Abstract 4424: Medatree enabling single cell copy number lineage tracing and functional discovery

Abstract

Abstract Copy number alterations (CNAs) play an important role in cancer development and treatment. CNA profiles in single tumor cells preserve archaeological records of a tumor evolution, enabling retrospective lineage tracing. Although current single-cell sequencing technologies have enabled accurate, high-throughput acquisition of single-cell integer CNA profiles, existing computational methods are neither intelligent, i.e., unaware of CNA evolution mechanisms such as breakage-fusion-bridge, nor scalable to current datasets consisting of thousands of cells. To address this challenge, we developed a novel, efficient computational approach, Medatree (minimal event distance aneuploidy tree) to infer a cell lineage tree from the whole genome CNA profiles of a large number (N>1,000) of cells obtained from patient samples. A tree inferred by Medatree describes the minimal number of single-copy gain or loss events that may have occurred during the evolution, the genealogical relations amongst individual cells, and the significant CNAs associated with lineage expansion. We confirmed through simulation that Medatree substantially improves lineage tracing accuracy (e.g., 0.73 vs. 0.31) over conventional phylogeny methods (e.g., maximum parsimony). We then applied Medatree to single-cell DNA sequencing (scDNA-seq) data obtained from 20 triple-negative breast cancer patients (TNBCs), some of which had longitudinal biopsies through the course of chemotherapy. The trees constructed by Medatree corroborated well with the timing and phenotypes of the biopsies. They also revealed branches with variable CNA rates, likely associated with differentially altered genome instability. We further identified significantly copy-number altered genes through permutation of the single-cell CNA profiles. Compared to results obtained by applying convention methods (phylogenetic trees and GISTIC) on the same datasets, Medatree detected more genes (1.6-fold) that have been functionally associated with breast cancer development and treatment outcome, proving the power of single-cell sequencing. Moreover, it revealed potential convergent CNA (such as gain of IGF1R) occurring in parallel lineages of a tumor sample in 3 out of 20 TNBCs. Similar gains in discovery power were observed when applying Medatree on CNA profiles derived from single-cell RNA sequencing (scRNA-seq) data of 12 multiple myeloma patients. In summary, we developed a first-of-its-kind approach Medatree that enables large-scale retrospective CNA lineage tracing from single-cell copy number profiles obtained from single-cell DNA and RNA-sequencing data of patient samples. Wide application of Medatree is expected to accelerate the study of copy number evolution during tumor development and lead to discovery of novel targets for diagnoses and treatments. The source code of Medatree is available at https://github.com/KChen-lab/Medatree. Citation Format: Fang Wang, Qihan Wang, Jincheng Han, Vakul Mohanty, Shaoheng Liang, Darlan Conterno Minussi, Ruli Gao, Li Ding, Nicholas Navin, Ken Chen. Medatree enabling single cell copy number lineage tracing and functional discovery [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 4424.