BGI-Qingdao / TGS-GapCloser

A gap-closing software tool that uses long reads to enhance genome assembly.
GNU General Public License v3.0
172 stars 12 forks source link

ERROR: line 645: 16242 Aborted (core dumped) #74

Closed Bank-tidy closed 9 months ago

Bank-tidy commented 9 months ago

Dear Developer,

Thank you for creating this practical software. I've encountered a memory issue while using it and would like to consult you. Here is my code:

tgsgapcloser \
    --scaff 100k_Chr7.fasta \
    --reads hifi.fasta \
    --output result_hifi \
    --minmap_arg '-ax map-hifi -t 10 -K 80M' \
    --thread 30 \
    --ne \
    >pipe.log 2>pipe.err

Initially, I input the complete genome sequence. The pipe.error file displays:

tgsgapcloser: line 645: 16242 Aborted (core dumped)$GapCloser --ont_reads_a $FINAL_READS --contig2ont_paf $OUT_PREFIX.fill.paf --min_match=$MIN_MATCH --min_idy=$MIN_IDY --prefix $OUT_PREFIX > $OUT_PREFIX.fill.log 2>&1

Following your suggestion of 5.5Gb xxx.paf records a 32Gb peak memory, I extracted the chromosome 7 genome with gap regions, resulting in a paf file of only 6.8G. My server has 512G of running memory, but I still encounter an error:

tgsgapcloser: line 645: 16242 Aborted (core dumped)$GapCloser --ont_reads_a $FINAL_READS --contig2ont_paf $OUT_PREFIX.fill.paf --min_match=$MIN_MATCH --min_idy=$MIN_IDY --prefix $OUT_PREFIX > $OUT_PREFIX.fill.log 2>&1 Should I reduce the -t parameter to 10 and increase the -K parameter to 80M, or do you have any other suggestions?

The additional noteworthy point is that in the hifi.fill.log, it shows: "tgsgapcloser: ../biocommon/align_common/align_result.cpp:134: void BGIQD::ALIGN_COMMON::ExtraInfo::InitFromStr(const string&): Assertion `items.size() == 3' failed."

Here is my log screenshot image image image

Best regards!

cchd0001 commented 9 months ago

Could you please provide the head 1000 line of tomato_hifi.fill.paf ?

Bank-tidy commented 9 months ago

@cchd0001 Of course, thank you,but this file is a bit big. The first 10 lines of the file have reached 70M. If it is convenient, I can send it to you by email? image image

cchd0001 commented 9 months ago

Oh, please do not use-a for minimap2 because it set the output format as SAM ( but we expect PAF ).

Please change

--minmap_arg '-ax map-hifi -t 10 -K 80M'

to

--minmap_arg '-x map-hifi -t 10 -K 80M'

Bank-tidy commented 9 months ago

@cchd0001 Oh, I understand, thank you for your answer!