Preprocessing of RSV Gene Expression Profiles for DARPA Project Hero Group Graduate Program in Bioinformatics Electrical Engineering and Computer Science University of Michigan Aut 31, 2008 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 1. Preprocessing method ===================================== Expression intensities are computed from CEL files using Michigan custom chip definition file (CCDF) version 10. This CCDF is based on human genome Build 36, Version 2, Sep 14, 2006 by National Center for Biotechnology Information (NCBI). The following parameters are used when computing probeset level expression measures: * Background correction: Robust Multichip Average (RMA) convolution * Normalization: Quantile normalization * Summarization: Median polish More details please see: Rafael. A. Irizarry, Benjamin M. Bolstad, Francois Collin, Leslie M. Cope, Bridget Hobbs and Terence P. Speed (2003), Summaries of Affymetrix GeneChip probe level data Nucleic Acids Research 31(4):e15 Note the intensity values are in Log base 2 scale. 2. Summary of samples ===================================== Total number of arrays: 164 3. Quality assessment ===================================== The residual plots show excellent data quality with only 8 chips having minor to mild artifacts. (see attached PowerPoint). 4. Files ===================================== 1) darpa.rsv.rma.(164).ccdf.txt RMA preprocessed singal intensity for complete set of 164 gene expression profiles. 2) residual.plots.zip Archive file including all the residual plots for visual examination. 3) darpa.rsv.rma.(164).residual.ppt Slides showing chips with minor to mild artifacts. 4) darpa.rsv.key.txt Key mapping chips, samples and subjects. 5. Others ===================================== If you have any question regarding this dataset, please email Prof Alfred Hero at 'hero@umich.edu' or Yongsheng Huang at 'huangys@umich.edu'.