Haplotype based copy number

Haplotype-based copy number analysis using TitanCNA for 10X Genomics data

Here is a schematic of the approach and modification of the allele-based copy number.

(a) Workflow for analysis of haplotype-based copy number. (i) Phased heterozygous SNP sites and haplotype block information were identified in the normal sample from output generated by LongRanger. (ii) The allele read counts of the two haplotypes were extracted for each SNP site from the tumor. (iii) Haplotype read counts were added across SNPs within 10 kb bins to generate haplotype coverage.
(b) Copy number profile in a representative sample on chr 6 is shown using molecule-based coverage (top, “molecule” refers to the unique barcode from linked-read information), SNP-based allelic fraction (middle), and haplotype-based fraction (bottom). The major haplotype fraction is shown (> 0.5), and the minor haplotype fraction is also shown for comparison (< 0.5). The haplotype phase blocks (from LongRanger) are shown with alternating colors. The haplotype fraction HFi at bin i is a continuous value modeled using a Gaussian distribution with mean and variance 2, where is the 3-component mixture representing tumor purity and subclonal heterogeneity (Ha et al., 2014). By contrast, allele fractions AFj at SNP j is modeled using a Binomial distribution for reference counts and depth Nj.