Dimension reduction for classification with gene expression microarray data.

Jian J Dai; Linh Lieu; David Rocke

doi:10.2202/1544-6115.1147

Home/Archive/PubMed

PubMed · 16646870

Dimension reduction for classification with gene expression microarray data.

Jian J Dai·Linh Lieu·David Rocke

View original source ↗Export metadata ↓

Abstract

An important application of gene expression microarray data is classification of biological samples or prediction of clinical and other outcomes. One necessary part of multivariate statistical analysis in such applications is dimension reduction. This paper provides a comparison study of three dimension reduction techniques, namely partial least squares (PLS), sliced inverse regression (SIR) and principal component analysis (PCA), and evaluates the relative performance of classification procedures incorporating those methods. A five-step assessment procedure is designed for the purpose. Predictive accuracy and computational efficiency of the methods are examined. Two gene expression data sets for tumor classification are used in the study.

Explore related subjects

Colonic Neoplasms Gene Expression Profiling Least-Squares Analysis Leukemia Logistic Models Neoplasms Oligonucleotide Array Sequence Analysis Principal Component Analysis Regression Analysis Comparative Study Evaluation Study Journal Article Research Support, N.I.H., Extramural Research Support, Non-U.S. Gov't Research Support, U.S. Gov't, Non-P.H.S.

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Jian J Dai, Linh Lieu, David Rocke. 2006-02-24. Dimension reduction for classification with gene expression microarray data.. https://doi.org/10.2202/1544-6115.1147

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related citations

PubMed2007

Racial disparities in colon cancer. Primary care endoscopy as a tool to increase screening rates among minority patients.

Colon cancer is a condition whose far-reaching effects have been well documented nationally and within the state of South Carolina. Fortunately, the disease is amenable to both primary and secondary prevention through screening colonoscopy. Despite the conceptual simplicity of recommending colonoscopy, barriers exist to universal (or even widespread) screening. Currently the infrastructure necessary to achieve screening goals set by the American Cancer Society (ACS), the American College of Gastroenterology (ACG), and the South Carolina Department of Health and Environmental Control (DHEC) has not been established. At current rates of training gastroenterologists, the medical community will not be able to come close to achieving widespread screening. Given the discrepancy between the public health benefit of achieving the goals and the deaths that have occurred because of the resource shortfall, we propose alternative measures to screen the at-risk population for consideration. This need is most acute in the black community, in which where screening rates tend to be lower and polyps have been found to progress more quickly than among white populations. In South Carolina, one model has used primary care physicians as the labor force to provide routine screening colonoscopy for their own patients. This model makes screening much more accessible to minority patients, as the wait is shorter and the cost typically lower. In combination with a faith-based partnership with minority religious organizations, this model has begun to make needed inroads toward addressing the disparities associated with colon cancer. Cancer 2007. (c) 2006 American Cancer Society.

Colonic Neoplasms↗

PubMed2007

Influence of caseload and surgical speciality on outcome following surgery for colorectal cancer: a review of evidence. Part 2: long-term outcome.

OBJECTIVE: We reviewed recent literature to assess the impact of hospital caseload, surgeon's caseload and education on long-term outcome following colorectal cancer surgery. METHOD: We searched the MEDLINE and Cochrane Library databases for relevant literature starting from 1992. We selected hospital caseload, surgeon's caseload and surgeon's education, type of hospital, and surgeon's experience as variables of interest. Measures of outcome were recurrence-free survival and overall survival, and for rectal cancer frequency of permanent stoma. We reviewed the 34 studies according to tumour location: colonic cancer, rectal cancer, or colorectal cancer. We described the studies individually and performed a meta-analysis whenever it was considered appropriate. RESULTS: For colonic cancer, overall survival improved with increasing hospital caseload, odds ratio (OR) 1.22 [95% confidence interval (CI) 1.16-1.28], and surgeon's education. For rectal cancer, overall survival improved with increasing hospital caseload, OR 1.38 (95% CI 1.19-1.60), and, possibly by surgeon' education and experience. Cancer-free survival was strongly influenced by surgeon's education. The colostomy rate was less in high caseload hospitals, OR 0.76 (95% CI 0.68-0.85). For colorectal cancer, overall survival improved with surgeon's education. CONCLUSION: The data have provided evidence that long-term survival following colorectal cancer surgery in general improved significantly with increasing hospital caseload and surgeon's education.

Colonic Neoplasms↗

PubMed2007

Effect of directed training on reader performance for CT colonography: multicenter study.

PURPOSE: To define the interpretative performance of radiologists experienced in computed tomographic (CT) colonography and to compare it with that of novice observers who had undergone directed training, with colonoscopy as the reference standard. MATERIALS AND METHODS: Physicians at each participating center received ethical committee approval and followed the committees' requests regarding informed consent. Nine experienced radiologists, nine trained radiologists, and 10 trained technologists from nine centers read 40 CT colonographic studies selected from a data set of 51 studies and modeled to simulate a population with positive fecal occult blood test results: Studies were obtained in eight patients with cancer, 12 patients with large polyp, four patients with medium polyp, and 27 patients without colonic lesions. Findings were verified with colonoscopy. An experienced radiologist used 50 endoscopically validated studies to train novice observers before they were allowed to participate. Observers used one software platform to read studies over 2 days. Responses were collated and compared with the known diagnostic category for each subject. The number of correctly classified subjects was determined for each observer, and differences between groups were examined with bootstrap analysis. RESULTS: Overall, 28 observers read 1084 studies and detected 121 cancers, 134 large polyps, and 33 medium polyps; 448 healthy subjects were categorized correctly. Experienced radiologists detected 116 lesions; trained radiologists and technologists detected 85 and 87 lesions, respectively. Overall accuracy of experienced observers (74.2%) was significantly better than that of trained radiologists (66.6%) and technologists (63.2%). There was no significant difference (P=.33) between overall accuracy of trained radiologists and that of technologists; however, some trainees reached the mean performance achieved by experienced observers. CONCLUSION: Experienced observers interpreted CT colonographic images significantly better than did novices trained with 50 studies. On average, no difference between trained radiologists and trained technologists was found; however, individual performance was variable and some trainees outperformed some experienced observers.

Colonic Neoplasms↗