Hoadley, K. A. et al. Cell 173, 291–304 (2018).

The Cancer Genome Atlas (TCGA), a program funded by the US National Institutes of Health (NIH), has ended over a decade after its inception. The ambitious project coordinated the multi-omic sequencing and clinical annotation of approximately 10,000 tumors across 33 cancer types. The resulting resource, publicly available through the NIH Genomic Data Commons (https://gdc.cancer.gov/), includes data on DNA sequence, DNA methylation, RNA and microRNA expression, and reverse-phase protein array binding, and makes it possible to rigorously characterize the molecular features of cancers. A set of 27 analysis papers associated with the final data release (https://www.cell.com/consortium/pancanceratlas) organize our current understanding by cell-of-origin patterns, oncogenic processes and signaling pathways. A flagship paper by Hoadley et al. shows that tumors cluster by histology, tissue type and anatomic origin when classified on the basis of nearly any data type. Integrated data analysis by iCluster defines various pan-cancer groupings of similar tumor types, which will help to orient future work on diagnostics and therapeutic development.