Fig. 1: HapTree-X framework compared to read-based phasing. | Nature Communications

Fig. 1: HapTree-X framework compared to read-based phasing.

From: Improved haplotype inference by exploiting long-range linking and allelic imbalance in RNA-seq datasets

Fig. 1

Traditional whole-genome sequencing (WGS) based phasing methods (top panel) depend on sequence contiguity and thus require a pair of SNPs (in red) to be connected through a common read that overlaps both in order to be phased. RNA-seq reads provide longer distance phasing capability due to long introns in the genome that are spliced-out in the sequenced transcript fragments (middle panel), yet SNPs that are far apart within the transcript due to long homozygous exonic regions are still difficult to phase using RNA-seq reads. Our HapTree-X framework (lower panel) overcomes this limitation by integrating RNA-seq reads and differential allele-specific expression (DASE) available from the RNA-seq data into a single probabilistic framework for haplotype phasing. For genes that display differential haplotypic expression (DHE), the majority of alleles can be phased together to obtain a single haplotype block for the entire gene. Depending on the DHE and depth-coverage, DASE-based phasing performs accurate haplotype reconstruction, without requiring paired-end or long reads, maintaining or improving on accuracy independent of gene/exon lengths as long as differential haplotypic expression is consistent across the loci being phased.

Back to article page