Skip to content

Collapse

Collapsed contigs are commonly observed in polyploid hybrids due to the presence of highly similar homologous regions introduced through hybridization. These regions pose significant challenges for direct de novo assembly using current computational approaches (e.g. hifiasm). To address this limitation, we developed a strategy comprising two steps:
(1) Collapsed contigs detection: Identification of candidate contigs (copy number ≥2) through integrated analysis of sequencing depth profiles from HiFi, ONT, and Pore-C data.
(2) Collapsed contigs rescuing: Duplicating and putting collapsed contigs into correctly groups.

Warning

This methodology’s particular efficacy in resolving localized collapsed regions can not resolve the collapsed regions that nearly whole chromosomes.

Collapsed contigs detection#

  • Custom mapping 
    minimap2 -cx map-hifi -I 16g --secondary=no -t 40 draft.asm.fasta hifi.fastq.gz | pigz -p 10 -c > hifi.align.paf.gz

cphasing-rs paf2depth hifi.align.paf.gz -w 5000 -s 1000 -o hifi.align.depth
cphasing collapse from-depth hifi.align.depth
  • Directly use the hitig results
    output.collapsed.contigs.list
  • Custom mapping 
    minimap2 -cx map-ont -I 16g -t 40 --secondary=no draft.asm.fasta ont.fastq.gz | pigz -p 10 -c > ont.align.paf.gz

cphasing-rs paf2depth hifi.align.paf.gz -w 5000 -s 1000 -o ont.align.depth
cphasing collapse from-depth ont.align.depth
  • Directly use the hitig results
    output.collapsed.contigs.list

The porec.align.paf.gz generated from cphasing mapper. 

cphasing-rs paf2depth porec.align.paf.gz -w 5000 -s 1000 -o porec.align.depth
cphasing collapse from-depth porec.align.depth

The GFA file generated from hifiasm record the read number, which can be used to identify the collapsed contigs. 

cphasing collapse from-gfa hifi.p_utg.noseq.gfa.gz

Collapsed contigs rescuing#

    cphasing collapse rescue 3.hyperpartition/porec.align.porec.q1.e5m.hg draft.asm.contigsizes 3.hyperpartition/output.clusters.txt contigs.collapsed.contig.list -n 4 -at 3.hyperpartition/draft.asm.allele.table

Note

Currently, we output the rescue result into a collapsed.rescue.clusters.txt file, which the user executes the next step of 4.scaffolding manually. And the input file can directly use the previous.

After scaffolding#

  • Generate new contig-level fasta and agp, which renamed the duplicated contigs (e.g. utg000001l -> utg000001_d2) The name of collapsed contigs in agp is need to rename for subsequence processes. 
    cphasing agp2fasta groups.rescued.agp draft.asm.fasta --contigs > draft.dup.fasta
cphasing collapse agp-dup groups.rescued.agp > groups.dup.agp
  • Generate a new pairs.gz or pairs.pqs file 
    cphasing collapse pairs-dup sample.pairs.pqs collapsed.rescue.contigs.list -o sample.dup.pairs.pqs
  • Curation by juicebox 
    cphasing utils agp2assembly groups.dup.agp > groups.dup.assembly
cphasing pairs2mnd sample.dup.pairs.pqs -o sample.dup.mnd.txt
  • Rename 
    cphasing rename -r ref.fa -f draft.dup.fasta -a groups.dup.agp -t 40
  • Plot 
    cphasing pairs2cool sample.dup.pairs.pqs sample.dup.pairs.pqs/_contigsizes sample.10k.cool 
cphasing plot -a groups.dup.agp -m sample.10k.cool -o sample.500k.wg.png