Figure 2

CLONET key elements. (A) For each genomic segment, the AF distribution of informative SNPs follows a bimodal distribution that depends on the proportion of reference bases in the active allele (N_ref) and on the percentage of neutral reads (β) (left panel; 60X coverage data used For high values of β the distribution becomes unimodal as a function of the coverage (right panel; N_ref = 0.5). (B) For a genomic segment, the relationship between Adm.local and β is a hyperbola whose parameter values are governed by the segment copy number (CN). (C) Genomic segment representation in the β versus Log R space depends on the combination of the copy number of the parental alleles. Aneuploidy causes a shift while Adm.global results in shrinkage of the allowed subspaces. Violet dots indicate genomic segments with equal number of parental alleles (that is, 1/1, 2/2, 3/3,…). (D) The sketch illustrates the alternative allele proportion (AP) of a mono-allelic REARR defined by the genomic coordinates bp₁ and bp₂ The aberrant allele is represented by two types of reads: reads with abnormal insert size (gray) and reads that span both sides of a breakpoint (orange for bp₁ and red for bp₂). The non-aberrant allele exhibits two matching classes: paired-end crossing reads where one end maps entirely to one side and the other entirely to the other side correspond to aberrant gray reads. Single end reads that span the breakpoint (overlapping) are a complementary match to aberrant orange and red reads. The clonality is proportional to the proportion of aberrant reads around the breakpoint junction. Ad hoc quality filters are applied for removing sequencing artifacts around REARRs (right inset). First, the non-aberrant allele coverage is the minimum between the coverage in bp₁ and bp₂. Then, a REARR is adequate for analysis if the AP is statistically indistinguishable if we consider overlapping or crossing reads (proportion test with limit value 0.1).