Abstract LB234: Identifying cancer vulnerabilities associated with changes in chromatin accessibility by simultaneously profiling hundreds of cancer cell lines with ATAC-seq

Abstract

Abstract The goal of the Cancer Dependency Map project (DepMap) is to identify the landscape of cancer vulnerabilities by systematically profiling hundreds of cell lines representing the heterogeneity of human cancer. Cancer cells are known to undergo significant alterations in the cancer genome including changes in the accessible chromatin regions (ACRs). These alterations could lead to novel cancer vulnerabilities but chromatin accessibility has not been systematically profiled in DepMap. Profiling hundreds of cell lines individually for ACRs would require large resources. Here we show that we can efficiently profile hundreds of cell lines using two single-cell approaches developed by our labs. SHARE-seq is a single-cell high-throughput combinatorial indexing method to measure both chromatin accessibility (scATAC-seq), and gene expression (scRNA-seq) in the same single cell. MIX-Seq is an approach that leverages the ability to pool hundreds of unique cancer cell lines and then apply single-cell genomics across cell populations from each cell line, resolving its identity using single nucleotide polymorphism (SNP) profiles. Here, we used a combination of MIX-seq (pooling cell lines) and SHARE-seq (scATAC-seq and RNA-seq) to simultaneously profile 500 unique cell lines covering 110 lineages within 22 cancer types at the single-cell level. Encouragingly, in a UMAP representation of the data, cancer cell lines cluster by lineage. Interestingly, we identified hundreds of thousands of peaks within cis-regulatory elements (CREs), 1/3 of which are unique to cancer cell lines and not present in normal cell types. Using published tools, we calculated an ATAC-seq gene activity prediction score per cell line. To identify cancer vulnerabilities associated with ACRs changes, we combined the ATAC-seq gene-activity predictions with RNAseq-based features to construct machine-learning models of cancer dependency scores. We found that the prediction accuracy for multiple dependencies improved with the addition of the ATAC-seq-based scores, indicating that chromatin accessibility provides information beyond gene expression. Additional analyses are underway to create other features. Overall, we have generated a large high-quality dataset to simultaneously map the chromatin accessibility in hundreds of cancer cell lines. Importantly, the chromatin accessibility adds to other omics profiling to identify biomarkers of response and understand how changes in chromatin can lead to new vulnerabilities. Citation Format: Patricia Borck, Kirsty Wienand, Fabiana Duarte, Alvin Qin, Ruochi Zhang, Juliana Babu, Simone Zhang, Samuel Maffa, Max Horlbeck, Rojesh Shrestha, Joshua Dempster, Jennifer Roth, Catarina Campbell, Jason Buenrostro, Francisca Vazquez. Identifying cancer vulnerabilities associated with changes in chromatin accessibility by simultaneously profiling hundreds of cancer cell lines with ATAC-seq [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(7_Suppl):Abstract nr LB234.