Fig. 1.

scDNA-seq artifacts. scDNA-seq analysis is complicated by a combination of imbalanced amplification of homologous alleles (allelic imbalance) and early artifacts that affect a large fraction of initial DNA. a MDA is a non-linear amplification process. DNA strands displaced by replication are immediately available for repeated rounds of replication, which can lead to imbalanced amplification between homologous alleles. Single cell sequencing depth is the sum of allele-specific sequencing depths, represented by a stacked depth plot. Pink: sequencing depth of maternal allele; blue: paternal allele. b Routine DNA damage during extraction protocols can disproportionately impact single cell DNA. Bulk DNA damage is mostly washed out since spontaneous errors are unlikely to recur independently on multiple molecules. However, damage to a single cell’s genome can affect a large amount (25%; single-stranded error) of the initial template. In an idealized MDA process, MDA replicates all four initial strands (two molecules) of DNA to produce eight strands (four molecules). A random polymerase misincorporation error would affect 1 out of 8 DNA strands, affecting 12.5% of the DNA. c Allelic imbalance affects the VAF of both true mutations and artifacts. Over amplification of an allele harboring early single-stranded damage (red) can inflate the artifact to a mutation-like VAF while a true mutation (green) can be reduced to low VAF. d Without allelic imbalance, heterozygous SNV VAFs would be tightly distributed around 50% due to random sampling effects. However, allelic imbalance causes VAFs in scDNA-seq to be significantly over-dispersed symmetrically around 50%