Bowtie vs BWA

[sarbashis]

Hi All,
I have illumina genomic sequencing data of a bacteria. When I aligned with bowtie only 1% aligned but using BWA (gap aligned) 76% aligned. The alignment showing lots of 1 nucleotide gaps on almost every 100 bp distance. Any suggestion what actually happening and how I handle this. Thank in advance

[GW_OK]

What's your estimated fold coverage?

[sarbashis]

@GW_OK
Could you kindly tell me how I calculate estimated fold coverage?

[Richard Finney]

Bowtie may still not support indels when doing alignment. I'm not sure, this may have been fixed (they "interoperate with indel callers", however).

http://www.google.com/search?hl=en&q=bowtie+indel+support+site%3Aseqanswers.com&aq=f&aqi=&aql=&oq=

[colindaven]

What is the genome size ? How much coverage were you expecting ? As Richard says I would aim for an indel capable aligner.

For bacteria I've had best results with Stampy or Novoalign to date.

[GW_OK]

@GW_OK
Could you kindly tell me how I calculate estimated fold coverage?

How many bases sequenced divided by estimated (or known) genome size equals fold coverage.

If you have something ridiculously high I can see where you might run into problems.

But the fact that bowtie doesn't do indels is also a very, very good point.

[sarbashis]

@GW_OK,
Thanks for you help. The genome size is approx 5MB and my fold coverage is 262.4. Using BWA, my average read depth (sum(depth in all position)/ Genome size ) is 161.8.

@colindaven,
Novoalign is not open source. I have tried stampy-1.0.13 but it showing error Could not import the maptools module. I am not able to find out which module I have to install.

[GW_OK]

That's a pretty high depth of coverage. You might try subsampling your data and see how that works.

[sarbashis]

@GW_OK, Thank you for your response. I will do a sampling of the data. I will take sample size of 60x coverage. Could you tell me why high depth coverage is not good.

[rskr]

I can't believe anyone would use a gapless aligner, for anything.

[GW_OK]

With high depth of coverage I think you can run into the error rate of the sequencer itself, leading to lots and lots of artifacts, like your every-other-nucleotide phenomenon.

[sarbashis]

@GW_OK, I have done sampling on the basis of very high quality reads and got good alignment. earlier what I got are actually sequencing error. Thank a lot.

[yana]

Bowtie may still not support indels when doing alignment. I'm not sure, this may have been fixed (they "interoperate with indel callers", however).

http://www.google.com/search?hl=en&q...&aqi=&aql=&oq=

Hi Richard,

May we have a bit more information to do "gap alignment" with bowtie?
If it's possible, it will be a great reason to replace my aligner! (bwa to bowtie)

Thanks in advance,