SEQanswers

Go Back   SEQanswers > Bioinformatics > Bioinformatics



Similar Threads
Thread Thread Starter Forum Replies Last Post
Segmentation fault, cuffdiff labrat73 Bioinformatics 12 03-22-2018 11:00 AM
Segmentation error (core dumped) in velvetg Lakshmi_ISU Bioinformatics 10 08-30-2013 05:52 PM
soap segmentation fault scami Bioinformatics 6 04-17-2012 06:08 AM
SOAP:segmentation fault Mansequencer Bioinformatics 1 11-19-2010 10:47 AM
see segmentation fault in soapsnp guoweiguowei@gmail.com Bioinformatics 3 04-21-2010 09:15 PM

Reply
 
Thread Tools
Old 10-11-2013, 08:33 PM   #1
syintel87
Member
 
Location: Universe

Join Date: Dec 2012
Posts: 81
Default [velvetg] segmentation fault?

Hello,

- I am running "velvet".
- Velvet that I am using is the latest version.
- After running velvetg, I've got error message "segmentation fault".
- I think the memory is enough: I can use either 768 GB or 128 GB machine.

The commands that I used are the below.

Quote:
make 'CATEGORIES=4'
Quote:
srun -p bigmem ./velveth OUTPUT_1 31 -fastq -shortPaired 02_L7.fastq 02_L8.fastq -fastq -shortPaired2 05_L7.fastq 05_L8.fastq -fastq -shortPaired3 08_L7.fastq 08_L8.fastq -fastq -shortPaired4 11_L7.fastq 11_L8.fastq > output1.out 2>&1 &
556825727 sequences found
...
[107491.832312] Inputting sequence 556000000 / 556825727
[107637.486093] === Sequences loaded in 104746.688101 s
[107637.491597] Done inputting sequences
[107637.491604] Destroying splay table
[107650.200950] Splay table destroyed

Quote:
srun -p bigmem -c 16 ./velvetg OUTPUT_1 -read_trkg yes -amos_file yes -unused_reads yes -exp_cov 10 -ins_length 3000 -ins_length_sd 1000 -ins_length2 6000 -ins_length_sd 1000 -ins_length3 9000 -ins_length_sd 1000 -ins_length4 11000 > output1_g.out 2>&1 &
srun: error: node92: task 0: Segmentation fault



I've got this error message. Would you please give me a piece of advice to resolve my problem? Thank you in advance.
syintel87 is offline   Reply With Quote
Old 10-12-2013, 10:50 AM   #2
winsettz
Member
 
Location: US

Join Date: Sep 2012
Posts: 91
Default

Quote:
Originally Posted by syintel87 View Post
Code:
srun -p bigmem -c 16 ./velvetg OUTPUT_1 -read_trkg yes -amos_file yes -unused_reads yes -exp_cov 10 -ins_length 3000 -ins_length_sd 1000 -ins_length2 6000 -ins_length_sd 1000 -ins_length3 9000 -ins_length_sd 1000 -ins_length4 11000 > output1_g.out 2>&1 &
srun: error: node92: task 0: Segmentation fault



I've got this error message. Would you please give me a piece of advice to resolve my problem? Thank you in advance.
Depending on the size of your files, might compiling with BIGASSEMBLY be required? Not sure if that'll solve your problem. Have you tried running the assembly with the shorter paired end libraries alone, and iteratively testing with the addition of your 6kb and then 9kb libraries.

Not sure if your ins_length*_sd is formed correctly. It should be ins_length_sd, ins_length2_sd, ins_length3_sd, ins_length4_sd (which you omitted). I don't think this would cause your segfault though.
winsettz is offline   Reply With Quote
Old 10-12-2013, 01:41 PM   #3
AdrianP
Senior Member
 
Location: Ottawa

Join Date: Apr 2011
Posts: 130
Default

Torsten Seemann, who contributes to the velvet project has answered this question which was addressed to the velvet mailing list:

Quote:
(Soyeon - I have replied this to the Velvet Users mailing list. That is the most appropriate place to ask these questions.)

The most likely problem is 556825727 sequence found. That is half a billion reads. For k=31 that is a LOT OF NODES. You have probably ran out of memory.

I see you are using srun (SLURM) on a cluster. How do you know that the velvet you compiled (on the head node?) is the one your job is using?

You are using 4 categories. Do you have 4 separate DNA libraries that you sequenced? Or is your data just split into multiple files?
AdrianP is offline   Reply With Quote
Old 10-12-2013, 04:47 PM   #4
winsettz
Member
 
Location: US

Join Date: Sep 2012
Posts: 91
Default

Quote:
Originally Posted by AdrianP View Post
Torsten Seemann, who contributes to the velvet project has answered this question which was addressed to the velvet mailing list:
I would assume that having four different insert sizes would suggest four different libraries. I guess he didn't say how much memory his job is taking, "bigmem" is vague. Velvet's also pretty memory-hungry, other assemblers are more efficient in memory utilization. -create_binary on the later builds of velvet can bring down memory utilization somewhat, but choosing a larger K and loading less data will be helpful as well.

One could imagine a situation where a job requests so much memory that it sits in queue forever, never quite getting enough freemem to kick off unless it's the weekend.

Last edited by winsettz; 10-12-2013 at 05:10 PM.
winsettz is offline   Reply With Quote
Reply

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off




All times are GMT -8. The time now is 12:54 AM.


Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2020, vBulletin Solutions, Inc.
Single Sign On provided by vBSSO