The pan-cancer proteome atlas, a mass spectrometry-based landscape for discovering tumor biology, biomarkers, and therapeutic targets

Most cancer proteomics studies to date have focused on a single cancer type. We report The Pan-Cancer Proteome Atlas (TPCPA) based on data-independent acquisition mass spectrometry, to better understand cancer biology and identify therapeutic targets and biomarkers. TPCPA includes 9,670 proteins derived from 999 primary tumors representing 22 cancer types. We describe pan-cancer and cancer type-enriched proteins with extensive external annotation, prioritizing candidate drug targets and biomarkers. Relevant for proteolysis-targeting chimeras, we identify E3-ubiquitin ligases highly expressed in specific tumor types, including HERC5 (esophageal cancer) and RNF5 (liver cancer). Co-expression analysis reveals 13 modules, including unexpected hub proteins as potential drug targets (e.g., GFPT1, LRPPRC, PINK1, DOCK2, and PTPN6). Analysis of 195 colorectal cancers identifies protein markers for RNA-based consensus molecular subtypes (CMSs) and two immune subtypes with prognostic value. We report a cancer type classifier for identification of cancers of unknown primary origin. All TPCPA data can be queried in a dedicated web resource. • Pan-cancer proteome across 18 solid and 4 liquid cancers from 999 human samples • Co-expressed pan-cancer proteins and biology with potential clinical utility • CMS proteins and immune subtypes with prognostic value in colorectal cancer • Multi-cancer classifier for identification of metastasis of unknown primary origin Knol et al. report a large pan-cancer proteome landscape based on high-throughput, single-shot mass spectrometry across 22 cancer types including both liquid ( n = 4) and solid cancers ( n = 18), revealing co-expression-based phenotypes and immune phenotypes and defining top 25 ranked proteins as potential biomarkers and drug targets for each cancer type as well as a multi-cancer classifier.