HELP: BFAST and SOLID PE

[kenietz]

Hi guys,
im sorry to bother you again with BFAST but im completely lost already. I would really appreciate some help.

I have SOLID PE data 50x35bp. I did the following:

1. created the CS index

2. created reads.fastq file from the F3 and F5 files:

@427_47_96
T3...2..2302232.3...3.0332212.22.211.220222...3.222
+
%!!!%!!%&&(%%%!%!!!(!&%)%&%(!%%!'%%!'%&%%%!!!%!%%%
@427_47_96
T.0.32222332.21222.22220.2..32322002
+
!+!(&%(%&%%!)%%%%!%%%%(!%!!%%%%%&*%

3. bfast match -f hg19.fa -r reads.fastq -A 1 -s 1 -e 250000 -w 0 -n 10 > reads.bmf
4. bfast localalign -f hg19.fa -m reads.bmf -A 1 -s 1 -e 250000 -n 10 -U > reads.baf
5. bfast postprocess -f hg19.fa -i reads.baf -a 3 -A 1 -Y 0 -n 10 -O 1 > reads.sam
6. samtools view -bS reads.sam > reads.bam
7. samtools sort reads.bam reads.sorted
8. samtools flagstat reads.sorted.bam

However i always get this in my flagstat:

500000 + 0 in total (QC-passed reads + QC-failed reads)
0 + 0 duplicates
175413 + 0 mapped (35.08%:-nan%)
500000 + 0 paired in sequencing
250000 + 0 read1
250000 + 0 read2
0 + 0 properly paired (0.00%:-nan%)
0 + 0 with itself and mate mapped
175413 + 0 singletons (35.08%:-nan%)
0 + 0 with mate mapped to a different chr
0 + 0 with mate mapped to a different chr (mapQ>=5)

No properly paired reads. It doesnt matter what kind of options i give to BFAST it always ends without properly paired reads.

During postprocessing i get this error which i have no idea what it means hence no idea what can i do in order to prevent it:

Postprocessing...
************************************************************
Estimating paired end distance...
Found only 0 distances to infer the insert size distribution
************************************************************
In function "GetPEDBins": Warning[OutOfRange]. Variable/Value: b->numDistances.
Message: Not enough distances to infer insert size distribution.
***** Warning *****
************************************************************

Am i doing something wrong? How should i proceed? Please help!

Thank you very much in advance!!!

[nilshomer]

The postprocess utility is the key here, since it cannot find ends to pair. Have you varied the "-S/-P" options to see if you can get any paired?

[kenietz]

I know postprocess is the problem. I tried all possible things. Using only -Y 0 or only -S 1 and -P 0. Btw, when using only -S 1 and -P 0 the command shows that is not using pairing. Something is wrong here:
Postprocessing...
************************************************************
Estimating paired end distance...
Found only 0 distances to infer the insert size distribution
************************************************************
In function "GetPEDBins": Warning[OutOfRange]. Variable/Value: b->numDistances.
Message: Not enough distances to infer insert size distribution.
***** Warning *****
************************************************************

Last night i started a job using bfast-0.6.5a with the same set of seqs and using the index made with bfast-0.7.a and it finished correctly. I see properly paired reads. So i think ver 0.7a or even git version 0.7b are broken for paired end. I think cos the sam file looks weird. Column number 9 where should be insert size was always zero for the reads that mapped:

427_49_140 89 chr19 10689591 27 50M = 10689591 0
427_49_140 165 chr19 10689591 0 * chr19 10689591 0

So thats why the message above. But i dont know how to solve that issue. So i will stick to ver 0.6.5.

[kenietz]

or the error could be in bfast localalign. i dont know exactly.