LRmeth: a logistic regression approach for inferring tissue of origin from tumor DNA methylation profiles

LRmeth is a novel framework to infer a tumor's tissue of origin using DNA methylation profiles. This dataset accompanies the LRmeth manuscript and contains information about the CpGs utilized in this study. All CpGs come from the Illumina 450K array, and this study largely restricted the number of CpGs to the top 1% most variable positions across 12 different tumor types. Each CpG in the spreadsheets is also annotated with the nearest gene name and genic feature. Each Excel workbook is split into "hypomethylated" and "hypermethylated" markers since this study utilized both. high_variance_cpgs.txt -- a plain text file of the 2921 high-variance CpGs used in this study. TCGA_healthy_markers_ALL.xlsx -- spreadsheet of healthy tissue markers using all ~290,000 CpGs. TCGA_DMP_markers_ALL.xlsx -- spreadsheet of DMPs by tissue using all ~290,000 CpGs. TCGA_DMP_markers_HIGHVAR.xlsx -- spreadsheet of DMPs by tissue using the high-variance CpGs only.