Abstract 4097: cBioPortal for cancer genomics

Abstract cBioPortal is a widely used platform for interactive visualization and analysis of large-scale multimodal cancer datasets. It provides cohort exploration tools, such as OncoPrint, mutation “lollipop” plots, survival and enrichment analyses, detailed patient-level views, and integrated variant annotations to support interpretation. The public instance of cBioPortal (https://www.cbioportal.org) serves >40,000 unique visitors globally each month. It hosts data from >500 studies, all also available through the cBioPortal Datahub. In 2025, we added 38 new studies (∼35,000 samples). We also added Tumor Break Load (TBL) scores across PCAWG, CCLE, and all 32 TCGA Pan-Cancer Atlas studies. More than 99 cBioPortal instances are deployed at institutions and companies worldwide. cBioPortal partners with AACR Project GENIE to provide access to the GENIE cohort in a dedicated instance (https://genie.cbioportal.org). Users can explore >268,000 clinically sequenced samples from 20 institutions, as well as GENIE Biopharma Collaborative (BPC) cohorts with detailed clinical annotations, including NSCLC (∼2,000 samples), colorectal cancer (∼1,500), and breast cancer (∼1,200), with more to come. Over the past year, cBioPortal progressed along two complementary directions. First, we introduced a chat-based interface for natural-language data exploration, reflecting ongoing efforts to utilize AI, specifically large language models, to augment traditional query and visualization workflows in cBioPortal. Second, we released several core platform enhancements: 1) we improved the performance for large cohorts by switching the backend database from MySQL to ClickHouse, an OLAP (Online Analytical Processing) database; 2) the Plots tab can now visualize variant allele frequencies, connect multiple samples from the same patient, and has more flexible categorical sorting; 3) variant interpretation has been strengthened through integration of AlphaMissense predictions; 4) we released a redesigned About page highlighting the year’s accomplishments and future roadmap. It is worth noting that many of these features were developed using AI-assisted technologies, which are increasingly standard practice in software engineering. cBioPortal is open source (https://github.com/cBioPortal) and developed collaboratively by groups at Memorial Sloan Kettering Cancer Center, Dana-Farber Cancer Institute, Children’s Hospital of Philadelphia, Princess Margaret Cancer Centre, Bilkent University, SE4BIO, and The Hyve. We welcome contributions from the cancer research community. Citation Format: Ino de Bruijn, Tali Mazor, Gaofei Zhao, Manda Wilson, Avery Wang, Floris Vleugels, Pim van Nierop, Henk-Jan van den Ham, S. Onur Sumer, Jessica Singh, Baby A. Satravada, Oleguer Plantalech, Angelica Ochoa, Zain-ul-Abideen Nasir, Ramyasree Madupuri, Pieter Lukasse, Aaron Lisman, James Lindsay, Xiang Li, Bryan Lai, Ritika Kundra, Priti Kumari, Sowmiyaa Kumar, Tim Kuijpers, James Ko, Zeynep Karagöz, Karthik Kalletla, Prasanna K Jagannathan, Jason Hwee, Guizela Huelsz Prince, Charles Haynes, Benjamin Gross, Zhaoyuan Fu, Ruslan Forostianov, Calla Chennault, Rima AlHamad, Ugur Dogrusoz, Allison Heath, Adam C. Resnick, Trevor J. Pugh, Chris Sander, Jianjiong Gao, Nikolaus Schultz, Ethan Cerami. cBioPortal for cancer genomics [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 4097.