Warnings In Bowtie Mapping
1
1
Entering edit mode
11.1 years ago
bashwin.u ▴ 40

Hello,

I am trying to use bowtie on small synthetic data for short read mapping. My command is ./bowtie -p 8 -t -S hg19 -1 synthetic_sample1.fq -2 synthetic_sample2.fq > bowtie.sam. The alignment stats is really bad.

Seeded quality full-index search: 00:03:12
# reads processed: 1000003
# reads with at least one reported alignment: 129 (0.01%)
# reads that failed to align: 999874 (99.99%)
Reported 129 paired-end alignments to 1 output stream(s)
Time searching: 00:03:17
Overall time: 00:03:17
[samopen] no @SQ lines in the header.
[sam_read1] missing header? Abort!

My error log says the following almost with every read I think. So the error log is a long list of similar errors

Warning: Exhausted best-first chunk memory for read chrY_25607312_25607822_7:0:0_4:0:0_13d4/1 (patid 986138); skipping read

When I googled the warning, I saw some suggestions of using --chunkmbs while running bowtie. I am not really sure what that does. I coudn't understand it from the manual. Still I used it with this command as it was suggested in one of the forums ./bowtie -p 8 -t --chunkmbs 256 -S hg19 -1 synthetic_sample1.fq -2 synthetic_sample2.fq > bowtie.sam. Then my error log says

Time loading reference: 00:00:01
Time loading forward index: 00:00:01
Time loading mirror index: 00:00:02
Error: Could not allocate ChunkPool of 268435456 bytes
Warning: Exhausted best-first chunk memory for read Error: Could not allocate ChunkPool of 268435456 bytes
Warning: Exhausted best-first chunk memory for read /var/spool/torque/mom_priv/jobs/2227.bora.qfab.org.SC: line 13:   374 Segmentation fault      ./bowtie -p 16 -t -x -S --chunkmbs 256 genome -1 synthetic_sample1.fq -2 synthetic_sample2.fq > bowtie_1node.sam

Can anyone please help me fix this problem?

PS: I ran BWA on the same data and I got 91% mapped reads. I used it on 1 node and then upto 5 nodes. I am trying to do the same thing with bowtie.

Thanks

bowtie mapping genome bwa paired-end • 8.2k views
ADD COMMENT
1
Entering edit mode
11.1 years ago
Matt LaFave ▴ 310

--chunkmbs essentially sets aside more memory for Bowtie to keep track of different alignment options, but the manual indicates that it only makes a difference when --best is enabled (the option that presents the "best" reads first). It seems strange that you'd get that warning without actually using --best. Were there any other options you used that weren't mentioned above?

As for the other problem - the Could not allocate ChunkPool of 268435456 bytes bit - that sounds like it may be a problem external to bowtie. It seems like your system is saying "you asked for 256 MB, but I don't have that much memory to allocate to you", so it throws an error. In both situations, maybe it's just the case that bowtie can't get the RAM it needs to do the job.

ADD COMMENT
0
Entering edit mode

I dont think RAM might be a problem. Because I used the same script which I used for BWA and it worked fine. I have given 10gb memory for the job. No I didn't miss any flags. That is the script which I used to execute Bowtie. I left to mention this. I did get a sam file from the first execution even after getting these many warnings. But the same file is very small; only 256 lines.

ADD REPLY

Login before adding your answer.

Traffic: 2977 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6