BWA using reads in FASTA format

[aggp11]

Hi,

Does BWA work if we provide reads in FASTA format instead of FASTQ?

Thanks,
Praful

[fubar]

Hi,

Does BWA work if we provide reads in FASTA format instead of FASTQ?

Thanks,
Praful

I'm sure you have already read the documentation for BWA, since it would be inconsiderate to waste bandwidth with this question otherwise? Since you seem to have missed it, no - fasta does not contain the base call quality information crucial for reliable short read alignment.

[lh3]

Actually bwa happens to accept reads in the fasta format.

[aggp11]

I'm sure you have already read the documentation for BWA, since it would be inconsiderate to waste bandwidth with this question otherwise? Since you seem to have missed it, no - fasta does not contain the base call quality information crucial for reliable short read alignment.

Fubar,

Thanks a lot for your reply. Like lh3 said, BWA seemed to work with reads in FASTA format. However, I have to look into the alignment in detail to see how reliable it is since the FASTA doesn't have the quality scores.

Praful

[aggp11]

Actually bwa happens to accept reads in the fasta format.

Thanks you lh3. I tried it and bwa did the alignment. I hve to still look into the details of the alignment.

[fubar]

Thanks you lh3. I tried it and bwa did the alignment. I hve to still look into the details of the alignment.

I'm glad it worked for you but perplexed since the documentation at http://bio-bwa.sourceforge.net/bwa.shtml seems to make it clear that bwa accepts (and indexes) reference genomes as fasta but requires the reads as fastq. Perhaps you have discovered some previously undiscovered benefits of not reading documentation? Does anyone know where this is actually documented?

eg, from the current release command line for single ended mapping, in.fq seems to suggest fastq as the input reads whereas the index command wants in.fasta as the documentation says.

iaas1:/data/ext_src$ bwa samse
Usage: bwa samse [-n max_occ] [-f out.sam] [-r RG_line] <prefix> <in.sai> <in.fq>
iaas1:/data/ext_src$ bwa index
Usage: bwa index [-a bwtsw|div|is] [-c] <in.fasta>

I guess it might assign a uniform quality score to all reads which probably allows some sort of alignment like old fashioned CAP3 and friends used to do - but it's missing all the important information about how trustworthy the reads are - might not matter if they're all good but will likely lead to errors due to uninformed overconfidence in poor reads.

[marcowanger]

I'm glad it worked for you but perplexed since the documentation at http://bio-bwa.sourceforge.net/bwa.shtml seems to make it clear that bwa accepts (and indexes) reference genomes as fasta but requires the reads as fastq. Perhaps you have discovered some previously undiscovered benefits of not reading documentation? Does anyone know where this is actually documented?

eg, from the current release command line for single ended mapping, in.fq seems to suggest fastq as the input reads whereas the index command wants in.fasta as the documentation says.

iaas1:/data/ext_src$ bwa samse
Usage: bwa samse [-n max_occ] [-f out.sam] [-r RG_line] <prefix> <in.sai> <in.fq>
iaas1:/data/ext_src$ bwa index
Usage: bwa index [-a bwtsw|div|is] [-c] <in.fasta>

I guess it might assign a uniform quality score to all reads which probably allows some sort of alignment like old fashioned CAP3 and friends used to do - but it's missing all the important information about how trustworthy the reads are - might not matter if they're all good but will likely lead to errors due to uninformed overconfidence in poor reads.

lh3 is the author of BWA ......

[ashuchawla]

Could I ask what is the command to use a FASTA file instead of a FASTQ file to align reads with a genome using BWA?

Actually bwa happens to accept reads in the fasta format.