SAM header missing in BFAST

[cdry7ue]

Nils,
I am using the command
bfast postprocess -f ref.fa -i run8.baf > run8.sam
I get the sam file but the header is missing. Is there a qucik way to get the header in there too.
thanks
-Ashish

[nilshomer]

Note: I moved this to its own thread since it had nothing to do with the previous thread.

The header is printed by BFAST and should not be missing. Could you post "head run8.sam"? Were there any warning/error messages?

Nils

[cdry7ue]

Nils,
My bad, there is no problem with BFAST. I was trying to have RSamtools package read a sam file and it was producing an error, apparently it only reads bam.

> X<-readBamGappedAlignments('run8_12.sam')
Error in .io_bam(.scan_bam, file, index, reverseComplement, tmpl, param = param) :
SAM/BAM header missing or empty
file: 'X:\Ashish\nextgen\ion3\bfast_processed\run8_12.sam'

However it works once it is turned into bam using samtools
samtools view -bS run8_default.sam >run8_default.bam

X<-readBamGappedAlignments('run8_default.bam')

Nils I was wondering if there is a way to filter reads from the .baf file based on the real
alignment quality score.(Also is there anyway this score can be seen in the sam file to filter on later ?) E.g. I would want to keep the read with score 3550, while getting rid of the read with -250 in score column(see below).

@MNZKY:4:144 0 1
GTG..bla.. :>>..bla..bla.. 1
1 1 + 3550 255 103 gtgc…bla..bla
@MNZKY:4:127 0 1
GTG..bla.. ))..bla..bla.. 1
1 1 + -250 255 108 gtgc…bla..bla

Thanks for your quick response....

[nilshomer]

Take a look at "bfast postprocess". You can also remove low quality alignments from the SAM by using the AS tag.

[cdry7ue]

I was wondering if there is scripts that are available to do that. If not I could always write my own to shave off lines with less than desired AS values. Btw are there any other dependencies in SAM like number of lines etc in the header?

[cdry7ue]

Nils,
I was wondering if the type of quality score encoding would alter the workings of BFast in any way. I am guessing as long as it is of PHRED type, the offset of 33 or 64 (Illumina) shouldnt really matter, since you are probably using some type of additive scheme in you local smith waterman scoring algorithm?
Thanks
Ashish

[nilshomer]

Nils,
I was wondering if the type of quality score encoding would alter the workings of BFast in any way. I am guessing as long as it is of PHRED type, the offset of 33 or 64 (Illumina) shouldnt really matter, since you are probably using some type of additive scheme in you local smith waterman scoring algorithm?
Thanks
Ashish

It doesn't matter in BFAST, but since the quality string is directly outputted to the SAM file, it may not be compliant if the qualities are in +64 format. My advice is to always have them in Sanger PHRED format.