Abstract A015: Inferring phylogenetic trees based on copy number aberrations from methylation array

Abstract Background: The development of methylation array platforms has allowed large scale studies on cancer epigenomes. Apart from intended analyses on methylation patterns, it is also possible to infer copy number profiles by leveraging the signal from methylation array data. To disentangle complex mixtures of phylogenetically related clones, it is best to jointly analyze different samples (different regions, timepoints, etc.) from the same tumor. However, current methods for studying copy number aberrations (CNAs) from methylation arrays consider tumor samples independently. In this project, we introduce Multi-Sample EPIC Phylogenetic Reconstruction (MultiSEPhyR), aiming to investigate the copy number profiles and clonal compositions from multi-sample methylation array data. Material and methods: As part of the GCGR (Glioma Cellular Genetic Resource; gcgr.org.uk) project, we analyze 53 samples from 26 cases with multiple methylation array data of primary glioblastomas, tumor-derived cell lines and xenograft re-derived cell lines. Copy Number Aberrations (CNAs) called from the methylation array data are validated using WGS data from a subset of the cohort. By jointly clustering the segmented LogR ratios from related samples, we can define tumor clones based on clustered CNA events. We use a two-step approach, where we firstly identify clonal events to obtain sample purity and tumor ploidy and use these values to subsequently infer the copy number state and the clonality of the remaining events. Results and Discussions: On a simulated dataset, MultiSEPhyR could accurately estimate sample purity and identify both clonal and sub-clonal CNA events. We then apply our method on GCGR dataset and show similar extent of intra-tumor heterogeneity in primary glioblastomas and derived cell lines. We also investigate clonal dynamics in each case. In the cell lines, while the heterogeneity is well captured, some sub-clonal CNA events may exhibit significantly different clonality compared to the primary glioblastomas. For example, we observe a clonal expansion of the clones harboring a deletion in chromosome 6 in xenograft re-derived cell lines. Conclusion We present a systematic way to call CNAs and resolve clonal compositions from methylation array data, leading to a cost-efficient way to study tumor evolution. Citation Format: Chuling Ding, Javier Herrero. Inferring phylogenetic trees based on copy number aberrations from methylation array [abstract]. In: Proceedings of the AACR Special Conference on the Evolutionary Dynamics in Carcinogenesis and Response to Therapy; 2022 Mar 14-17. Philadelphia (PA): AACR; Cancer Res 2022;82(10 Suppl):Abstract nr A015.