PhyloWGS: reconstructing subclonal composition and evolution from whole-genome sequencing of tumors.

Deshwar AG, Vembu S, Yung CK, Jang GH, Stein L, Morris Q. · 2015

Read PhyloWGS: reconstructing subclonal composition and evolution from whole-genome sequencing of tumors. by Deshwar AG, Vembu S, Yung CK, Jang GH, Stein L, Morris Q. (2015) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc4359439.txt.

Introduction

Amit Deshwar, Shankar Vembu, Christina Yung, Gun Ho Jang, Lincoln Stein, and Quaid Morris presented a software tool called PhyloWGS in Genome Biology in 2015. Tumors are mixes of cell groups, each with its own set of mutations, grown from a single ancestor. Reading a tumor's whole genome gives tangled data from all groups at once. The tool aims to unmix that data and rebuild each group's makeup and family tree.

PhyloWGS uses how often point mutations appear, plus how often larger structural changes appear, across one or more tumor samples. It adds a careful fix for spots where copy number changes would otherwise skew the mutation rates. Tests showed that fix greatly improves rebuilding compared with older methods. The software is free and open source, so other labs can run and adapt it.

Catalog record

CREATORDeshwar AG, Vembu S, Yung CK, Jang GH, Stein L, Morris Q.
YEAR2015
COLLECTIONPAPERS
FORMATTXT
FILE SIZE72 KB
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC4359439

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

PhyloWGS: Reconstructing subclonal composition and evolution from whole-genome sequencing of tumors By Amit G Deshwar, Shankar Vembu, Christina K Yung, Gun Ho Jang, Lincoln Stein, Quaid Morris Published in: Genome Biology DOI: 10.1186/s13059-015-0602-8 ABSTRACT Abstract Tumors often contain multiple subpopulations of cancerous cells defined by distinct somatic mutations. We describe a new method, PhyloWGS, which can be applied to whole-genome sequencing data from one or more tumor samples to reconstruct complete genotypes of these subpopulations based on variant allele frequencies (VAFs) of point mutations and population frequencies of structural variations. We introduce a principled phylogenic correction for VAFs in loci affected by copy number alterations and we show that this correction greatly improves subclonal reconstruction compared to existing methods. PhyloWGS is free, open-source software, available at https://github.com/morrislab/phylowgs. Electronic supplementary material The online version of this article (doi:10.1186/s13059-015-0602-8) contains supplementary material, which is available to authorized users. BACKGROUND Tumors contain multiple, genetically diverse subclonal populations of cells that have evolved from a single progenitor population through successive waves of expansion and selection [1-3]. Reconstructing their evolutionary histories can help identify characteristic driver mutations associated with cancer development and progression [4,5], and can provide insight into how tumors might respond to treatment [6,7]. In some cases, it is possible to genotype the subpopulations present in a tumor, while reconstructing its history, using the population frequencies of mutations that distinguish these subclonal populations [2,8-21]. Increasingly, tumors are being characterized using whole-genome sequencing (WGS) of bulk tumor samples [22] and few automated methods exist to perform this reconstruction on the basis of these data reliably. Subclonal reconstruction algorithms attempt to infer the population structure of heterogeneous tumors based on the measured variant allelic frequency (VAF) of their somatic mutations. Some methods perform this reconstruction based solely on single nucleotide variants or small indels (collectively known as simple somatic mutations or SSMs) [16-19,21,23]. Others use changes in read coverage to identify genomic regions with an average ploidy that differs from normal, which they explain using inferred copy number variations (CNVs) that affect some of the cells in the sample [15,20,24,25]. The low read depth of current WGS complicates subclonal reconstruction. Until recently, subclonal populations (i.e., subpopulations) were defined based on accurate estimates of the proportion of cells with each mutation (i.e., their population frequency), which, for individual SSMs, are only available through targeted resequencing where the read depths are orders of magnitude higher than typical WGS depths [17,18,23]. However, preliminary evidence suggests that the much larger number of mutations detected by WGS can compensate for their decreased read depth [26]. In contrast, CNVs affect large, multi-kilobase-sized or megabase-sized regions of the genome, which allow the average copy number of these regions to be accurately estimated with WGS. Unfortunately, CNV-based subclonal reconstruction is more difficult than SSM-based reconstruction because of the need to estimate simultaneously population frequency and new copy number for each CNV. Most CNV-based methods only attempt to infer the copy number status of the clonal cancerous population [24,25] that contains the mutations shared by all of the cancerous cells. The few CNV-based methods [15,20] that attempt to resolve more than one cancerous subpopulation are practically limited to a small number (often two) of subpopulations. In contrast, SSM-based methods applied to targeted resequencing data can reliably resolve many more cancerous subpopulations [16-18,23]. However, it remains unclear what the limits of WGS-based automated subclonal reconstruction are. Another open question is how to combine CNVs and SSMs when doing reconstruction. CNVs overlapping SSMs can interfere with SSM-based reconstruction because they complicate the relationship between VAF and population frequency. Although some methods attempt to model the impact of CNVs on the allele frequency of overlapping SSMs [17-19,27], these methods have significant restrictions. For example, several of these methods [17,18] make the unrealistic assumption that every cell either contains the structural variation and the mutation or neither. Also, no method places structural variations in a phylogenetic tree, which is important for studying the evolution of cancerous genomes. We describe PhyloWGS, the first method designed for complete subclonal phylogenic reconstruction of both CNVs and SSMs from WGS of bulk tumor samples. Unlike all previous methods, PhyloWGS appropriately corrects SSM population frequencies in regions overlapping CNVs and is fast enough to perform reconstruction of at least five cancerous subpopulations based on thousands of mutations. We present results on subclonal reconstruction problems that cannot be correctly reconstructed using previous methods. We also probe the relationship between WGS read depth and the number of subpopulations that PhyloWGS can recover. Finally, we demonstrate that even in the absence of reliable CNV estimates, it is still feasible to perform automated subclonal composition reconstruction based on SSM frequency data at typical WGS read depths (30 to 50 ×), even for highly rearranged genomes where less than 2% of the SSMs lie in regions of normal copy number. Open-source, free software implementing PhyloWGS is available under the GNU General Public License v3 [28]. Previous work Figure 1 provides an overview of an evolving tumor, the measurement of somatic VAFs and the resulting subclonal reconstruction process. Panel (i) of this figure shows a visualization of the evolution of a tumor over time as non-cancerous cells (subpopulation A, shown in grey) are replaced by, at first, one clonal cancerous population (subpopulation B, shown in green), which then further develops into multiple cancerous subpopulations (C and D, shown in blue and yellow, respectively). Tumor cells define new subpopulations by acquiring new oncogenic mutations that allow their descendants to expand relative to the other tumor subpopulations. Each circle in panel (i) refers to a subpopulation. We associate subpopulations with the set of shared somatic mutations that distinguish it from its parent subpopulation (or, in the case of A, from the germ line (or reference) genome); this mutation set is indicated by the corresponding lower case letter (e.g. mutation set b first appears in subpopulation B). However, each subpopulation also inherits all of its parent’s mutations; the subclonal lineage of a mutation is the set of all subpopulations that contain it (e.g., the subclonal lineage of a is A, B, C and D). [Figure: The development of intratumor heterogeneity and subclonal reconstruction. Tumor composition over time (i), the resulting distribution of variant allele frequencies (VAFs) (ii), the result of successful inference of the VAF clusters (iii), and the desired output of subclonal inference (iiii). SSM, simple somatic mutation; VAF, variant allelic frequency.] In general, the subpopulation-defining mutation sets include more than one mutation. Cancerous cells often have increased mutation rates, and even non-cancerous cells accumulate somatic mutations at a rate of 1.1 per cell division [29]. As such, subpopulations are defined not only by the small number of oncogenic ‘driver’ mutations that support rapid expansion but also by a larger number of ‘passenger’ mutations acquired before the driver mutation(s). The selective sweeps that cause subpopulation expansion increase the population frequency of both driver and passenger SSMs, driving them to having indistinguishable population frequencies [30,31]. However, sampling and technical noise in sequencing means that the observed VAFs are distributed around the true value for a subpopulation. Panel (ii) shows an example histogram of measured VAFs for SSMs found in a heterogeneous tumor sample. Subclonal reconstruction algorithms define mutation sets, and their associated

Read the complete text