Haplotype-Resolved Assembly for Synthetic Long Reads Using a Trio-Binning Strategy

Abstract

The accuracy and completeness of genome haplotyping are crucial for characterizing the relationship between human disease susceptibility and genetic variations, especially for the heterozygous variations. However, most of current variations are unphased genotypes, and the construction of long-range haplotypes remains challenging. We introduced a de novo haplotype-resolved assembly tool, HAST that exports two haplotypes of a diploid species for synthetic long reads with trio binning.It generates parental distinguishing k-mer libraries, partitions reads from the offspring according to the unique markers, and individually assembles them to resolve the haplotyping problem. Based on the stLFR co-barcoding data of an Asian as well as his parental massive parallel sequencing data, we utilized HAST to recover both haplotypes with a scaffold N50 of >11 Mb and an assembly accuracy of 99.99995% (Q63). The complete and accurate employment of long-range haplotyping information provided sub-chromosome level phase blocks (N50 ~13 Mb) with 99.6% precision and 94.1% recall on average. We suggest that the accurate and efficient approach accomplishes the regeneration of the haplotype chromosomes with trio binning, thus promoting the determination of haplotype phase, the heterosis of crossbreeding, and the formation of autopolyploid and allopolyploid.Page 3 haplotyping at its source. However, it is technically impractical for most species.Another strategy refers to the whole-chromosome isolation, for instance, chromosome sorting 6 , which is not applicable to large sample size due to its time-consuming and labor-intensive procedure. Haplotyping information can also be straightforwardly obtained from the sequencing of fragmented or whole DNA molecules. This approach requires long-range information to reconstruct the haplotypes from the biological or physical markers or overlapping relations of the segmented homologous chromosomes, including synthetic long reads (SLR, such as MGI stLFR co-barcoding reads and 10X genomics linked reads) 7-10 , single-molecule long reads (Pacific Biosciences (PacBio) or Oxford Nanopore (ONT) long reads) 11, 12 , Hi-C 13, 14 and optical mapping (Bionano) 15 . In practice the long-range information is employed to recover diploid or polyploid genomes with read-mapping-based, or de novo assemblybased methods.Currently, to attribute the successive genotypes to an arbitrary haplotype, sequencing reads are generally mapped to the reference genomes and then the alignment information recovers the heterozygous genotypes at each polymorphic site 16,17 .Nevertheless, the read-mapping-based methods are hampered by the lack of perfect reference sequence and the inability to detect large structural variations. Long reads 18, 19 or SLR 7,8,20 can extend the phase blocks to ~1Mb, but it is still complex to span regions of low heterozygous density or gaps in the reference. Assembly-based methods usually generate longer phase blocks across homozygous regions of the

A database and comprehensive analysis of the algae genomes

“…In addition, trio-binning strategy which is used for haploid assembly, is also a feasible way to remove variable exogenous contamination through the batch effect of sequencing data, and obtain a relatively pure algae genome. HAST [40], a recently published software, may realize this vision if there is suitable data. Alternatively, we may open mind more widely, change our mind and turn the question of 'how to separate algae genome with contamination' to 'what species do these sequences belong to, no matter it is algae or bacteria', that means we should change from common genomic mind to meta genomic mind.…”

Symbiont-Screener: a reference-free filter to automatically separate host sequences and contaminants for long reads or co-barcoded reads by unsupervised clustering

“…Meanwhile, the pedigree may provide unique inherent information. Recent studies have illustrated that the trio-binning strategy with long reads or co-barcoded reads enables haplotype-aware genome assembly with high accuracy and completeness, which promotes studies of genomic diseases and evolutionary relationships 22–24 . However, there are few specific decontamination tools designed for error-prone long reads or sparse co-barcoded reads, and current local-alignment-based methods originally designed for short reads or pre-assembled contigs could be less efficient and more time-consuming.…”