search

BGI-Qingdao / TGS-GapCloser

A gap-closing software tool that uses long reads to enhance genome assembly.
GNU General Public License v3.0
179 stars 13 forks source link

Candidate aborts with an error #9

Closed felixgrewe closed 4 years ago

felixgrewe commented 4 years ago

I want to close gaps in SSPACE-Longread scaffolds of a wtbdg2 nanopore assembly (no Illumina polishing possible).

When I run TGS-GapCloser without error correction (--ne) it finished, but when I use error correction with racon it stoped at step 2.2:

pipe.err [...] TGS-GapCloser.sh: line 373: 113308 Aborted (core dumped) $Candidate --ont_reads_a $TGS_READS --contig2ont_paf $OUT_PREFIX.sub.paf --candidate_max 10 --candidate_shake_filter --candidate_merge < $TMP_INPUT_SCAFF_INFO > $OUT_PREFIX.ont.fasta 2> $OUT_PREFIX.cand.log

.cand.log [...] TGSGapCandidate INFO CST 2020/4/28 1:7:26 : LoadPAF finish. used wall clock : 217 seconds, cpu time : 217.104172 seconds TGSGapCandidate INFO CST 2020/4/28 1:7:26 : LoadScaffInfo start now ... TGSGapCandidate INFO CST 2020/4/28 1:7:31 : LoadScaffInfo finish. used wall clock : 5 seconds, cpu time : 5.176731 seconds TGSGapCandidate INFO CST 2020/4/28 1:7:31 : ParseAllGap start now ... terminate called after throwing an instance of 'std::out_of_range' what(): map::at

adonis316 commented 4 years ago

Hi, We have seen this error when the input TGS long reads are in fastq format before. Would you please provide us your whole command (TGS-GapCloser.sh. XXXXXXX)?

Are you trying to use wtdbg2 assembled contigs to close the gaps in SSPACE-Longread scaffolds? Two points I would like to note:

  1. The assembled contigs should be ~1X coverage of the genome, and the coverage is not sufficent for you to do the error correction byRacon.
  2. The input assembled contigs however, could be error corrected by wtdbg2 using ONT reads. You can use the corrected contigs to close gaps without any further correction.
felixgrewe commented 4 years ago

Thanks for you immediate help.

Yes, I have an wtdbg2 assembly of nanopore only reads, used then SSPACE-Longread for scaffolding, and tried to close gaps with TGS-GapCloser.

Regarding your suggestions: 1) I should have a mean coverage of 50x. 2) Thank you for this suggestion, I'll continue this route.

~/git/TGS-GapCloser/TGS-GapCloser.sh  \
        --scaff  [path-to]/scaffolds.fasta \
        --reads  [path-to]/[ont-reads].fastq \
        --output wtbdg2-SSPACE \
        --racon [path-to]/bin/racon \
        >pipe.log 2>pipe.err
adonis316 commented 4 years ago

The input long reads should be in FASTA format. Please convert your ONT reads from FASTQ to FASTA first. Let me know if it works for you.