mahulchak / quickmerge

A simple and fast metassembler and assembly gap filler designed for long molecule based assemblies.
GNU General Public License v3.0
200 stars 31 forks source link

Segmentation fault with larger heterozygous input #42

Closed ViriatoII closed 5 years ago

ViriatoII commented 5 years ago

Hi,

I'm having segmentation fault at the quickmerger stage after using the merge_wrapper. Nucmer and everything before worked fine and created normal outputs, including oneline.fa files.

-rw-r--r-- 1 guerrer QGGP    2456765 Jul 22 13:23 anchor_summary_out.txt
-rw-r--r-- 1 guerrer QGGP 1049205219 Jul 22 13:22 hybrid_oneline.fa
-rw-r--r-- 1 guerrer QGGP          0 Jul 22 13:23 merged_out.fasta
-rw-r--r-- 1 guerrer QGGP   91857760 Jul 22 13:22 out.delta
-rw-r--r-- 1 guerrer QGGP    9563554 Jul 22 13:22 out.rq.delta
-rw-r--r-- 1 guerrer QGGP    3982704 Jul 22 13:23 param_summary_out.txt
-rw-r--r-- 1 guerrer QGGP 1343593069 Jul 22 13:22 self_oneline.fa

The quickmerge error must have to do with my input self_oneline.fa, it's the only new thing.

Now, instead of rerrunning the whole wrapper, I'm just running this command:

quickmerge -d out.rq.delta -q hybrid_oneline.fa -r self_oneline.fa -hco 5.0 -c 1.5 -l 0 -ml 5000 -p out

Results:

   File                   size        First line(fasta header)
hybrid_oneline.fa      1049205219        >1
self_oneline.fa        845417800         >tig00035160                Success!

hybrid_oneline.fa      1049205219        >1
self_oneline.fa        1343593069        >tig00000003_pilon     Segmentation fault (core dumped)

The only difference is the size and origin of the self file. The successful one is a canu assembly after using purge haplotigs to eliminate haplotigs (alternative diploid contigs ). The unsuccessful one is a canu assembly but without using purge haplotigs (thus it is heterozygous/diploid).

Is my problem derived from the input sizes? Or maybe from heterozygosity?

mahulchak commented 5 years ago

You can check if it is due to file size by running the command with

/usr/bin/time -v

It'll show you the memory usage for the two different assembly sizes.

Generally I don't recommend use of heterozygous assemblies for assembly merging. Better to use the assembly without haplotigs.

On Mon, Jul 22, 2019, 17:40 ViriatoII notifications@github.com wrote:

Hi,

I'm having segmentation fault at the quickmerger stage after using the merge_wrapper. Nucmer and everything before worked fine and created normal outputs, including oneline.fa files.

-rw-r--r-- 1 guerrer QGGP 2456765 Jul 22 13:23 anchor_summary_out.txt -rw-r--r-- 1 guerrer QGGP 1049205219 Jul 22 13:22 hybrid_oneline.fa -rw-r--r-- 1 guerrer QGGP 0 Jul 22 13:23 merged_out.fasta -rw-r--r-- 1 guerrer QGGP 91857760 Jul 22 13:22 out.delta -rw-r--r-- 1 guerrer QGGP 9563554 Jul 22 13:22 out.rq.delta -rw-r--r-- 1 guerrer QGGP 3982704 Jul 22 13:23 param_summary_out.txt -rw-r--r-- 1 guerrer QGGP 1343593069 Jul 22 13:22 self_oneline.fa

The quickmerge error must have to do with my input self_oneline.fa, it's the only new thing.

Now, instead of rerrunning the whole wrapper, I'm just running this command:

quickmerge -d out.rq.delta -q hybrid_oneline.fa -r self_oneline.fa -hco 5.0 -c 1.5 -l 0 -ml 5000 -p out

Results:

File size First line(fasta header) hybrid_oneline.fa 1049205219 >1 self_oneline.fa 845417800 >tig00035160 Success!

hybrid_oneline.fa 1049205219 >1 self_oneline.fa 1343593069 >tig00000003_pilon Segmentation fault (core dumped)

The only difference is the size and origin of the self file. The successful one is a canu assembly after using purge haplotigs to eliminate haplotigs (but without running pillon). The unsuccessful one is a piloned canu assembly but without using purge haplotigs (thus it is heterozygous).

Is my problem derived from the input sizes? Or maybe from heterozygosity?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/mahulchak/quickmerge/issues/42?email_source=notifications&email_token=ABZQH2BYBZNVY4OFW52Q653QAWPOJA5CNFSM4IFX54M2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HATZKPA, or mute the thread https://github.com/notifications/unsubscribe-auth/ABZQH2H4FA3Y2ELYJJEC6PDQAWPOJANCNFSM4IFX54MQ .