Identifying and removing haplotypic duplication in primary genome assemblies

Abstract

Motivation Rapid development in long-read sequencing and scaffolding technologies is accelerating the production of reference-quality assemblies for large eukaryotic genomes. However, haplotype divergence in regions of high heterozygosity often results in assemblers creating two copies rather than one copy of a region, leading to breaks in contiguity and compromising downstream steps such as gene annotation. Several tools have been developed to resolve this problem. However, they either focus only on removing contained duplicate regions, also known as haplotigs, or fail to use all the relevant information and hence make errors. Results Here we present a novel tool, purge_dups, that uses sequence similarity and read depth to automatically identify and remove both haplotigs and heterozygous overlaps. In comparison with current tools, we demonstrate that purge_dups can reduce heterozygous duplication and increase assembly continuity while maintaining completeness of the primary assembly. Moreover, purge_dups is fully automatic and can easily be integrated into assembly pipelines. Availability and implementation The source code is written in C and is available at https://github.com/dfguan/purge_dups. Supplementary information Supplementary data are available at Bioinformatics online.

HiCanu: accurate assembly of segmental duplications, satellites, and allelic variants from high-fidelity long reads

“…they do not preserve phase across their entire length). Canu also does not assign contigs to haplotypes, and requires postprocessing with a tool such as Purge_dups (Guan et al 2020) to split the diploid assembly into primary and alternate alleles. While recent studies have successfully integrated HiFi data with additional long-range linkage information Porubsky et al 2019), we do not expect that significant improvements in phasing can be achieved by HiFi-only assemblies without an increase of HiFi read lengths.…”

“…Similar to many modern assemblers, when faced with a diploid genome, HiCanu outputs contigs as "pseudo-haplotypes" that preserve local allelic phasing but may switch between haplotypes across longer distances. A single set of contigs representing all resolved alleles is output regardless of ploidy, and additional processing with a tool such as Purge_dups (Guan et al 2020) is required to partition the contigs into primary and alternate allele sets. Figure 1.…”

“…The large fraction of duplicated BUSCO (Waterhouse et al 2018) genes also supported the hypothesis that the assemblies captured alleles from both haplotypes ( Supplementary Table 2). To facilitate like-for-like comparison of all assemblies, we ran Purge_dups (Guan et al 2020) to identify alleles and divide the assemblies into primary and alternate contig sets (Methods). Assembly statistics were then computed for both contig sets and the results summarized in Table 1.…”

Towards complete and error-free genome assemblies of all vertebrate species

“…VGP assemblies exceeding Q40 contained fewer frameshift errors, as predicted 76 , and therefore we recommend targeting a minimum QV of 40 whenever possible. Haplotype phasing and false duplications are the most underdeveloped measures, presumably because this is an under-appreciated area of need, but recent tools, developed here and elsewhere 27,48,51,53,77 , are helping to address this. The six broad quality categories in the 1st column are split into sub-metrics in the 2nd column.…”

Comparative Genomic Insights into the Longhorned Tick, Haemaphysalis Longicornis

“…The k-mer frequencies were estimated with 21-mers using khist.sh (one of the BBTools v38.49 modules, [17]). Redundant heterozygous regions were removed with an identity cutoff of 60 using two rounds of Purge_dups v1.0.0 [16]. Non-redundant assembly was polished with Illumina short reads using two rounds of NextPolish v1.1.0 [17].…”