An overview of mutational and copy number signatures in human cancer

Abstract

The genome of each cell in the human body is constantly under assault from a plethora of exogenous and endogenous processes that can damage DNA. If not successfully repaired, DNA damage generally becomes permanently imprinted in cells, and all their progenies, as somatic mutations. In most cases, the patterns of these somatic mutations contain the tell-tale signs of the mutagenic processes that have imprinted and are termed mutational signatures. Recent pan-cancer genomic analyses have elucidated the compendium of mutational signatures for all types of small mutational events, including (1) single base substitutions, (2) doublet base substitutions, and (3) small insertions/deletions. In contrast to small mutational events, where, in most cases, DNA damage is a prerequisite, aneuploidy, which refers to the abnormal number of chromosomes in a cell, usually develops from mistakes during DNA replication. Such mistakes include DNA replication stress, mitotic errors caused by faulty microtubule dynamics, or cohesion defects that contribute to chromosomal breakage and can lead to copy number (CN) alterations (CNAs) or even to structural rearrangements. These aberrations also leave behind genomic scars which can be inferred from sequencing as CN signatures and rearrangement signatures. The analyses of mutational signatures of small mutational events have been extensively reviewed, so we will not comprehensively re-examine them here. Rather, our focus will be on summarising the existing knowledge for mutational signatures of CNAs. As studying CN signatures is an emerging field, we briefly summarise the utility that mutational signatures of small mutational events have provided in basic science, cancer treatment, and cancer prevention, and we emphasise the future role that CN signatures may play in each of these fields. © 2022 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

Keywords: cancer; copy number signatures; genomics; mutational signatures.

Figure 1

Examples of CN profiles derived from karyotyping and DNA sequencing. Mock karyotype (left), logR and BAF tracks (middle), and CN profile (right) for (A) a diploid and (B) an aneuploid genome. Maternal (blue) and paternal (orange) chromosomes are displayed separately and phased throughout. CN, copy number. The aneuploid genome is whole‐genome doubled (WGD) and includes losses prior to (chr1 + 6 + 19) and following (chr3 + 12 + 14) WGD, as well as gains prior to (chr5 + 20) and following (chr8 + 18) WGD. Additionally, a chromothripsis event after WGD involving both chr10 and chr16 is included.

Figure 2

CN summary vectors for the two profiles displayed in Figure 1. (A) Diploid and (B) aneuploid, as per the summary methodology of Steele et al [24]. LOH status and total CN categories are displayed above the bar plots. Segment size categories are displayed below the bar charts.

Figure 3

CN signature methodology (mechanism‐agnostic). (A) Allele‐specific CN profiles (top panels; y‐axis = allele‐specific CN, orange = minor CN, blue = major CN) have their segments classified by loss of heterozygosity (LOH) status, total CN, and segment size (middle panels; see legend in figure) to generate CN feature counts (bottom panels). Example profiles are shown for a diploid genome, a diploid genome with historic chromosomal instability, a genome double version of the previous genome, and a genome with extensive LOH (in order left to right). (B) CN feature counts for a dataset are combined into matrix of CN summaries (left), which are decomposed using non‐negative matrix factorisation, or another appropriate method, into a matrix of signature definitions (right) and a matrix of attributions of each of those signatures in the samples of the given dataset (middle).

Figure 4

Post hoc mapping of CN signatures to the genome. (A) CN profile of a dedifferentiated liposarcoma (DDLPS) with characteristic amplification of chr12 including MDM2. Orange indicates minor CN; blue indicates major CN. The bottom panel indicates regions of the genome that have been mapped to a signature (in grey); regions of the genome unattributed to a signature are in white. (B) CN summary vector for the DDLPS in A. See legend of Figure 2 for the ordering of CN classes. (C) Signature definitions for the three signatures attributed to this sample. CN classes from B can be explained by each of the three signatures. (D) Heatmap of the likelihood of signatures (y‐axis) explaining the segments in a genomic bin of the window (blue), where here the bins are the chromosomes of the human genome (x‐axis). Global attributions of the three signatures to this sample are coloured in orange. Signature 8, a signature associated with chromothripsis amplification, is attributed to chromosomes 12, 1, and 2, due to the highly segmented patterns seen on those chromosomes. Once the chromosomal likelihoods are obtained, the maximum likelihood signature for each segment based on its CN class can be assigned, giving the assignments shown in the bottom panel of A.

Figure 5

Schematic diagram of hypothetical genome configurations for two chromosomes (left; blue = chromosome A, red = chromosome B) and their associated CN and rearrangement profiles (y‐axis = allele‐specific CN, x‐axis = genome, orange = minor CN, blue = major CN, green = rearrangement). Specific structural alterations to the genome may lead to only CN alterations (whole‐genome doubling; WGD), only rearrangements (translocation), or a combination of both (loss). Further, the consequences of the same structural alterations (WGD and a loss) may depend on the order in which those structural alterations occurred; a loss followed by WGD (Loss → WGD) will lead to loss of heterozygosity for the lost segment, whereas the reverse ordering (WGD → Loss) will retain heterozygosity for that segment, which may be important when assessing second hits to tumour suppressor genes.