1. Signal dataset - 300 tissue-specific overexpressor genes in human and mouse genomes. Background dataset - 5,000 human and mouse underexpressor genes. Format: <accession number> <gene name> <location> 2. EI optimization: -- optimized TF tissue-specficity weights, human and mouse; -- TF occurrence and importance, human and mouse; -- Plots of 10 most important TFs, human and mouse; 3. Database of EI cENHs (candidate tissue-speicific enhancers). -- 7,187 human cENHs (text file; 773kb); -- 5,585 mouse cENHs (text file; 546kb); Format: <cENH location> <type> <list of tissue-specificity:score> <locus location>. 4. Database of EI TSDs (tissue-specific demarcations). -- 587,795 human TSDs (gzipped file; 46Mb); -- 593,549 mouse TSDs (gzipped file; 36Mb); Format: <cENH location> <type> <list of tissue-specificity:score> <locus location>. 5. Relatedness of human tissues is available as a web-table. Relatedness is defined as the percent overlap in highly expressed genes. Note: NCBI Build 35 human (hg17) and mouse (mm7) assemblies were used in the study. |