Title: Generating simulated SNP array and sequencing data to assess genomic segmentation algorithms

Authors: Mark R. Zucker; Kevin R. Coombes

Addresses: Memorial Sloan Kettering Cancer Center, Department of Epidemiology and Biostatistics, New York City, New York, 10065, USA ' The Ohio State University Wexner Medical Center, Department of Biomedical Informatics, Columbus, OH 43210, USA

Abstract: We developed a tool, implemented in an R package called true and accurate clone generator (TACG), to simulate 'ground truth' and realistic SNP array and single nucleotide variant (SNV) data. We present TACG and use it to assess several different approaches to segmentation of copy number data from SNP arrays, with a particular interest in detecting copy number variations (CNVs) in cancer samples. We demonstrate that DNAcopy, an algorithm using circular binary segmentation, generally performs best, which is in agreement with previous research. We determine the conditions under which it and other methods break down. In particular, we assess how characteristics like clonal heterogeneity, presence of nested CNVs, and the type of aberration affect algorithm accuracy. The simulations we generated proved to be useful in determining not just the comparative overall accuracy of different algorithms, but also in determining how their efficacy is affected by the biological characteristics of samples from which the data was generated.

Keywords: SNP arrays; copy number variation; cancer; simulations; segmentation; algorithms; Hidden Markov Models; circular binary segmentation; genomics; whole exome sequencing.

DOI: 10.1504/IJCBDD.2020.113822

International Journal of Computational Biology and Drug Design, 2020 Vol.13 No.5/6, pp.438 - 453

Received: 16 Sep 2019
Accepted: 21 Apr 2020
Published online: 31 Mar 2021 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article