bbmap error rate - subs - indels

[danova]

HI,
I´m running a pacbio experiment where i´m testing a mock against a set of pacbio reads.
I´m looking for the error rate , subs, and indwells. Exactly like the Pacbioalign output.

Is it possible to get the same information from each sequence in the reference database (actually my reference database contains 10 16S sequences)?

/usr/bin/java -ea -Xmx26g -cp bbmap/current align2.BBMapPacBio ref=mock_16s.fa in=reads_of_insert.fasta
out=mapped.sam mhist=mhist.txt ehist=ehist.txt indelhist=indelhist.txt

Match Rate: NA NA 97.3943% 41104155
Error Rate: 99.5618% 28853 2.4870% 1049594
Sub Rate: 97.2774% 28191 1.2914% 545011
Del Rate: 87.9124% 25477 0.9838% 415211
Ins Rate: 41.4700% 12018 0.2118% 89372
N Rate: 30.0932% 8721 0.1188% 50127

[Brian Bushnell]

If I am understanding correctly, you want the error rates broken down individually per reference sequence. BBMap can't do that directly. I think the best approach would be to split the data with Seal and then perform 10 mappings, like this:

seal.sh in=reads_of_insert.fasta ref=mock_16s.fa ambig=all k=27 edist=1 pattern=out_%.fasta

That will create one output file per reference sequence. Next, make one reference file per reference sequence:

16S_1.fasta
16S_2.fasta
...etc

And map 10 times:

/usr/bin/java -ea -Xmx26g -cp bbmap/current align2.BBMapPacBio ref=16S_1.fasta in=out_16S_1.fasta out=mapped_1.sam mhist=mhist_1.txt ehist=ehist_1.txt indelhist=indelhist_1.txt

[danova]

thanks Brian,
Looks like a good approach.