HaploMerger: Reconstructing allelic relationships for polymorphic diploid genome assemblies

Abstract

Whole-genome shotgun assembly has been a long-standing issue for highly polymorphic genomes, and the advent of nextgeneration sequencing technologies has made the issue more challenging than ever. Here we present an automated pipeline, HaploMerger, for reconstructing allelic relationships in a diploid assembly. HaploMerger combines a LASTZChainNet alignment approach with a novel graph-based structure, which helps to untangle allelic relationships between two haplotypes and guides the subsequent creation of reference haploid assemblies. The pipeline provides flexible parameters and schemes to improve the contiguity, continuity, and completeness of the reference assemblies. We show that HaploMerger produces efficient and accurate results in simulations and has advantages over manual curation when applied to real polymorphic assemblies (e.g., 4%-5% heterozygosity). We also used HaploMerger to analyze the diploid assembly of a single Chinese amphioxus (Branchiostoma belcheri ) and compared the resulting haploid assemblies with EST sequences, which revealed that the two haplotypes are not only divergent but also highly complementary to each other. Taken together, we have demonstrated that HaploMerger is an effective tool for analyzing and exploiting polymorphic genome assemblies.

Nemertean and phoronid genomes reveal lophotrochozoan evolution and the origin of bilaterian heads

“…Gaps in the scaffolds were filled with GapCloser (v1.12-r6). Redundant allele scaffolds were removed with HaploMerger (2_20151106) 60 . Genome assembly quality was assessed with N(X) graphs using QUAST (v3.1) ( Supplementary Fig.…”

dipSPAdes: Assembler for Highly Polymorphic Diploid Genomes

“…Currently, HaploMerger (Huang et al, 2012) is the only available (and practical) tool for assembling HP genomes from short reads. Thus, we benchmarked the performance of dipSPAdes and HaploMerger in generating consensus contigs (HaploMerger build 20120810 was used).…”

“…The second approach (Huang et al, 2012;Vinson et al, 2005) is to generate haplocontigs using a conventional assembly algorithm and to further reconstruct allelic relationships between haplotypes based on pairwise contig alignments (recently, an advanced algorithm for generating haplocontigs based on the overlap graph approach was proposed in Donmez and Brudno, 2011). In reality, such approaches generate a mixture of haplocontigs and consensus contigs since the degree of polymorphism varies along the HP genomes, and it is often difficult to generate haplocontigs in genomic regions with low polymorphism rates.…”

“…One can see that in this case, HaploMerger hardly improved on the original fragmented assembly by SPAdes (slightly increasing N50 from 5392 to 5876). This limitation [acknowledged in (Huang et al, 2012)] often prevents application of HaploMerger for NGS data where N50 is typically small. To further benchmark various assemblers, we mixed two libraries of S. commune and computed contigs from the graph that was constructed from the resulting mixed library.…”

Draft Genome Sequence of the Aquatic Fungus Hymenoscyphus tetracladius Strain NNIBRFG329

“…The quality of the reads was checked using FastQC v0.10.1 (http://www.bioinformatics.babraham.ac.uk/projects/fastqc) and was preprocessed using Trimmomatic v0.30 for paired-end sequencing (6) and NextClip v1.3 for mate pair sequencing (7). The reads were assembled using Canu v1.6 for PacBio reads (8) and SOAPdenovo v2.04 for Illumina reads (9) and were merged using HaploMerger2 v.20151124 (10). The draft genome of H.…”