bowtie memory warning

[crh]

Hi

I'm mapping solid reads using bowtie and received the following warnings for MANY reads. Mapping % is very low as well.
Am I running out of memory?

command line:
bowtie -C -S -k 3 --best --un unmapped_list genome_index -f seq.csfasta -Q seq.qual bowtie.sam



Warning: Exhausted best-first chunk memory for read 301_728_1008_F3 (patid 11858997); skipping read

# reads processed: 19323816
# reads with at least one reported alignment: 4408352 (22.81%)
# reads that failed to align: 14915464 (77.19%)
Reported 5326165 alignments to 1 output stream(s)

Thanks

charles

[kopardev]

I get similar warnings when I use paired end data. But if I align read1 and read2 separately, I do not get any warnings. Solutions/pointers are appreciated.
Thanks,
Vishal

[gntc]

Yes, I believe you are running out of memory. If you use --chunkmbs 200 you should stop getting the warning. This may help improve your alignment statistics.

[kopardev]

In fact, I have tried using --chunkmbs 1000 and still managed to get these errors ... only for paired-end data though.

[gntc]

Try higher I guess. 200 worked for me when working with non-paired end data.

[crh]

bowtie memory warning

All

Thanks for the advice:

Using --chunkmbs 200 eliminated the memory warnings.

I'm still not mapping a significant fraction of the reads. I'm going to trim the unmapped colorspace reads from 3'end to see if I can get any of these to map - other suggestions?

bowtie -C -S -k 3 --best


# reads processed: 19323816
# reads with at least one reported alignment: 4408352 (22.81%)
# reads that failed to align: 14915464 (77.19%)
Reported 5326165 alignments to 1 output stream(s)

[gladexp]

Bowtie: Memory error warning

Same as the quoted question. When I run the Bowtie command as below to generate SAM with the 1 best hit from fasta input, it has such warning messages,

$ bowtie -f -S -m 1 -k 1 --best h_sapiens_37_asm Input.fa > Output.sam
------------------
Warning: Exhausted best-first chunk memory for read seq_5023_1 (patid 5023); skipping read
Warning: Exhausted best-first chunk memory for read seq_5041_1 (patid 5041); skipping read
Warning: Exhausted best-first chunk memory for read seq_20071_1 (patid 20071); skipping read
......
------------------

Which flag options should I use?

Many thanks.

Ben

Hi

I'm mapping solid reads using bowtie and received the following warnings for MANY reads. Mapping % is very low as well.
Am I running out of memory?

command line:
bowtie -C -S -k 3 --best --un unmapped_list genome_index -f seq.csfasta -Q seq.qual bowtie.sam



Warning: Exhausted best-first chunk memory for read 301_728_1008_F3 (patid 11858997); skipping read

# reads processed: 19323816
# reads with at least one reported alignment: 4408352 (22.81%)
# reads that failed to align: 14915464 (77.19%)
Reported 5326165 alignments to 1 output stream(s)

Thanks

charles

[jay]

I had this problem as well. I tried --chunkmbs 200 as well, and it didn't help, but --chunkmbs 2000 seems like it might have fixed the bug being reported, but I am still getting very low alignmment proportions (5% of reads).