }, { "creatorType": "author", "firstName": "Gil", "lastName": "Mor" }, { "creatorType": "author", "firstName": "Kathryn", "lastName": "Stone" }, { "creatorType": "author", "firstName": "David", "lastName": "Ward" }, { "creatorType": "author", "firstName": "Kenneth", "lastName": "Williams" }, { "creatorType": "author", "firstName": "Hongyu", "lastName": "Zhao" } ], "abstractNote": "Motivation: Novel methods, both molecular and statistical, are urgently needed to take advantage of recent advances in biotechnology and the human genome project for disease diagnosis and prognosis. Mass spectrometry (MS) holds great promise for biomarker identification and genome-wide protein profiling. It has been demonstrated in the literature that biomarkers can be identified to distinguish normal individuals from cancer patients using MS data. Such progress is especially exciting for the detection of early-stage ovarian cancer patients. Although various statistical methods have been utilized to identify biomarkers from MS data, there has been no systematic comparison among these approaches in their relative ability to analyze MS data. Results: We compare the performance of several classes of statistical methods for the classification of cancer based on MS spectra. These methods include: linear discriminant analysis, quadratic discriminant analysis, k-nearest neighbor classifier, bagging and boosting classification trees, support vector machine, and random forest (RF). The methods are applied to ovarian cancer and control serum samples from the National Ovarian Cancer Early Detection Program clinic at Northwestern University Hospital. We found that RF outperforms other methods in the analysis of MS data. Supplementary information: http://bioinformatics.med.yale.edu/proteomics/BioSupp1.html", "publicationTitle": "Bioinformatics", "publisher": "", "place": "", "date": "September 1, 2003", "volume": "19", "issue": "13", "section": "", "partNumber": "", "partTitle": "", "pages": "1636-1643", "series": "", "seriesTitle": "", "seriesText": "", "journalAbbreviation": "", "DOI": "10.1093/bioinformatics/btg210", "citationKey": "", "url": "http://bioinformatics.oxfordjournals.org/cgi/content/abstract/19/13/1636", "accessDate": "2009-04-30T15:09:23Z", "PMID": "", "PMCID": "", "ISSN": "", "archive": "", "archiveLocation": "", "shortTitle": "", "language": "", "libraryCatalog": "HighWire", "callNumber": "", "rights": "", "extra": "", "tags": [ { "tag": "MALDI" }, { "tag": "SVM" } ], "collections": [ "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T09:52:36Z" } }, { "key": "UZ375B3S", "version": 1, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/UZ375B3S", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/UZ375B3S", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "Whiteley et al.", "parsedDate": "2008-10-01", "numChildren": 0 }, "data": { "key": "UZ375B3S", "version": 1, "itemType": "journalArticle", "title": "Blood Biomarkers in the Diagnosis of Ischemic Stroke: A Systematic Review", "creators": [ { "creatorType": "author", "firstName": "William", "lastName": "Whiteley" }, { "creatorType": "author", "firstName": "Mei-Chiun", "lastName": "Tseng" }, { "creatorType": "author", "firstName": "Peter", "lastName": "Sandercock" } ], "abstractNote": "Background and Purpose-- The diagnosis of ischemic stroke can be difficult. CT may be normal in the early stages of ischemic stroke or in patients with minor symptoms and MR is not always possible. Many blood markers have been proposed for the diagnosis of stroke in the acute setting. Methods and Results-- We have systematically reviewed the diagnostic literature and found 21 studies testing 58 single biomarkers and 7 panels of several biomarkers. Although all show either a high sensitivity or specificity, there are limitations in the design and reporting of all the studies that mean no biomarker can be recommended for use in clinical practice. Conclusions-- We make recommendations for the design and reporting of studies of diagnostic blood biomarkers in stroke.", "publicationTitle": "Stroke", "publisher": "", "place": "", "date": "Octobre 1, 2008", "volume": "39", "issue": "10", "section": "", "partNumber": "", "partTitle": "", "pages": "2902-2909", "series": "", "seriesTitle": "", "seriesText": "", "journalAbbreviation": "", "DOI": "10.1161/STROKEAHA.107.511261", "citationKey": "", "url": "http://stroke.ahajournals.org/cgi/content/abstract/39/10/2902", "accessDate": "2008-10-07T09:47:09Z", "PMID": "", "PMCID": "", "ISSN": "", "archive": "", "archiveLocation": "", "shortTitle": "Blood Biomarkers in the Diagnosis of Ischemic Stroke", "language": "", "libraryCatalog": "HighWire", "callNumber": "", "rights": "", "extra": "", "tags": [ { "tag": "panel of biomarkers" } ], "collections": [ "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T09:52:36Z" } }, { "key": "ASGVNGUR", "version": 1, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/ASGVNGUR", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/ASGVNGUR", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "Wang et al.", "parsedDate": "2004", "numChildren": 0 }, "data": { "key": "ASGVNGUR", "version": 1, "itemType": "conferencePaper", "title": "Boosted PRIM with Application to Searching for Oncogenic Pathway of Lung Cancer", "creators": [ { "creatorType": "author", "firstName": "Pei", "lastName": "Wang" }, { "creatorType": "author", "firstName": "Young", "lastName": "Kim" }, { "creatorType": "author", "firstName": "Jonathan", "lastName": "Pollack" }, { "creatorType": "author", "firstName": "Robert", "lastName": "Tibshirani" } ], "abstractNote": "Boosted PRIM (Patient Rule Induction Method) is a new algorithm developed for two-class classification problems. PRIM is a variation of those Tree-Based methods ( [4] Ch9.3), seeking box-shaped regions in the feature space to separate different classes. Boosted PRIM is to implement PRIM-styled weak learners in Adaboost, one of the most popular boosting algorithms. In addition, we improve the performance of the algorithm by introducing a regularization to the boosting process, which supports the perspective of viewing boosting as a steepest-descent numerical optimization by Jerry Friedman [3]. The motivation for Boosted PRIM is to solve the problem of \"searching for oncogenic pathways\" based on array-CGH (Comparative Genomic Hybridization) data, though the algorithm itself is suitable for general classification problems. We illustrate the performance of the method through some simulation studies as well as an application on a lung cancer array-CGH data set.", "proceedingsTitle": "Proceedings of the 2004 IEEE Computational Systems Bioinformatics Conference", "conferenceName": "", "publisher": "IEEE Computer Society", "place": "", "date": "2004", "eventPlace": "", "volume": "", "issue": "", "numberOfVolumes": "", "pages": "604-609", "series": "", "seriesNumber": "", "DOI": "", "ISBN": "0-7695-2194-0", "citationKey": "", "url": "http://portal.acm.org/citation.cfm?id=1019336", "accessDate": "2009-03-06T09:34:42Z", "ISSN": "", "archive": "ACM", "archiveLocation": "", "shortTitle": "", "language": "", "libraryCatalog": "", "callNumber": "", "rights": "", "extra": "", "tags": [ { "tag": "machine learning" }, { "tag": "patient rule induction method (PRIM)" } ], "collections": [ "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T09:52:36Z" } }, { "key": "AT9MICET", "version": 1, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/AT9MICET", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/AT9MICET", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "Vitzthum et al.", "parsedDate": "2005", "numChildren": 0 }, "data": { "key": "AT9MICET", "version": 1, "itemType": "journalArticle", "title": "Proteomics: from basic research to diagnostic application. A review of requirements & needs", "creators": [ { "creatorType": "author", "firstName": "Frank", "lastName": "Vitzthum" }, { "creatorType": "author", "firstName": "Fritz", "lastName": "Behrens" }, { "creatorType": "author", "firstName": "N Leigh", "lastName": "Anderson" }, { "creatorType": "author", "firstName": "John H", "lastName": "Shaw" } ], "abstractNote": "For several years proteomics research has been expected to lead to the finding of new markers that will translate into clinical tests applicable to samples such as serum, plasma and urine: so-called in vitro diagnostics (IVDs). Attempts to implement technologies applied in proteomics, in particular protein arrays and surface-enhanced laser desorption ionization time-of-flight mass spectrometry (SELDI-TOF MS), as IVD instruments have initiated constructive discussions on opportunities and challenges inherent in such a translation process also with respect to the use of multi-marker profiling approaches and pattern signatures in IVD. Taking into account the role that IVD plays in health care, we describe IVD requirements and needs. Subject to stringent costs versus benefit analyses, IVD has to provide reliable information about a person's condition, prognosis or risk to suffer a disease, thus supporting decisions on treatment or prevention. It is mandatory to fulfill requirements in routine IVD, including disease prevention, diagnosis, prognosis, and treatment monitoring or follow up among others. To fulfill IVD requirements, it is essential to (1) provide diagnostic tests that allow for definite and reliable diagnosis tied to a decision on interventions (prevention, treatment, or nontreatment), (2) meet stringent performance characteristics for each analyte (in particular test accuracy, including both precision of the measurement and trueness of the measurement), and (3) provide adequate diagnostic accuracy, i.e., diagnostic sensitivity and diagnostic specificity, determined by the desired positive and negative predictive values which depend on disease frequency. The fulfillment of essential IVD requirements is mandatory in the regulated environment of modern diagnostics. Addressing IVD needs at an early stage can support a timely and effective transition of findings and developments into routine diagnosis. IVD needs reflect features that are useful in clinical practice. This helps to generate acceptance and assists the implementation process. On the basis of IVD requirements and needs, we outline potential implications for clinical proteomics focused on applied research activities.", "publicationTitle": "Journal of Proteome Research", "publisher": "", "place": "", "date": "2005", "volume": "4", "issue": "4", "section": "", "partNumber": "", "partTitle": "", "pages": "1086-1097", "series": "", "seriesTitle": "", "seriesText": "", "journalAbbreviation": "J. Proteome Res.", "DOI": "10.1021/pr050080b", "citationKey": "", "url": "http://www.ncbi.nlm.nih.gov/pubmed/16083257", "accessDate": "2008-09-10T13:05:37Z", "PMID": "", "PMCID": "", "ISSN": "1535-3893", "archive": "", "archiveLocation": "", "shortTitle": "Proteomics", "language": "", "libraryCatalog": "NCBI PubMed", "callNumber": "", "rights": "", "extra": "PMID: 16083257", "tags": [ { "tag": "panel of biomarkers" } ], "collections": [ "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T09:52:36Z" } }, { "key": "AF4IT2QJ", "version": 1, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/AF4IT2QJ", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/AF4IT2QJ", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "Visintin et al.", "parsedDate": "2008", "numChildren": 0 }, "data": { "key": "AF4IT2QJ", "version": 1, "itemType": "journalArticle", "title": "Diagnostic Markers for Early Detection of Ovarian Cancer", "creators": [ { "creatorType": "author", "firstName": "Irene", "lastName": "Visintin" }, { "creatorType": "author", "firstName": "Ziding", "lastName": "Feng" }, { "creatorType": "author", "firstName": "Gary", "lastName": "Longton" }, { "creatorType": "author", "firstName": "David C.", "lastName": "Ward" }, { "creatorType": "author", "firstName": "Ayesha B.", "lastName": "Alvero" }, { "creatorType": "author", "firstName": "Yinglei", "lastName": "Lai" }, { "creatorType": "author", "firstName": "Jeannette", "lastName": "Tenthorey" }, { "creatorType": "author", "firstName": "Aliza", "lastName": "Leiser" }, { "creatorType": "author", "firstName": "Ruben", "lastName": "Flores-Saaib" }, { "creatorType": "author", "firstName": "Herbert", "lastName": "Yu" }, { "creatorType": "author", "firstName": "Masoud", "lastName": "Azori" }, { "creatorType": "author", "firstName": "Thomas", "lastName": "Rutherford" }, { "creatorType": "author", "firstName": "Peter E.", "lastName": "Schwartz" }, { "creatorType": "author", "firstName": "Gil", "lastName": "Mor" } ], "abstractNote": "Purpose: Early detection would significantly decrease the mortality rate of ovarian cancer. In this study, we characterize and validate the combination of six serum biomarkers that discriminate between disease-free and ovarian cancer patients with high efficiency. Experimental Design: We analyzed 362 healthy controls and 156 newly diagnosed ovarian cancer patients. Concentrations of leptin, prolactin, osteopontin, insulin-like growth factor II, macrophage inhibitory factor, and CA-125 were determined using a multiplex, bead-based, immunoassay system. All six markers were evaluated in a training set (181 samples from the control group and 113 samples from OC patients) and a test set (181 sample control group and 43 ovarian cancer). Results: Multiplex and ELISA exhibited the same pattern of expression for all the biomarkers. None of the biomarkers by themselves were good enough to differentiate healthy versus cancer cells. However, the combination of the six markers provided a better differentiation than CA-125. Four models with <2% classification error in training sets all had significant improvement (sensitivity 84%-98% at specificity 95%) over CA-125 (sensitivity 72% at specificity 95%) in the test set. The chosen model correctly classified 221 out of 224 specimens in the test set, with a classification accuracy of 98.7%. Conclusions: We describe the first blood biomarker test with a sensitivity of 95.3% and a specificity of 99.4% for the detection of ovarian cancer. Six markers provided a significant improvement over CA-125 alone for ovarian cancer detection. Validation was performed with a blinded cohort. This novel multiplex platform has the potential for efficient screening in patients who are at high risk for ovarian cancer.", "publicationTitle": "Clinical Cancer Research", "publisher": "", "place": "", "date": "Février 15, 2008", "volume": "14", "issue": "4", "section": "", "partNumber": "", "partTitle": "", "pages": "1065-1072", "series": "", "seriesTitle": "", "seriesText": "", "journalAbbreviation": "Clin. Cancer Res.", "DOI": "10.1158/1078-0432.CCR-07-1569", "citationKey": "", "url": "http://clincancerres.aacrjournals.org/cgi/content/abstract/14/4/1065", "accessDate": "2008-03-06T08:23:00Z", "PMID": "", "PMCID": "", "ISSN": "", "archive": "", "archiveLocation": "", "shortTitle": "", "language": "", "libraryCatalog": "HighWire", "callNumber": "", "rights": "", "extra": "", "tags": [ { "tag": "cancer" }, { "tag": "cross-validation" }, { "tag": "logistic regression" }, { "tag": "panel of biomarkers" }, { "tag": "validation set" } ], "collections": [ "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T09:52:36Z" } }, { "key": "39BPP5TN", "version": 1, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/39BPP5TN", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/39BPP5TN", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "Vasconcelos et al.", "parsedDate": "2006-09", "numChildren": 0 }, "data": { "key": "39BPP5TN", "version": 1, "itemType": "journalArticle", "title": "A comparison of fatigue scales in postpoliomyelitis syndrome", "creators": [ { "creatorType": "author", "firstName": "Olavo M. Jr.", "lastName": "Vasconcelos" }, { "creatorType": "author", "firstName": "Olga A.", "lastName": "Prokhorenko" }, { "creatorType": "author", "firstName": "Kay F.", "lastName": "Kelley" }, { "creatorType": "author", "firstName": "Alexander H.", "lastName": "Vo" }, { "creatorType": "author", "firstName": "Cara H.", "lastName": "Olsen" }, { "creatorType": "author", "firstName": "Marinos C.", "lastName": "Dalakas" }, { "creatorType": "author", "firstName": "Lauro S.", "lastName": "Halstead" }, { "creatorType": "author", "firstName": "Bahman", "lastName": "Jabbari" }, { "creatorType": "author", "firstName": "William W.", "lastName": "Campbell" } ], "abstractNote": "OBJECTIVE: To examine the applicability and validity of traditional fatigue questionnaires in postpoliomyelitis syndrome (PPS) patients with disabling fatigue. DESIGN: Cross-sectional study. PPS and disabling fatigue were ascertained according to published criteria. Descriptiveness was determined using the McNemar test, and interscale z-score agreement was estimated with Pearson's coefficients. SETTING: PPS clinic. PARTICIPANTS: Fifty-six survivors of poliomyelitis: 39 met criteria for PPS, 25 of whom met criteria for disabling fatigue. INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURES: The Fatigue Severity Scale (FSS), visual analog scale (VAS) for fatigue, and Fatigue Impact Scale (FIS). RESULTS: Twenty-four patients scored 50% or higher on the scale range for FSS, compared with 19 patients for VAS for fatigue (P=.042), and 7 patients for FIS (P<.001). Scores for patients with disabling fatigue averaged 81.5%, 62%, and 40.9% of the scale range for FSS, VAS for fatigue, and FIS, respectively. Agreement was moderate between the FSS and VAS for fatigue (r=.45, P=.02), but low between FSS and FIS (r=.29, P=.15), and FIS and VAS for fatigue (r=.20, P=.33). Two sample t tests showed significant differences between those with disabling fatigue and those without, based on FSS scores (t=3.8, P<.001), but not for VAS for fatigue or FIS scores. CONCLUSIONS: FSS was the most descriptive of the instruments tested. Scores generated by the scales were not interchangeable. Of the 3 scales, FFS seemed to be the most informative for the clinical assessment of fatigue in patients with PPS.", "publicationTitle": "Archives of Physical Medicine and Rehabilitation", "publisher": "", "place": "", "date": "Sep 2006", "volume": "87", "issue": "9", "section": "", "partNumber": "", "partTitle": "", "pages": "1213-1217", "series": "", "seriesTitle": "", "seriesText": "", "journalAbbreviation": "Arch. Phys. Med. Rehabil.", "DOI": "10.1016/j.apmr.2006.06.009", "citationKey": "", "url": "http://www.ncbi.nlm.nih.gov/pubmed/16935057", "accessDate": "2009-03-16T12:07:30Z", "PMID": "", "PMCID": "", "ISSN": "0003-9993", "archive": "", "archiveLocation": "", "shortTitle": "", "language": "", "libraryCatalog": "NCBI PubMed", "callNumber": "", "rights": "", "extra": "PMID: 16935057", "tags": [ { "tag": "mcnemar" }, { "tag": "panel of biomarkers" } ], "collections": [ "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T09:52:36Z" } }, { "key": "UUT99G7D", "version": 1, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/UUT99G7D", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/UUT99G7D", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "Thompson and Zucchini", "parsedDate": "1989", "numChildren": 0 }, "data": { "key": "UUT99G7D", "version": 1, "itemType": "journalArticle", "title": "On the statistical analysis of ROC curves", "creators": [ { "creatorType": "author", "firstName": "M. L.", "lastName": "Thompson" }, { "creatorType": "author", "firstName": "W.", "lastName": "Zucchini" } ], "abstractNote": "We introduce a new accuracy index for receiver operating characteristic (ROC) curves, namely the partial area under the binormal ROC graph over any specified region of interest. We propose a simple but general procedure, based on a conventional analysis of variance, for comparing accuracy indices derived from two or more different modalities. The proposed method is related to and compared with existing methodology, and is illustrated by results from an experiment on optimization of density and contrast yielded by multiform photographic images used for scintigraphy.", "publicationTitle": "Statistics in Medicine", "publisher": "", "place": "", "date": "1989", "volume": "8", "issue": "10", "section": "", "partNumber": "", "partTitle": "", "pages": "1277-1290", "series": "", "seriesTitle": "", "seriesText": "", "journalAbbreviation": "", "DOI": "10.1002/sim.4780081011", "citationKey": "", "url": "http://dx.doi.org/10.1002/sim.4780081011", "accessDate": "2009-03-02T15:21:00Z", "PMID": "", "PMCID": "", "ISSN": "", "archive": "", "archiveLocation": "", "shortTitle": "", "language": "", "libraryCatalog": "Wiley InterScience", "callNumber": "", "rights": "", "extra": "", "tags": [ { "tag": "ROC Curve" } ], "collections": [ "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T09:52:36Z" } }, { "key": "2NUUVJUB", "version": 1, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/2NUUVJUB", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/2NUUVJUB", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "Steel et al.", "parsedDate": "2005-02", "numChildren": 0 }, "data": { "key": "2NUUVJUB", "version": 1, "itemType": "journalArticle", "title": "Methods of comparative proteomic profiling for disease diagnostics", "creators": [ { "creatorType": "author", "firstName": "Laura F.", "lastName": "Steel" }, { "creatorType": "author", "firstName": "Brian B.", "lastName": "Haab" }, { "creatorType": "author", "firstName": "Samir M.", "lastName": "Hanash" } ], "abstractNote": "aDepartment of Microbiology and Immunology, Drexel University College of Medicine, Doylestown, PA 18901, USA bThe Van Andel Research Institute, 333 Bostwick NE, Grand Rapids, MI 49503, USA cFred Hutchinson Cancer Research Center, Seattle, WA 98109, USA The recent development of numerous technologies for proteome analysis holds the promise of new and more precise methods for disease diagnosis. In this review, we provide an overview of some of these technologies including two-dimensional gel electrophoresis (2DE), historically the workhorse of proteomic analysis, as well as some newer approaches such as liquid phase separations combined with mass spectrometry, and protein microarrays. It is evident that each method has its own strengths and weaknesses and no single method will be optimal in all applications. However, the continuing development of innovative strategies for protein separation and analysis is providing a wealth of new tools for multi-dimensional protein profiling that will advance our capabilities in disease diagnostics and our understanding of disease pathology. Keywords: Comparative proteomic profiling; Disease diagnostics; Proteins biomarkers Nucleic acid based technologies have been widely used in studies of comparative gene expression profiling for biomarker discovery. However, it is essential that these studies also be carried out at the protein level. Proteins are the functional readout of genetic information and protein activity can be affected by many factors that are not reflected in the RNA transcript population (transcriptome). For instance, there can be a substantial discordance between mRNA abundance and protein expression levels [1] and [2]. Further, over 200 different post-translational modifications [3] can regulate protein function by altering properties such as interactions with other biomolecules or sub-cellular localization. In developing tools for disease diagnostics, it is also important to consider that many of the biological fluids that are relatively accessible for analysis, such as serum, urine, and saliva, are rich in protein but very poor sources of nucleic acids for assay. An ideal proteome screening methodology would combine high throughput capabilities with detection of as many protein products as possible in a sensitive, reproducible, and quantifiable manner. The wide-ranging biochemical heterogeneity of proteins makes it unlikely that any single separation and analysis method will be suitable for profiling the full proteome of any cell type, tissue, or biological fluid. In the following sections, we describe several of the tools that have been used or that are being developed for protein biomarker discovery and disease diagnostics, each with its own strengths as well as limitations. Several reviews of proteomic investigations in disease diagnosis have been published [4], [5], [6], [7], [8] and [9]. Here, we will emphasize recent studies that are illustrative of proteomic approaches currently being used. Two-dimensional gel electrophoresis (2DE), developed in the mid-1970s [10] and [11], was the first method to allow the resolution and simultaneous display of hundreds of proteins. Recent improvements in the implementation of this basic technology, together with the explosion of protein sequence information resulting from genomic studies, and the development of techniques for peptide analysis by mass spectrometry, have fueled the emergence of proteomics as a powerful tool for comparative gene expression profiling. The straightforward application of 2DE in disease proteomics is well-demonstrated by a large study aimed at the discovery of proteins that might serve as prognostic biomarkers for survival of lung cancer patients [12]. Proteins from lung tissue of 90 patients with lung adenocarcinoma were resolved by 2DE and 682 protein spots were quantified and statistically analyzed for correlations with patient survival. Using the top 20 proteins that showed a significant correlation with survival, it was possible to generate a risk index that was highly predictive of outcome for patients with early stage tumors. Of a total of 46 spots shown to correlate with survival, 33 were identified by mass spectrometry, providing information regarding biological changes associated with the tumor tissue. Importantly, one of the proteins identified, phosphoglycerate kinase 1 (PGK1), was detected in serum where it retained its strong correlation with patient survival, suggesting that it may prove useful in screens of disease progression. It is worth noting that for most of the proteins identified in this study, similar associations with survival were not found at the mRNA level when the same [13] or different [14] tumor sets were examined by microarray analysis [12]. This underscores the value of conducting studies at multiple levels of gene expression. 2DE has been used by many investigators to compare the protein complements of diseased and healthy tissue. Among the more recent studies is one in which proteins from peripheral blood mononuclear cells (PBMCs) were compared between healthy individuals and those with rheumatoid arthritis [15]. Twenty-nine differentially expressed protein spots were found in gels that could be used in hierarchical clustering for the accurate separation of healthy individuals from those with arthritis. Some of the proteins were identified by mass spectrometry and have known roles in inflammatory or autoimmune processes [15], showing that this approach will be of value both diagnostically and in helping to understand the disease pathology. Another recent study used 2DE and mass spectrometry to identify proteins whose expression is modulated in oral tongue squamous cell carcinoma [16]. Approximately 600 protein spots from ten pairs of matched tumor and surrounding non-tumor tissue were compared by 2DE and spots showing consistent differences were identified by peptide mass fingerprinting. Many of the observed changes could be explained in terms of tongue tumor pathology, the increased vascularization of the tumor tissue, or were proteins whose expression has been found to be modulated in other tumors as well [16]. While direct comparisons of tumor and non-tumor tissue are clearly informative, it is also true that interactions among the heterogeous cell types that comprise the tumor microenvironment are critical to disease progression [17]. In order to capture proteins engaged in this intercellular cross-talk, methods are being devised to sample fluids that contact diseased tissue. For instance, nipple aspirate fluid is being examined for markers of breast cancer [18] and urine is being studied for markers of urinary tract disease [19]. Proteins from fluid that directly perfuses breast tumor tissue were collected from supernatants of short term cultures of freshly excised tumor tissue [20]. These tumor interstitial fluid (TIF) proteins were characterized by 2DE with subsequent identification of 267 proteins by mass spectrometry, immunoblotting, or comparison to existing databases. Proteins were found representing many aspects of cellular metabolism, cell-cell interactions, and angiogenesis, indicating that TIF protein profiles will likely be a rich source of information related to the interplay between healthy and diseased cells, as well as to the body's defense response to the diseased tissue. A novel application of 2DE has been in the discovery of circulating autoantibodies in cancer patients. There is evidence of a humoral immune response against tumor antigens in some cancer patients that might be used in serum-based assays of disease progression or in the development of anticancer vaccines [21], [22] and [23]. Proteins isolated from tumor tissue or cell lines are resolved by 2DE and then transferred to membranes for immunoblotting against patient sera. Immunoglobulins present in the sera that have reactivity against tumor proteins can be detected in these 2D blots and the antigenic protein can be identified by mass spectrometry after alignment of the blot with a stained gel. Sera from patients with lung cancer [21] and [24], hepatocellular carcinoma [22], renal cell carcinoma [23] and [25], and neuroblastoma [26] have been studied in this way, and in each case, autoantibodies specific to a limited number of tumor proteins were found. Although it is incompletely understood why some tumor proteins become antigenic in a subset of patients, the antitumor antibodies often correspond to proteins that are overexpressed, mislocalized, or mutant in the tumor. There is also evidence that increased cytokine activity contributes to the development of autoantibodies in some patients [21]. Despite successes with 2DE, the method has many, often-described limitations. For instance, solubility problems can lead to an under-representation of hydrophobic membrane proteins, highly basic proteins are difficult to resolve in first dimension focusing gels, and the dynamic range of detection possible in gels can be exceeded by the dynamic range of protein abundance in samples, making the detection of low abundance proteins difficult. Many of these limitations are being addressed, both through improvements to the technology and by using 2DE together with other technologies to take advantage of the complementary strengths of each. New detergents are being used to extend the utility of 2DE to more of the low solubility proteins [27] and [28] and work continues to improve the resolution of basic proteins [29], [30] and [31]. Gels that focus proteins in a very narrow pH range in the first dimension, so-called zoom gels, can be used to increase the number of proteins resolved in the 2D system [32]. The problems posed by widely different levels of proteins in a sample are particularly notable in serum or plasma where proteins vary by as much as 12 orders of magnitude in abundance and a small number of proteins, including albumin, immunoglobulins, transferrin, haptoglobins, α1-antitrypsin, acid-1-glycoprotein, constitute as much as 80% of the total protein [33]. Affinity based methods are available for the specific removal of many of the abundant proteins, making minor but possibly informative proteins more accessible to detection and analysis [19] and [34]. Improvements in the detection of some serum proteins that can be gained by removal of albumin are illustrated in Fig. 1. It is also becoming evident that proteins such as albumin and the immunoglobulins can serve as carrier proteins, able to bind potentially useful biomarkers [35]. It will undoubtedly be beneficial to examine proteins and peptides that co-elute with the abundant proteins, as well as those left in the unselected population, in any separation strategy. Additional methods of sample fractionation prior to 2DE analysis are being used, essentially adding a third dimension to protein separations. More low abundance proteins become detectable when gels are loaded with proteins from individual fractions, rather than the total cell or tissue lysate. Of course, disadvantages to this approach are that it multiplies both the total number of gels required and the total amount of sample necessary for a given analysis as well as introducing the potential for protein loss or degradation as the number of sample handling steps increases. Nevertheless, liquid chromatographic separations, including those based on ion exchange, hydrophobic interactions, differential affinity, and size exclusion, have all proven useful in increasing the number of proteins resolved by 2DE (reviewed in [36]). Reversed-phase high performance liquid chromatography (RP-HPLC) was used to fractionate proteins from cultured human breast epithelial cells (HBL-100), cultured B-cells (BL60-2), and rat lung tissue [37]. Subsequent 2DE showed a reproducible fractionation that allowed detection of proteins not clearly visible in gels of unfractionated cell lysates, including some that were experimentally induced by the apoptotic agent staurosporine [37]. Solution phase isoelectric focusing has also been applied as a prefractionation step. A procedure for microscale solution isoelectrofocusing (musol-IEF) has been developed that uses a series of small volume chambers to form discrete pH zones for the high resolution separation of proteins based on pI [38]. Experiments with mouse serum [38] and human breast cancer cell extracts [39] have demonstrated that prefractionation by this procedure increases the loading capacity and greatly enhances the resolution possible with narrow pH range first dimension IPG strips. Another form of sample prefractionation can be achieved at the cellular level. Tumor specimens invariably contain mixed populations of cells, with variable proportions of diseased and normal cells, as well as mixed cell types naturally occuring in the tissue. Clearly, protein expression differences arising from the disease state could be masked by the hetergeneity of the sample. Laser capture microdissection (LCM) allows precise dissection, so that malignant cells or their non-malignant counterparts can be cleanly separated from neighboring cells in biopsy material [40]. The value of LCM in proteomic profiling is illustrated in a recent study of pancreatic ductal adenocarcinoma (PDAC), where both normal and malignant ductal epithelial cells represent only a small percentage of the tumor mass [41]. When 2DE was used to compare proteins from non-malignant pancreatic tissue with those from normal ductal cells collected by LCM, there were numerous differences, presumably due to the small contribution made by ductal cells to the heterogeneous undissected tissue. LCM was then used to prepare populations enriched in normal or malignant ductal cells from pancreatic tumors. Nine differentially expressed proteins that varied consistently between the normal and malignant ductal cells could be detected by 2DE of the LCM collected samples [41]. In a second example, LCM was used to help in profiling proteins from matched normal ductal/lobular units and ductal carcinoma in situ (DCIS) of the breast [42]. In this study, too, distinct protein profiles were generated by 2DE of proteins isolated from frozen tissue sections or LCM collected epithelial cells, and the two methods of tissue sampling produced only partially overlapping lists of differentially expressed proteins [42]. LCM is a highly labor intensive procedure that yields limited amounts of material and so it is not suitable for screening large numbers of samples. Nevertheless, it sharply focuses comparisons of proteins found in a subset of cells from heterogenous tissue, and proteins identified in cell populations obtained by LCM can be pursued in larger sample sets by other analytical techniques, such as immunohistochemistry. A serious bottleneck in the evaluation of 2D gels is the delineation of protein spot boundaries in the gel image, and the matching of spots in a series of gels so that quantitative comparisons can be made. Even with specialized imaging equipment and sophisticated software, the process requires time-consuming manual editing. This problem is exacerbated by gel-to-gel differences that arise from unavoidable minor variations in the efficiency of protein entry into the IPG strip, the transfer of proteins from the first to the second dimension gel, or in local areas of the gel composition itself. Two-dimensional difference gel electrophoresis (2D DIGE) is an analytical strategy designed to minimize these problems, making sample-to-sample comparisons easier and more accurate, as well as reducing the number of gels required to evaluate a series of samples [43] and [44], reviewed in [45]. In 2D DIGE, different size and charge matched fluorescent dyes, such as Cy3 and Cy5 derivatives, are used to covalently label the proteins of two samples that are to be compared. The labeled protein samples are mixed together and then resolved in a single 2D gel. Fluorescent signal from the Cy3 and Cy5 dyes can be imaged separately, and the ratio of labeling can be determined for each spot, allowing quantitative comparisons between the samples to be made within individual spots in the image. Since the samples are run in a single gel, differences due to technical variations are avoided and the process of gel matching is eliminated. An internal standard can be added to the mix, comprised of a combination of equal amounts of each sample in the comparison series, labeled with a third fluorescent dye, such as Cy2 [46]. This refines the accuracy of quantitation and helps in making comparisons among multiple samples. 2D DIGE has been applied to a model system of breast cancer [47]. Protein expression patterns were compared between a cell line established from human breast luminal epithelium (HB4a) and a derivative cell line that overexpresses ErbB-2 (HBc3.6). Several proteins showing deregulation in the HBc3.6 cell line could be identified by mass spectrometry and are known to be associated with changes in cell morphology, proliferation, cell transformation, or metastasis [47]. In another study, proteins from colonic tumor tissue and nearby normal mucosa from six colorectal adenocarcinoma patients were compared by 2D DIGE [48]. Over 1500 spots were resolved and quantitatively analyzed by this method, yielding 52 discrete proteins, identified by mass spectrometry, that showed consistent differences between normal and cancerous tissue [48]. 2D DIGE has also been used to reveal differential protein expression in infiltrating ductal carcinoma of the breast (IDCA) [49]. The comparative power of 2D DIGE was combined with the specificity of LCM to discover potential markers of esophageal carcinoma [50]. Cancerous and normal squamous epithelial cells were dissected from frozen esophageal tissue sections and proteins were compared by DIGE. Numerous protein spots were found to vary more than three-fold in expression and are candidate markers of esophageal cancer. Two of the proteins were identified by mass spectrometry and their differential expression in normal and cancer cells was confirmed by immunoblotting, demonstrating the feasibilty of this approach [50]. There is a great deal of interest at the present time in developing gel-free systems for protein analysis because of their potential for multiplexing [51] and [52]. An analogy may be made to DNA sequencing, notably as utilized in the genome project which received a considerable boost when the switch from gel-based approaches to a gel-free technology took place. Multi-modular combinations of HPLC, liquid-phase isoelectric focusing (IEF), and capillary electrophoresis (CE) provide various options to develop high-resolution orthogonal 2D liquid phase-based strategies for the separation of complex mixtures of proteins. Such strategies include SEC–CE or SEC–RPLC as used by Jorgenson's group to fractionate protein mixtures in Escherichia coli lysates [53] and [54]. Le Coutre analyzed E. coli membrane proteins with affinity chromatography, followed by on-line RPLC–MS [55]. Feng reported the use of ion-exchange chromatography (IEC) followed by on-line eight-channel parallel RPLC–ESI-MS to purify recombinant proteins in a high-throughput fashion [56]. A major advantage of liquid separations is that proteins are maintained in solution that allows on-line intact protein characterization by MS as well as protein recovery. Our group developed a novel 2D IEF-RPLC system to fractionate or resolve large numbers of cellular proteins. These protein fractions were recovered and applied to protein biochips to determine their antigenicity in cancer [52], [57] and [58]. The capacity of the 2D separation system in practice is limited to resolving no more than 10,000 protein forms according to Giddings’ model, if each dimension has a capacity of 100; that capacity may not be sufficient to achieve complete resolution of a cell or tissue proteome. It is, therefore, beneficial to reduce sample complexity as much as possible. With the emergence of soft ionization techniques such as fast atom bombardment (FAB), matrix-assisted laser desorption ionization (MALDI), and electrospray ionization (ESI) more than a decade ago [59], [60] and [61], biological mass spectrometry (Bio-MS) has become a standard tool for protein analysis [62]. Biological samples subjected to mass spectrometry consist of three major types: (1) tissues; (2) cell populations; and (3) biological fluids. Innovations in mass spectrometry continue to have a substantial impact on proteomics. Nano-electrospray techniques [63] and [64] combined with a hybrid quadrupole time-of-flight mass spectrometer tandem mass analyzer (ESI Q-TOF MS/MS) enable extensive fragmentations to produce collision-induced dissociation (CID) spectra that allow unambiguous protein identification by peptide sequence tags through protein sequence database searches. High-throughput proteomic analysis may also be performed with a MALDI Q-TOF MS/MS tandem instrument [65] and [66] and MALDI TOF-TOF MS/MS tandem mass spectrometry [67]. A new ion source for Fourier-transform ion cyclotron resonance mass spectrometry (FTICR-MS) enables quick changes between MALDI and ESI modes [68]. Mass spectrometry in conjunction with proteomics, has been utilized primarily for protein identification. However, it is possible to profile tissues and biological fluids directly using mass spectrometry. The potential of mass spectrometry to yield comprehensive profiles of peptides and proteins in biological fluids without the need to first carry out protein separations has attracted interest. In principle, such an approach would be highly suited for clinical applications because of reduced sample requirements and high throughput. This approach is currently popularized, particularly for serum analysis, by the technology referred to as surface-enhanced laser desorption ionization (SELDI) [7]. Proteins from a patient sample are captured by various types of surfaces with different properties including adsorption, partition, electrostatic interaction, or affinity chromatography. Although such surfaces are referred to as “chips”, they should not be confused with microarrays as they do not involve any type of arraying. Aside from the use of SELDI, the direct analysis of tissues or biological fluids may be simply accomplished using standard matrix-assisted laser desorption ionization without the use of proprietary surfaces. Some quite noteworthy findings have been reported using SELDI. They include the ability to accurately diagnose ovarian, prostate, breast, and other types of cancer with minimal sample requirement and with high throughput. A study of ovarian cancer that has attracted considerable attention demonstrated the ability of SELDI in combination with an algorithm, to correctly identify all cancer patients, including those with limited stage I disease [69]. MALDI mass spectrometry has been utilized in an innovative fashion to profile tissues in situ. A recent study utilized this approach to classify lung tumors based on their proteomic profile [70]. Proteomic spectra were obtained for 79 lung tumors and 14 normal lung tissues. More than 1600 protein peaks were detected from histologically selected 1 mm diameter regions of single frozen sections from each tissue. Class-prediction models based on differentially expressed peaks enabled the classification of lung cancer histologies, distinction between primary tumors and metastases to the lung from other sites, and classification of nodal involvement with 85% accuracy. The major drawbacks of direct analysis of tissues or biological fluids by MALDI or SELDI are the preferential detection of proteins with a lower molecular mass and the difficulty in determining the identity of proteins whose masses are measured because of lack of correspondence between the masses detected and those predicted for corresponding proteins, due to post-translational modifications. There have been some concerns regarding the significance of the diagnostic patterns uncovered using SELDI because the molecules monitored in serum using this approach are likely to be present at concentrations many fold higher than traditional cancer biomarkers. Such markers, therefore, are unlikely to originate from the tumor and thus are considered to be epiphenomena of cancer produced by other organs in response either to the presence of cancer or to a generalized condition of the cancer patient such as debilitation or acute-phase reaction [71]. Thus, the role of MALDI and MALDI surfaces in profiling biological fluids remains to be determined. Antibody and protein arrays offer an attractive complement to separation and mass spectrometry methods for comparative proteomics research. Various technologies for probing binding interactions on arrays of immobilized antibodies, proteins or peptides are in development and use. Each technology has its own advantages, disadvantages, and optimal applications. These methods and their applications in comparative proteomics are reviewed here. Antibody arrays are useful for measuring the abundance of multiple, specific proteins in low sample volumes. Antibody array methods are particularly well suited to profiling many candidate biomarkers in large sets of biological samples, such as serum, to identify individual proteins or groups of proteins that statistically associate with a particular condition. The multiplex capability of antibody arrays allows both the efficient testing of many individual candidate markers and also the evaluation of the use of multiple markers in combination. The use of multiple markers in combination may in some cases have higher diagnostic accuracy than individual markers. Since microarray experiments are generally rapid to run and easy to analyze, large clinical studies are possible, enabling the validation of multiple new or candidate markers. Various technological implementations of antibody array experiments have been demonstrated. A variety of substrates and methods of antibody attachment have been used, such as passive adsorption of antibodies onto membranes [72], [73], [74] and [75], poly-l-lysine coated glass [76] and [77], or hydrogels [75] and [78], covalent linkage to amine-reactive coated glass [77], [79] and [80], or linkage of biotinylated antibodies to streptavidin-coated glass [81]. The best choice of surface is not yet firmly established and may depend on the application or the detection method used. Factors to consider in evaluating surfaces are reproducibility and consistency in both the background and the signals, the signal levels relative to the background levels, and the ability of the surface to maintain the antibodies in their properly folded, reactive forms. A variety of detection formats also have been employed. Sandwich assays, using a pair of antibodies specific for every target, have been developed in a chip format for the multiplexed detection of cytokines [80], [82], [83], [84] and [85]. Sandwich assays have the potential for very high specificity and sensitivity of detection. Rolling-circle amplification (RCA) [82] and [84], tyramide signal amplification [80], and fluorescence [85] have been used as detection methods for multiplexed sandwich assays. RCA significantly enhances fluorescence signal and reduces detection limits in comparison to non-amplified fluorescence methods. Its advantages for microarray assays are that it is an isothermal process and that the amplification products are covalently attached to the spot of origin—a factor important for planar, multiplexed assays. An alternative to sandwich assays are “label-based” assays, in which the proteins to be detected are labeled with tags that allow detection after capture by immobilized antibodies. A benefit of the label-based assays is that only one antibody per target is required (as opposed to two antibodies per target for a sandwich assay), making the development and testing of assays for new targets straightforward. This capability will be important for research in which multiple rare or newly discovered proteins are to be probed. Another advantage of the label-based method is that competitive assays are possible, since two different samples, a test sample and a reference sample, can be co-incubated on an array. Competitive assays can lessen the requirement to match the concentrations of analytes to a particular linear range for each analyte. This feature may be important when a multiplexed assay measures different analytes in widely varying concentration ranges [86]. Competitive assays could be performed using a labeled reference sample and an unlabeled test sample [86], or both the test and reference samples could be labeled, each with its own distinguishable label [76]. Labeled proteins have been detected by fluorescence [74], [78] and [87], RCA [75], or colorimetric methods [73]. RCA detection of labeled proteins was developed as a means to improve the detection sensitivity of the label-based antibody microarray assay [75]. Two pools of proteins were, respectively, labeled with biotin and digoxigenin and co-incubated on antibody microarrays. The biotin-labeled proteins were detected by RCA with green fluorescence and the digoxigenin-labeled proteins were detected by RCA with red fluorescence. The fluorescence was enhanced up to 30-fold relative to non-amplified fluorescence, and the reproducible detection of low-abundance proteins in serum samples was demonstrated. Several reports have demonstrated the application of antibody microarrays to cancer proteomics research. Portions of frozen tumor specimens isolated by laser capture microdissection (LCM) were probed by antibody arrays to identify proteins both in the tumor tissue and in the surrounding stroma that had levels correlating with advancement of disease [73]. A similar study probed proteins in LCM-isolated tissue from hepatocellular carcinoma tumors and the surrounding environment, identifying proteins that may be associated with that disease [74]. Proteins in cultured colon carcinoma cells were profiled by antibody arrays to identify proteins that may be regulated in response to radiation exposure [77]. In a novel application, microarrays of antibodies spotted onto nitrocellulose specifically captured cells expressing specific membrane antigens [72]. Suspensions of leukocytes isolated from the blood of leukemia patients were incubated on microarrays of antibodies recognizing various CD antigens, and quantification of the bound cells by dark field microscopy identified antigens that accurately discriminated CLL lymphocytes from normal lymphocytes. Another useful application for antibody arrays is to profile sets of proteins in blood serum or other readily sampled biological fluids to identify candidate markers for cancer diagnosis. Such an application was demonstrated in the reproducible and accurate measurement of multiple proteins in serum samples from prostate cancer patients and controls [78]. The clustering of antibody measurements and ELISA measurements from four replicate experimental sets measuring 53 different serum samples (Fig. 2) shows that replicate microarray measurements are highly reproducible and that the ELISA and microarray measurements substantially agree. A set of five candidate biomarkers was derived from the study with statistically different levels between cases and controls. This result established the feasibility and value of multiplexed serum biomarker detection. The further application and development of the above methods are sure to yield valuable results in cancer proteomics research. Protein and peptide arrays are complementary to antibody arrays. They are useful for probing the interactions of protein and peptides with other antibodies, proteins, or other molecules. The methods and applications of these technologies are discussed below. “Reverse phase” protein arrays recently have proven useful for probing the abundance of specific proteins in sets of biological samples. Protein lysates from cell culture or tissue samples are spotted in microarrays onto nitrocellulose membranes. A labeled antibody specific for a particular protein is incubated on a microarray, and quantification of the bound antibody reveals the amount of that protein in each of the samples. Therefore, reverse phase array experiments measure a single protein in many samples, in contrast to antibody array experiments that measure many proteins in one sample. Several demonstrations of the use of the technology for profiling proteins in cancer have appeared. The technology was used to measure proteins relevant to apoptosis pathways in malignant and normal prostate tissue [88], to investigate defects in signaling in ovarian cancer tissues [89], and to profile multiple proteins in 60 cancer cell lines used by the National Cancer Institute to screen compounds for anticancer activity [90]. Protein arrays also have been made from purified or semi-purified proteins (as opposed to whole-cell lysates). High-throughput expression and purification methods were used to produce proteins, and the arrayed proteins were used to probe specific binding interactions. One study looked at the interactions of calmodulin- and phospholipid-interacting proteins with arrayed yeast proteins that had been expressed and purified from 5800 open reading frames [91]. An efficient method to produce arrays of proteins is to spot individual bacterial colonies of a cDNA library onto membranes, induce the colonies for protein expression, and lyse the cells on the membrane [92], [93] and [94]. These arrays may be most useful for measuring protein–protein and protein–small molecule interactions, and may eventually find application in diagnostics and comparative proteomics. Another method to produce proteins for arrays is to separate whole-cell lysates into the component protein fractions using multi-dimensional liquid chromatography. Multiple modes of separation in succession (for example, ion-exchange chromatography followed by reverse phase) have high resolving power, and liquid phase methods allow convenient fraction collection. As previously suggested [58], protein fractions separated by liquid chromatography and spotted onto microarrays could be used for the parallel interrogation of thousands of proteins. An advantage of using proteins taken from their native states is that modifications and alterations to the proteins are present, in contrast to proteins expressed in foreign systems, such as in bacterial or insect cells, that may not have correct post-translational modifications. A valuable application of such microarrays is the study of immune responses in cancer patients [95], [96] and [97]. Arrays of tumor-derived proteins were incubated with sera from cancer patients and controls, and the level of antibody binding to each protein fraction identified proteins that may commonly elicit immune responses in prostate cancer [95] and [96] and in lung cancer [97]. Circulating tumor-specific antibodies may be valuable for cancer diagnostics. Peptide microarrays also have been powerfully used to study and characterize immune responses [98] and [99]. Sets of peptides from candidate targets of autoantibodies in various autoimmune diseases were collected and arrayed, and the arrays were incubated with sera from patients with autoimmune diseases such as autoimmune encephalomyelitis or multiple sclerosis. The detection of antibody binding at each peptide revealed the specificity of the autoimmune response in each patient. The mapping of immunoreactivity in the autoimmune patients could be used for diagnosis, prognosis, and tailoring of antigen-specific tolerizing therapy. The above applications and technologies demonstrate the value of antibody, protein and peptide microarray methods. Further improvements to the technologies and dissemination of the methods should broaden their use and impact in biological research. Work from the authors’ laboratories is funded, in part, by the National Cancer Institute Early Detection Research Network. L.F.S. receives additional support from NIH grant AI055988 and Nucleonics Inc. B.B.H. acknowledges support from the Cancer Research and Prevention Foundation, the Department of Defense (DAMD17-03-1-0044) and the Van Andel Research Institute.", "publicationTitle": "Journal of Chromatography B", "publisher": "", "place": "", "date": "February 2005", "volume": "815", "issue": "1-2", "section": "", "partNumber": "", "partTitle": "", "pages": "275-284", "series": "", "seriesTitle": "", "seriesText": "", "journalAbbreviation": "J. Chromatogr. B Analyt. Technol. Biomed. Life Sci.", "DOI": "10.1016/j.jchromb.2004.10.072", "citationKey": "", "url": "http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6X0P-4DWGWGT-2&_user=779890&_rdoc=1&_fmt=&_orig=search&_sort=d&view=c&_acct=C000043220&_version=1&_urlVersion=0&_userid=779890&md5=c73b6905592c751fa48463573a7e1add", "accessDate": "2009-04-15T12:19:40Z", "PMID": "", "PMCID": "", "ISSN": "", "archive": "", "archiveLocation": "", "shortTitle": "", "language": "", "libraryCatalog": "ScienceDirect", "callNumber": "", "rights": "", "extra": "", "tags": [ { "tag": "panel of biomarkers" } ], "collections": [ "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T09:52:36Z" } }, { "key": "IRXZ2DDT", "version": 1, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/IRXZ2DDT", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/IRXZ2DDT", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "Stead et al.", "parsedDate": "2008-03-01", "numChildren": 0 }, "data": { "key": "IRXZ2DDT", "version": 1, "itemType": "journalArticle", "title": "Information quality in proteomics", "creators": [ { "creatorType": "author", "firstName": "David A.", "lastName": "Stead" }, { "creatorType": "author", "firstName": "Norman W.", "lastName": "Paton" }, { "creatorType": "author", "firstName": "Paolo", "lastName": "Missier" }, { "creatorType": "author", "firstName": "Suzanne M.", "lastName": "Embury" }, { "creatorType": "author", "firstName": "Cornelia", "lastName": "Hedeler" }, { "creatorType": "author", "firstName": "Binling", "lastName": "Jin" }, { "creatorType": "author", "firstName": "Alistair J. P.", "lastName": "Brown" }, { "creatorType": "author", "firstName": "Alun", "lastName": "Preece" } ], "abstractNote": "Proteomics, the study of the protein complement of a biological system, is generating increasing quantities of data from rapidly developing technologies employed in a variety of different experimental workflows. Experimental processes, e.g. for comparative 2D gel studies or LC-MS/MS analyses of complex protein mixtures, involve a number of steps: from experimental design, through wet and dry lab operations, to publication of data in repositories and finally to data annotation and maintenance. The presence of inaccuracies throughout the processing pipeline, however, results in data that can be untrustworthy, thus offsetting the benefits of high-throughput technology. While researchers and practitioners are generally aware of some of the information quality issues associated with public proteomics data, there are few accepted criteria and guidelines for dealing with them. In this article, we highlight factors that impact on the quality of experimental data and review current approaches to information quality management in proteomics. Data quality issues are considered throughout the lifecycle of a proteomics experiment, from experiment design and technique selection, through data analysis, to archiving and sharing.", "publicationTitle": "Briefings in Bioinformatics", "publisher": "", "place": "", "date": "Mars 1, 2008", "volume": "9", "issue": "2", "section": "", "partNumber": "", "partTitle": "", "pages": "174-188", "series": "", "seriesTitle": "", "seriesText": "", "journalAbbreviation": "Brief Bioinform.", "DOI": "10.1093/bib/bbn004", "citationKey": "", "url": "http://bib.oxfordjournals.org/cgi/content/abstract/9/2/174", "accessDate": "2008-04-18T15:16:59Z", "PMID": "", "PMCID": "", "ISSN": "", "archive": "", "archiveLocation": "", "shortTitle": "", "language": "", "libraryCatalog": "HighWire", "callNumber": "", "rights": "", "extra": "", "tags": [], "collections": [ "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T09:52:36Z" } }, { "key": "E3EWZPRR", "version": 1, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/E3EWZPRR", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/E3EWZPRR", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "Salzberg", "parsedDate": "1997", "numChildren": 0 }, "data": { "key": "E3EWZPRR", "version": 1, "itemType": "journalArticle", "title": "On Comparing Classifiers: Pitfalls to Avoid and a Recommended Approach", "creators": [ { "creatorType": "author", "firstName": "Steven", "lastName": "Salzberg" } ], "abstractNote": "An important component of many data mining projects is finding a good classification algorithm, a process that requires very careful thought about experimental design. If not done very carefully, comparative studies of classification and other types of algorithms can easily result in statistically invalid conclusions. This is especially true when one is using data mining techniques to analyze very large databases, which inevitably contain some statistically unlikely data. This paper describes several phenomena that can, if ignored, invalidate an experimental comparison. These phenomena and the conclusions that follow apply not only to classification, but to computational experiments in almost any aspect of data mining. The paper also discusses why comparative analysis is more important in evaluating some types of algorithms than for others, and provides some suggestions about how to avoid the pitfalls suffered by many experimental studies.", "publicationTitle": "Data Mining and Knowledge Discovery", "publisher": "", "place": "", "date": "1997", "volume": "1", "issue": "3", "section": "", "partNumber": "", "partTitle": "", "pages": "317-328", "series": "", "seriesTitle": "", "seriesText": "", "journalAbbreviation": "", "DOI": "10.1023/A:1009752403260", "citationKey": "", "url": "http://dx.doi.org/10.1023/A:1009752403260", "accessDate": "2009-10-06T07:16:32Z", "PMID": "", "PMCID": "", "ISSN": "", "archive": "", "archiveLocation": "", "shortTitle": "On Comparing Classifiers", "language": "", "libraryCatalog": "SpringerLink", "callNumber": "", "rights": "", "extra": "", "tags": [], "collections": [ "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T09:52:36Z" } }, { "key": "G7W4TGJN", "version": 1, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/G7W4TGJN", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/G7W4TGJN", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "Saeys et al.", "parsedDate": "2007-10-01", "numChildren": 0 }, "data": { "key": "G7W4TGJN", "version": 1, "itemType": "journalArticle", "title": "A review of feature selection techniques in bioinformatics", "creators": [ { "creatorType": "author", "firstName": "Yvan", "lastName": "Saeys" }, { "creatorType": "author", "firstName": "Inaki", "lastName": "Inza" }, { "creatorType": "author", "firstName": "Pedro", "lastName": "Larranaga" } ], "abstractNote": "Feature selection techniques have become an apparent need in many bioinformatics applications. In addition to the large pool of techniques that have already been developed in the machine learning and data mining fields, specific applications in bioinformatics have led to a wealth of newly proposed techniques. In this article, we make the interested reader aware of the possibilities of feature selection, providing a basic taxonomy of feature selection techniques, and discussing their use, variety and potential in a number of both common as well as upcoming bioinformatics applications. Contact: yvan.saeys@psb.ugent.be Supplementary information: http://bioinformatics.psb.ugent.be/supplementary_data/yvsae/fsreview", "publicationTitle": "Bioinformatics", "publisher": "", "place": "", "date": "Octobre 1, 2007", "volume": "23", "issue": "19", "section": "", "partNumber": "", "partTitle": "", "pages": "2507-2517", "series": "", "seriesTitle": "", "seriesText": "", "journalAbbreviation": "", "DOI": "10.1093/bioinformatics/btm344", "citationKey": "", "url": "http://bioinformatics.oxfordjournals.org/cgi/content/abstract/23/19/2507", "accessDate": "2007-11-13T07:41:49Z", "PMID": "", "PMCID": "", "ISSN": "", "archive": "", "archiveLocation": "", "shortTitle": "", "language": "", "libraryCatalog": "HighWire", "callNumber": "", "rights": "", "extra": "", "tags": [ { "tag": "feature selection" }, { "tag": "review" } ], "collections": [ "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T09:52:36Z" } }, { "key": "87TSDID7", "version": 1, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/87TSDID7", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/87TSDID7", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "Ross et al.", "parsedDate": "2008-10-15", "numChildren": 0 }, "data": { "key": "87TSDID7", "version": 1, "itemType": "journalArticle", "title": "Chemosensitivity and stratification by a five monoclonal antibody immunohistochemistry test in the NSABP B14 and B20 trials", "creators": [ { "creatorType": "author", "firstName": "Douglas T", "lastName": "Ross" }, { "creatorType": "author", "firstName": "Chung-Yeul", "lastName": "Kim" }, { "creatorType": "author", "firstName": "Gong", "lastName": "Tang" }, { "creatorType": "author", "firstName": "Olga L", "lastName": "Bohn" }, { "creatorType": "author", "firstName": "Rodney A", "lastName": "Beck" }, { "creatorType": "author", "firstName": "Brian Z", "lastName": "Ring" }, { "creatorType": "author", "firstName": "Robert S", "lastName": "Seitz" }, { "creatorType": "author", "firstName": "Soonmyung", "lastName": "Paik" }, { "creatorType": "author", "firstName": "Joseph P", "lastName": "Costantino" }, { "creatorType": "author", "firstName": "Norman", "lastName": "Wolmark" } ], "abstractNote": "PURPOSE: To test the association between risk stratification and outcome in a prospectively designed, blinded retrospective study using tissue arrays of available paraffin blocks from the estrogen receptor-expressing, node-negative samples from the National Surgical Adjuvant Breast and Bowel Project B14 and B20 tamoxifen and chemotherapy trials. EXPERIMENTAL DESIGN: Tissue arrays were stained by immunohistochemistry targeting p53, NDRG1, SLC7A5, CEACAM5, and HTF9C. Risk stratification was done using predefined scoring rules, algorithm for combining scores, and cutoff points for low-risk, moderate-risk, and high-risk patient strata. RESULTS: In a univariate Cox model, this test was significantly associated with recurrence-free interval [HR, 1.3 (95% confidence interval, 1.1-1.6); P = 0.006]. In a multivariate model it contributed information independent of age, tumor size, and menopausal status (P = 0.007). The Kaplan-Meier estimates of the proportion of recurrence-free after 10 years were 73%, 86%, and 85% for the high-risk, moderate-risk, and low-risk groups (P = 0.001). The Kaplan-Meier estimates of the breast-cancer-specific-death rate were 23%, 10%, and 9% (P < 0.0001). Exploratory analysis in patients >/=60 years old showed Kaplan-Meier estimates of the proportion of recurrence-free of 78%, 89%, and 92%. Both high-risk and low-risk groups showed significant improvement on treatment with cytotoxic chemotherapy. CONCLUSIONS: Immunohistochemistry using five monoclonal antibodies assigns breast cancer patients to a risk index that was significantly associated with clinical outcome among the estrogen receptor-expressing, node-negative tamoxifen-treated patients. It seems that the test may be able to identify patients who have greater absolute benefit from adjuvant chemotherapy compared with unstratified patient populations. Exploratory analysis suggests that this test will be most useful in clinical decision making for postmenopausal patients.", "publicationTitle": "Clinical Cancer Research", "publisher": "", "place": "", "date": "Oct 15, 2008", "volume": "14", "issue": "20", "section": "", "partNumber": "", "partTitle": "", "pages": "6602-6609", "series": "", "seriesTitle": "", "seriesText": "", "journalAbbreviation": "Clin. Cancer Res", "DOI": "10.1158/1078-0432.CCR-08-0647", "citationKey": "", "url": "http://www.ncbi.nlm.nih.gov/pubmed/18927301", "accessDate": "2009-10-02T14:34:32Z", "PMID": "", "PMCID": "", "ISSN": "1078-0432", "archive": "", "archiveLocation": "", "shortTitle": "", "language": "", "libraryCatalog": "NCBI PubMed", "callNumber": "", "rights": "", "extra": "PMID: 18927301", "tags": [ { "tag": "panel of biomarkers" } ], "collections": [ "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T09:52:36Z" } }, { "key": "7W73H3AV", "version": 1, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/7W73H3AV", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/7W73H3AV", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "Rosengart et al.", "parsedDate": "2007", "numChildren": 0 }, "data": { "key": "7W73H3AV", "version": 1, "itemType": "journalArticle", "title": "Prognostic Factors for Outcome in Patients With Aneurysmal Subarachnoid Hemorrhage", "creators": [ { "creatorType": "author", "firstName": "Axel J.", "lastName": "Rosengart" }, { "creatorType": "author", "firstName": "Kim E.", "lastName": "Schultheiss" }, { "creatorType": "author", "firstName": "Jocelyn", "lastName": "Tolentino" }, { "creatorType": "author", "firstName": "R. Loch", "lastName": "Macdonald" } ], "abstractNote": "Background and Purpose The purpose of this study was to describe prognostic factors for outcome in a large series of patients undergoing neurosurgical clipping of aneurysms after subarachnoid hemorrhage (SAH). Methods Data were analyzed from 3567 patients with aneurysmal SAH enrolled in 4 randomized clinical trials between 1991 and 1997. The primary outcome measure was the Glasgow outcome scale 3 months after SAH. Multivariable logistic regression with backwards selection and Cox proportional hazards regression models were derived to define independent predictors of unfavorable outcome. Results In multivariable analysis, unfavorable outcome was associated with increasing age, worsening neurological grade, ruptured posterior circulation aneurysm, larger aneurysm size, more SAH on admission computed tomography, intracerebral hematoma or intraventricular hemorrhage, elevated systolic blood pressure on admission, and previous diagnosis of hypertension, myocardial infarction, liver disease, or SAH. Variables present during hospitalization associated with poor outcome were temperature >38{degrees}C 8 days after SAH, use of anticonvulsants, symptomatic vasospasm, and cerebral infarction. Use of prophylactic or therapeutic hypervolemia or prophylactic-induced hypertension were associated with a lower risk of unfavorable outcome. Time from admission to surgery was significant in some models. Factors that contributed most to variation in outcome, in descending order of importance, were cerebral infarction, neurological grade, age, temperature on day 8, intraventricular hemorrhage, vasospasm, SAH, intracerebral hematoma, and history of hypertension. Conclusions Although most prognostic factors for outcome after SAH are present on admission and are not modifiable, a substantial contribution to outcome is made by factors developing after admission and which may be more easily influenced by treatment.", "publicationTitle": "Stroke", "publisher": "", "place": "", "date": "Août 1, 2007", "volume": "38", "issue": "8", "section": "", "partNumber": "", "partTitle": "", "pages": "2315-2321", "series": "", "seriesTitle": "", "seriesText": "", "journalAbbreviation": "", "DOI": "10.1161/STROKEAHA.107.484360", "citationKey": "", "url": "http://stroke.ahajournals.org/cgi/content/abstract/38/8/2315", "accessDate": "2007-11-16T13:32:14Z", "PMID": "", "PMCID": "", "ISSN": "", "archive": "", "archiveLocation": "", "shortTitle": "", "language": "", "libraryCatalog": "HighWire", "callNumber": "", "rights": "", "extra": "", "tags": [ { "tag": "Cox proportional hazards" }, { "tag": "logistic regression" }, { "tag": "sah" } ], "collections": [ "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T09:52:36Z" } }, { "key": "XFA28FFU", "version": 1, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/XFA28FFU", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/XFA28FFU", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "Ressom et al.", "parsedDate": "2007-03-01", "numChildren": 0 }, "data": { "key": "XFA28FFU", "version": 1, "itemType": "journalArticle", "title": "Peak selection from MALDI-TOF mass spectra using ant colony optimization", "creators": [ { "creatorType": "author", "firstName": "H. W.", "lastName": "Ressom" }, { "creatorType": "author", "firstName": "R. S.", "lastName": "Varghese" }, { "creatorType": "author", "firstName": "S. K.", "lastName": "Drake" }, { "creatorType": "author", "firstName": "G. L.", "lastName": "Hortin" }, { "creatorType": "author", "firstName": "M.", "lastName": "Abdel-Hamid" }, { "creatorType": "author", "firstName": "C. A.", "lastName": "Loffredo" }, { "creatorType": "author", "firstName": "R.", "lastName": "Goldman" } ], "abstractNote": "Motivation: Due to the large number of peaks in mass spectra of low-molecular-weight (LMW) enriched sera, a systematic method is needed to select a parsimonious set of peaks to facilitate biomarker identification. We present computational methods for matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) spectral data preprocessing and peak selection. In particular, we propose a novel method that combines ant colony optimization (ACO) with support vector machines (SVM) to select a small set of useful peaks. Results: The proposed hybrid ACO-SVM algorithm selected a panel of eight peaks out of 228 candidate peaks from MALDI-TOF spectra of LMW enriched sera. An SVM classifier built with these peaks achieved 94% sensitivity and 100% specificity in distinguishing hepatocellular carcinoma from cirrhosis in a blind validation set of 69 samples. Area under the receiver operating characteristic (ROC) curve was 0.996. The classification capability of these peaks is compared with those selected by the SVM-recursive feature elimination method. Availability: Supplementary material and MATLAB scripts to implement the methods described in this article are available at http://microarray.georgetown.edu/web/files/bioinf.htm Contact: hwr@georgetown.edu Supplementary information: Supplementary data are available at Bioinformatics online.", "publicationTitle": "Bioinformatics", "publisher": "", "place": "", "date": "March 1, 2007", "volume": "23", "issue": "5", "section": "", "partNumber": "", "partTitle": "", "pages": "619-626", "series": "", "seriesTitle": "", "seriesText": "", "journalAbbreviation": "", "DOI": "10.1093/bioinformatics/btl678", "citationKey": "", "url": "http://bioinformatics.oxfordjournals.org/cgi/content/abstract/23/5/619", "accessDate": "2009-05-08T14:06:09Z", "PMID": "", "PMCID": "", "ISSN": "", "archive": "", "archiveLocation": "", "shortTitle": "", "language": "", "libraryCatalog": "HighWire", "callNumber": "", "rights": "", "extra": "", "tags": [], "collections": [ "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T09:52:36Z" } }, { "key": "2SN2P4F3", "version": 1, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/2SN2P4F3", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/2SN2P4F3", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "Reddy et al.", "parsedDate": "2008", "numChildren": 0 }, "data": { "key": "2SN2P4F3", "version": 1, "itemType": "journalArticle", "title": "Logical Analysis of Data (LAD) model for the early diagnosis of acute ischemic stroke", "creators": [ { "creatorType": "author", "firstName": "Anupama", "lastName": "Reddy" }, { "creatorType": "author", "firstName": "Honghui", "lastName": "Wang" }, { "creatorType": "author", "firstName": "Hua", "lastName": "Yu" }, { "creatorType": "author", "firstName": "Tiberius O", "lastName": "Bonates" }, { "creatorType": "author", "firstName": "Vimla", "lastName": "Gulabani" }, { "creatorType": "author", "firstName": "Joseph", "lastName": "Azok" }, { "creatorType": "author", "firstName": "Gerard", "lastName": "Hoehn" }, { "creatorType": "author", "firstName": "Peter L", "lastName": "Hammer" }, { "creatorType": "author", "firstName": "Alison E", "lastName": "Baird" }, { "creatorType": "author", "firstName": "King C", "lastName": "Li" } ], "abstractNote": "BACKGROUND:Strokes are a leading cause of morbidity and the first cause of adult disability in the United States. Currently, no biomarkers are being used clinically to diagnose acute ischemic stroke. A diagnostic test using a blood sample from a patient would potentially be beneficial in treating the disease.RESULTS:A classification approach is described for differentiating between proteomic samples of stroke patients and controls, and a second novel predictive model is developed for predicting the severity of stroke as measured by the National Institutes of Health Stroke Scale (NIHSS). The models were constructed by applying the Logical Analysis of Data (LAD) methodology to the mass peak profiles of 48 stroke patients and 32 controls. The classification model was shown to have an accuracy of 75% when tested on an independent validation set of 35 stroke patients and 25 controls, while the predictive model exhibited superior performance when compared to alternative algorithms. In spite of their high accuracy, both models are extremely simple and were developed using a common set consisting of only 3 peaks.CONCLUSION:We have successfully identified 3 biomarkers that can detect ischemic stroke with an accuracy of 75%. The performance of the classification model on the validation set and on cross-validation does not deteriorate significantly when compared to that on the training set, indicating the robustness of the model. As in the case of the LAD classification model, the results of the predictive model validate the function constructed on our support-set for approximating the severity scores of stroke patients. The correlation and root mean absolute error of the LAD predictive model are consistently superior to those of the other algorithms used (Support vector machines, C4.5 decision trees, Logistic regression and Multilayer perceptron).", "publicationTitle": "BMC Medical Informatics and Decision Making", "publisher": "", "place": "", "date": "2008", "volume": "8", "issue": "1", "section": "", "partNumber": "", "partTitle": "", "pages": "30", "series": "", "seriesTitle": "", "seriesText": "", "journalAbbreviation": "BMC Med. Inform. Decis. Mak.", "DOI": "10.1186/1472-6947-8-30", "citationKey": "", "url": "http://www.biomedcentral.com/1472-6947/8/30", "accessDate": "2009-01-28T13:26:44Z", "PMID": "", "PMCID": "", "ISSN": "1472-6947", "archive": "", "archiveLocation": "", "shortTitle": "", "language": "", "libraryCatalog": "BioMed Central and More", "callNumber": "", "rights": "", "extra": "", "tags": [ { "tag": "Logical analysis of data" }, { "tag": "panel of biomarkers" }, { "tag": "seldi" }, { "tag": "stroke" } ], "collections": [ "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T09:52:36Z" } }, { "key": "38T8GSM8", "version": 1, "library": { "type": "group", "id": 62016, "name": "xavier.robin", "links": { "alternate": { "href": "https://www.zotero.org/groups/xavier.robin", "type": "text/html" } } }, "links": { "self": { "href": "https://api.zotero.org/groups/62016/items/38T8GSM8", "type": "application/json" }, "alternate": { "href": "https://www.zotero.org/groups/xavier.robin/items/38T8GSM8", "type": "text/html" } }, "meta": { "createdByUser": { "id": 12953, "username": "Calimo", "name": "Xavier Robin", "links": { "alternate": { "href": "https://www.zotero.org/calimo", "type": "text/html" } } }, "creatorSummary": "Rai and Vitzthum", "parsedDate": "2006-08", "numChildren": 0 }, "data": { "key": "38T8GSM8", "version": 1, "itemType": "journalArticle", "title": "Effects of preanalytical variables on peptide and protein measurements in human serum and plasma: implications for clinical proteomics", "creators": [ { "creatorType": "author", "firstName": "Alex J", "lastName": "Rai" }, { "creatorType": "author", "firstName": "Frank", "lastName": "Vitzthum" } ], "abstractNote": "There is a wealth of knowledge in the field of in vitro diagnostics with regard to preanalytical variables and their impact on the determination of peptide and protein analytes in human serum and plasma. This information is applicable to clinical proteomics investigations, which utilize the same sample types. Studies have demonstrated that the majority of variations and errors in in vitro diagnostics seem to occur in the preanalytical phase prior to specimen analysis. Preanalytical processes include study design, compliance of the subjects investigated, compliance of the technical staff in adherence to protocols, choice of specimens utilized and sample collection and processing. These variables can have a dramatic impact on the determination of analytes and can affect result outcomes, reproducibility and the validity of investigations. By drawing analogies to in vitro diagnostics practices, specific variables that are likely to impact the results of proteomics studies can be identified. Recognition of such variables is the first step towards their understanding and, eventually, controlling their impact. In this article, we will review preanalytical variables, provide examples for their effects on the determination of distinct peptides and proteins and discuss potential implications for clinical proteomics investigations.", "publicationTitle": "Expert Review of Proteomics", "publisher": "", "place": "", "date": "Aug 2006", "volume": "3", "issue": "4", "section": "", "partNumber": "", "partTitle": "", "pages": "409-426", "series": "", "seriesTitle": "", "seriesText": "", "journalAbbreviation": "Expert Rev Proteomics", "DOI": "10.1586/14789450.3.4.409", "citationKey": "", "url": "http://www.ncbi.nlm.nih.gov/pubmed/16901200", "accessDate": "2009-05-28T12:43:40Z", "PMID": "", "PMCID": "", "ISSN": "1744-8387", "archive": "", "archiveLocation": "", "shortTitle": "Effects of preanalytical variables on peptide and protein measurements in human serum and plasma", "language": "", "libraryCatalog": "NCBI PubMed", "callNumber": "", "rights": "", "extra": "PMID: 16901200", "tags": [], "collections": [ "QN36N5W7" ], "relations": {}, "dateAdded": "2011-12-05T09:52:36Z", "dateModified": "2011-12-05T09:52:36Z" } } ]