|
UVA GeneChip/Microarray Bioinformatics Core provides advanced statistical and computational support and expertise in the analysis and storage of GeneChipTM and microarray data at the University of Virginia. The GMB’s functions and services are closely associated with the GeneChip/Microarray facilities at the UVA Biomolecular Research Facility (BRF). The GMB currently offers two major services:
The first service is provided for all UVA researchers conducting microarray studies under the general institutional support and the second service can be provided under either collaborative project or consulting arrangements. SERVICES User service on the UVA GEOSS database/analysis system Researchers need to utilize the GEOSS database system as the main web site for conducting in-house GeneChipTM and microarray experiments at the University of Virginia. UVA and collaborative researchers should use this system to design, request, and manage GeneChip/microarray studies. The web interface (http://genes.med.virginia.edu) is easy to use, and the site is secure. Several useful Bioinformatics tools have been incorporated into GEOSS for efficient data analysis. Quality control (QC) and normalization of GeneChip/microarray data While the BRF provides a standard validation procedure for the quality of RNA samples, labeling, and hybridization for all GeneChip/microarray studies, it is important to address additional QC issues. The GMB performs a QC procedure by graphical and statistical examination of all array data in each study, as well as non-linear local regression normalization based on the results from the QC analysis, and evaluation of statistical indices of specificity and sensitivity. If one plans to perform a further analysis using a third-party software program, such as GeneSpring, it is recommended that one use data from the GMB service. Statistical discovery of differential gene expression The GMB provides up-to-date analysis on differential-expression discovery based on well-validated and published approaches. These approaches are not yet available in commercial software packages. Analyses include local-pooled-error (LPE) test, significant analysis of microarray (SAM), Westfall and Young’s step-down permutation test (WY), heterogeneous error model (HEM) analysis, and ANOVA modeling. We can also provide results based on recent p-value adjustment methods designed for genome-wide microarray data, such as family-wise error rate (FWER) and false discovery rate (FDR). Clustering and classification analysis Various clustering and classification analysis with high-quality graphics output can be provided depending on the needs of each study. These include hierarchical clustering, Kmeans clustering, self-organizing map, linear and quadratic discriminant analysis, logistic analysis, and support vector machine, to name a few. Advanced bioinformatics analysis Advanced bioinformatics analysis for gene expression data is available from the GMB core. The list of analyses includes (but is not limited to) SNP chip analysis, genome-wide discovery of expression hot-spots, and advanced annotation/functional information analysis. Contact HyungJun Cho, Ph.D. (hcho@virginia.edu, 4-8547) for GeneChip/microarray data analysis. Contact Jodi Kanter (jlk3x, 4-2846) for the user service on the GEOSS database and analysis system, or visit http://genes.med.virginia.edu |