Fig. 1

The CNAqc framework for clonal and subclonal CNAs. a CNAqc integrates mutations, allele-specific CNAs, and tumor purity \(\pi\) to quality control (QC) clonal and subclonal tumor aneuploidy. Given an input purity error tolerance \(\epsilon >0\), CNAqc computes a score \(\lambda \in \mathfrak{R}\) that leads to a pass (\(|\lambda |\le \epsilon\)) or fail status. The score magnitude \(\lambda\) can be used to (i) adjust \(\pi\) and re-parametrise the copy number caller or (ii) to select among alternative ploidy and purity estimates (e.g., a diploid versus a tetraploid solution). b CNAqc considers clonal simple CNAs, defined as 1:0 (LOH), 1:1 (heterozygous diploid), 2:0 (copy-neutral LOH), 2:1 (triploid), and 2:2 (tetraploid) states, the most prevalent segments in PCAWG, a large pan-cancer cohort of primary tumors. All other CNAs are called complex. c,d VAFs for mutations sitting on 1:1 and 2:1 segments, when \(\pi =1\). For 2:1, there are two peaks of clonal mutations (33% and 66% VAF). Multiplicities determine whether a mutation sits on the amplified segment. e CNAqc equations predict VAF peaks for any CNA, the distance between data and expected peaks is at the core of our approach. The color scale ranges from 0 to 1 (low to high VAF) for the expected peaks. For complex CNAs, multiplicities range from 1 to the copies of each allele; for subclonal CNAs, peaks depend also on clone size, and the evolutionary (linear/ branching) model of evolution. f QC for mutations in a 2:1 segment with ~ 90% sample purity. The vertical dashed lines are expected VAF peaks matched in the shaded area, obtained normalizing \(\epsilon\) for segment ploidy, mutation multiplicity, and purity. Black dots are peaks detected by CNAqc; if they fall within the bandwidths, the QC result is pass (green status bar). QC results are computed per peak, per type of segment, and at the sample level