Characterizing the allele- and haplotype-specific copy number landscape of cancer genomes at single-cell resolution with CHISEL

Abstract

Single-cell barcoding technologies have recently been used to perform whole-genome sequencing of thousands of individual cells in parallel. These technologies provide the opportunity to characterize genomic heterogeneity at single-cell resolution, but their extremely low sequencing coverage (ă0.05X per cell) has thus far restricted their use to identification of the total copy number of large multi-megabase segments in individual cells. However, total copy numbers do not distinguish between the two homologous chromosomes in humans, and thus provide a limited view of tumor heterogeneity and evolution missing important events such as copy-neutral loss-of-heterozygosity (LOH).We introduce CHISEL, the first method to infer allele-and haplotype-specific copy numbers in single cells and subpopulations of cells by aggregating sparse signal across thousands of individual cells. We applied CHISEL to 10 single-cell sequencing datasets from 2 breast cancer patients, each dataset containing «2 000 cells. We identified extensive allele-specific copy-number aberrations (CNAs) in these samples including copy-neutral LOH, whole-genome duplications (WGDs), and mirrored-subclonal CNAs in subpopulations of cells. These allele-specific CNAs alter the copy number of genomic regions containing well-known breast cancer genes including TP53, BRCA2, and PTEN but are invisible to total copy number analysis. We utilized CHISEL's allele-and haplotype-specific copy numbers to derive a more refined reconstruction of tumor evolution: timing allele-specific CNAs before and after WGDs, identifying low-frequency subclones distinguished by unique CNAs, and uncovering evidence of convergent evolution. This reconstruction is supported by orthogonal analysis of somatic single-nucleotide variants (SNVs) obtained by pooling barcoded reads across clones defined by CHISEL.1 Single-cell DNA sequencing is a promising technology to quantify tumor heterogeneity and evolution with unprecedented resolution, enabling the identification of rare subpopulations of cells with distinct mutations and the inference of the evolutionary dynamics of cancer 1-4 . Recently, single-cell barcoding technologies, including Chromium Single Cell CNV Solution from 10X Genomics 5, 6 and direct library preparation 7, 8 , have been used to perform low-coverage whole-genome sequencing of thousands of individual cells in parallel, overcoming the limited number of cells and the amplification/coverage biases of previous techniques 4 . Due to technical and financial limitations, these technologies have extremely low sequencing coverage (ă0.05X per cell) which has thus far limited their application to the detection of large («3-5Mb) copy-number aberrations (CNAs) in individual cells. CNAs alter the number of copies of genomic regions, are frequent somatic mutations that drive cancer development 9-12 , play a crucial role in cancer treatment and prognosis 13,14 , and provide important markers for reconstruction of cancer evolution [15][16][17][18][19] .Since the human genome is diploid, ea...

Methods for copy number aberration detection from single-cell DNA-sequencing data

“…The same applies to ADO, as most of the existing CNA calling methods do not distinguish major and minor alleles (for a given SNV, the most common and less frequent alleles are called major and minor alleles, respectively). To the best of our knowledge, the only CNA detection method that is allele-specific is CHISEL [56], which we review below. Generally speaking, since CNAs occur at individual alleles, allelespecific CNA detectors have an advantage of inferring the evolutionary history of the cells.…”

“…Generally speaking, since CNAs occur at individual alleles, allelespecific CNA detectors have an advantage of inferring the evolutionary history of the cells. However, resolving allele-specific CNAs requires phasing, which is to distinguish CNAs based on the alleles at which they occur and group them by the alleles, a step that, for example, is accomplished by using SNVs in CHISEL [56]. However, scDNAseq's low coverage makes it very challenging to do phasing based on SNVs.…”

“…CHISEL [56] is the first allele-specific and haplotype-specific CNA detection tool for scD-NAseq. It utilizes a reference-based algorithm, Eagle2, to phase blocks in each bin, with the help of either a matched normal sample or a pseudo-normal sample derived from diploid cells.…”

“…CHISEL then groups cells into clones by hierarchical clustering. The application of CHISEL to 10 single-cell data sets in 2 breast cancer patients [56] demonstrated that, thanks in part to considering the BAF signal, it can (1) detect CNAs on some well-known breast cancer genes and (2) reconstruct a more refined tumor evolutionary scenario that has been orthogonally validated by SNVs. CHISEL is innovative in incorporating allelic signals to infer the clusters for bins and cells in addition to read depth.…”

STARCH: Copy number and clone inference from spatial transcriptomics data

“…First, the copy number model used in STARCH is fairly coarse with only three copy number states (Deletion, Neutral, Amplification); a larger number of copy number states could provide more accurate CNPs and clone assignments. One interesting direction is to extend STARCH to infer allele-specific copy number, as has been done for bulk DNA-seq [Ha et al, 2014, Zaccaria and Raphael, 2018, Nik-Zainal et al, 2012, single-cell DNA-seq [Garvin et al, 2015, Zaccaria andRaphael, 2019], and single-cell RNA-seq [Fan et al, 2018]. Allele-specific copy number provides a more accurate representation of the clonal structure of tumors, is essential for identifying events such as copy-neutral loss of heterozygosity, and could be helpful in quantifying allele-specific gene expression.…”