ruanjue / wtdbg2

Redbean: A fuzzy Bruijn graph approach to long noisy reads assembly
GNU General Public License v3.0
513 stars 94 forks source link

Will wtpoa-cns automatically exit after the run is complete? #50

Closed YangJianshun closed 5 years ago

YangJianshun commented 5 years ago

I run the command “./wtpoa-cns -t 16 -i prefix.ctg.lay -fo prefix.ctg.lay.fa”, but after generating the prefix.ctg.lay.fa file for a long time, wtpoa-cns does not Automatic termination, is this normal?

ruanjue commented 5 years ago

Abnormal. What's about the log information? Could you copy the text here.

YangJianshun commented 5 years ago

Abnormal. What's about the log information? Could you copy the text here.

$ /public/home/yangjs/software/wtdbg2/wtpoa-cns -t 16 -i BAC_nanopore.ctg.lay.gz -fo BAC_nanopore.final.fa

-- total memory 32727680.0 kB -- available 30977672.0 kB -- 4 cores -- Starting program: /public/home/yangjs/software/wtdbg2/wtpoa-cns -t 16 -i BAC_nanopore.ctg.lay.gz -fo BAC_nanopore.final.fa -- pid 31901 -- date Thu Dec 6 13:59:36 2018

11 contigs 800 edges

No longer change and don't quit But I did get this file ‘’BAC_nanopore.final.fa’’

ruanjue commented 5 years ago

Thanks Jianshun!

Please try new commit https://github.com/ruanjue/wtdbg2/commit/31cf61b4dfb1c3fdae56a6d9a48fa77442944f3f

Best, Jue

YangJianshun commented 5 years ago

Thanks Jianshun!

Please try new commit 31cf61b

Best, Jue

Thank you very much

zrlewis commented 5 years ago

Hi,

Jumping on this closed issue here because I think that I'm seeing the same behavior.

wtpoa-cns appears to hang and not finish. The end of the log file lists that it has gone through all of the contigs and edges, but the program does not exit.

I'm working off of the most recent commit

git show --oneline -s
5009cff fixed a bug in option -K

My command was:

wtpoa-cns -t 20 -i [filename].ctg.lay.gz -o [filename].ctg.lay.fa


For reference, my wtdbg2 command was:

wtdbg2 -t 20 -i concatenated_reads.fasta \
   -fo output_wtdbg2 \
   -p 0 -k 15 -AS 2 -s 0.05 -L 8000 \
   -g 3.1g \
   --edge-min 2 --rescue-low-cov-edges --aln-noskip \
   --ttr-cutoff-depth 20 \
   -K 1000.0

Here is the tail of the wtdbg2 output:

[Thu Dec  6 16:46:58 2018] building contigs
[Thu Dec  6 16:46:59 2018] searched 120585 contigs
[Thu Dec  6 16:46:59 2018] Estimated: TOT 2910323968, CNT 50730, AVG 57369, MAX 1501696, N50 165376, L50 4860, N90 21760, L90 23422, Min 5120
[Thu Dec  6 16:54:46 2018] output 50730 contigs
[Thu Dec  6 16:54:52 2018] Program Done
** PROC_STAT(TOTAL) **: real 106937.848 sec, user 2052869.090 sec, sys 25841.980 sec, maxrss 252911340.0 kB, maxvsize 279355532.0 kB
---
ruanjue commented 5 years ago

Dear zrlewis,

I cannot locate the bug from reading the code, for I haven't find a test data reproduce this problem.

If anyone find a small test data, I will continue to work on it.

Best, Jue

zrlewis commented 5 years ago

@ruanjue Thanks for looking into it. I'm trying a couple new assemblies from scratch using the most recent commit and will let you know if I still encounter the problem.

zrlewis commented 5 years ago

@ruanjue wtpoa-cns finished this time after running both wtdbg2 and consensus from the most recent commit. I had previously been running a slightly older wtdbg2 version. It was labeled Version: 2.2 (20181111), but I don't have the commit. That could have been the problem.

ruanjue commented 5 years ago

I believe there is still similar problem in wtpoa-cns, but have no idea where to find it. Let's talk about it when it re-occuring.

Jue

shri1984 commented 3 years ago

Hi, I am using latest copy of wtdbg2 from github. I also see a similar issue. wtpoa-cns reaches very end and then don't exit. i dont know whether this issue is related to the parameters I used in wtdbg2 run. I never issues before when I ran wtdbg2 with default settings.

ruanjue commented 3 years ago

It looked like some contig took very long time or endless. Try to split the input fle into multiple ones, and run wtpoa-cns seperately.

shri1984 commented 3 years ago

Thanks. You are right, there is one contig look endless.
I managed to split the ctg.lay file into multiple files and manage to finish them independently. However, when I counted the number of contigs in ctg.lay and concatenated (from independent files) .raw.fa. I observed a difference of 1 contig, and I isolated that contig which is not showing up in the final raw. fa and run it wtpoa-cns. It does not look like finishing any time soon. It is running for 2 hours for that one contig. any input?

ruanjue commented 3 years ago

If the contig file is small, please send it to me ruanjue.big(AT)qq.com.