Computational methods for chromosome-scale haplotype reconstruction

Abstract

High-quality chromosome-scale haplotype sequences of diploid genomes, polyploid genomes, and metagenomes provide important insights into genetic variation associated with disease and biodiversity. However, whole-genome short read sequencing does not yield haplotype information spanning whole chromosomes directly. Computational assembly of shorter haplotype fragments is required for haplotype reconstruction, which can be challenging owing to limited fragment lengths and high haplotype and repeat variability across genomes. Recent advancements in long-read and chromosome-scale sequencing technologies, alongside computational innovations, are improving the reconstruction of haplotypes at the level of whole chromosomes. Here, we review recent and discuss methodological progress and perspectives in these areas.

Features and applications of haplotypes in crop breeding

“…Haplotype is a combination of alleles for different polymorphisms (such as SNPs, insertions/deletions and other markers or variants) present on the same chromosome, which are inherited together with minimum chance of contemporary recombination 71 , 72 . Any individual has two haplotypes for a given stretch of chromosomal DNA; while at the population level, many haplotypes can be found for the same stretch 73 .…”

“…Fixing of higher error rate (~10%) in the long-read sequencing technologies compared to short-read sequencing methods (NGS methods) needs specific bioinformatics-mediated adjustments 92 . In this regard, many different phasing methods enabling haplotype construction/reconstruction from long-read sequencing data have been recently developed, such as reference-based phasing (molecular haplotyping, single-cell phasing, and polyploid phasing), de novo genome assembly (such as diploid and polyploid assembly) and strain-resolved metagenome assembly (de novo re-assembly, single nucleotide variant-based assembly, read and contig binning) 72 . Combination of these haplotype analysis methods with various computational tools such as WhatsHap, HapCut2, HapTree, WhatsHap- polyphase, Falcon phase, Hifiasm, SDip, POLYTE, DESMAN, MetaMaps, and ProxiMeta, has greatly enhanced the efficiency and precision in the identification of do novo and rare variants from the long-read sequencing data 72 .…”

“…In this regard, many different phasing methods enabling haplotype construction/reconstruction from long-read sequencing data have been recently developed, such as reference-based phasing (molecular haplotyping, single-cell phasing, and polyploid phasing), de novo genome assembly (such as diploid and polyploid assembly) and strain-resolved metagenome assembly (de novo re-assembly, single nucleotide variant-based assembly, read and contig binning) 72 . Combination of these haplotype analysis methods with various computational tools such as WhatsHap, HapCut2, HapTree, WhatsHap- polyphase, Falcon phase, Hifiasm, SDip, POLYTE, DESMAN, MetaMaps, and ProxiMeta, has greatly enhanced the efficiency and precision in the identification of do novo and rare variants from the long-read sequencing data 72 . Therefore, integrating the various phasing and bioinformatics tools with the long-read sequencing technologies has allowed us to fully exploit the potential of these sequencing approaches in haplotype construction 91 .…”

“…This evidenced that successful haplotype reconstruction in the polyploid species has a huge impact on breeding these crops in the future 96 . Recent research demonstrates the enormous potential of the TGS in resolving the accuracy issues in the haplotype identification, thereby increasing the scope of haplotypes for genetic studies in both animals and plants 72 . Hence, the TGS platforms offer promising alternative to obtain haplotype-related information from the genomes, and future affordability of these sequencing platforms will have a profound impact on plant research and breeding.…”

Filling gaps of genome scaffolds via probabilistic searching optical maps against assembly graph

“…The longer reads by third generation sequencing [ 3 , 4 ], and long-distance linking information by pair-end, mate-pair, or mapping technologies, will definitely help genome assembly methods to resolve the ambiguities incurred by repeats [ 5 ]. The study [ 6 ] provides an elaborated review on the methodological progresses and perspectives in the integration of short-range and long-range information for improving assembly contiguity.…”