bfast match problem on fedora 14

[sammy07]

Hi everybody,

I cant make bfast work on Fedora 14.
I'm using unpaired SOLID data of 50 bp length.
In version 0.6.4e, the solid2fastq step did not work properly for me. In the last version (0.6.4f git:50fb580ff23fe839b72b49924908159bd35d6f86) from the git repository, the bfast match step fails with the following error message:

...
In total read 1 contigs for a total of 3171952 bases
************************************************************
Reading /home/user/work/analysis/reads.1.fastq into a temp file.
bfast: ../bfast/RGMatch.c:155: RGMatchPrint: Assertion `m->readLength > 0' failed.

The performed steps were:
1) solid2fastq -n 10000000 -o reads datafile.csfasta datafile.qual
2) ~/tools/bfast/bfast/bfast fasta2brg -f refGenome.fna
~/tools/bfast/bfast/bfast fasta2brg -f refGenome.fna -A 1
3) bfast index -f refGenome.fna -m 11111... -w 14 -i 1 -A 1 -n 8
4) bfast match -f refGenome.fna -A 1 -r reads1.fastq -n 8 > tmp.bmf

I want to try it on a virtual machine running Fedora 13. Would be happy about any further suggestions... Thanks in advance!

[nilshomer]

Looks like a read with zero bases, could you check for that in your fastq file?

Hi everybody,

I cant make bfast work on Fedora 14.
I'm using unpaired SOLID data of 50 bp length.
In version 0.6.4e, the solid2fastq step did not work properly for me. In the last version (0.6.4f git:50fb580ff23fe839b72b49924908159bd35d6f86) from the git repository, the bfast match step fails with the following error message:

...
In total read 1 contigs for a total of 3171952 bases
************************************************************
Reading /home/user/work/analysis/reads.1.fastq into a temp file.
bfast: ../bfast/RGMatch.c:155: RGMatchPrint: Assertion `m->readLength > 0' failed.

The performed steps were:
1) solid2fastq -n 10000000 -o reads datafile.csfasta datafile.qual
2) ~/tools/bfast/bfast/bfast fasta2brg -f refGenome.fna
~/tools/bfast/bfast/bfast fasta2brg -f refGenome.fna -A 1
3) bfast index -f refGenome.fna -m 11111... -w 14 -i 1 -A 1 -n 8
4) bfast match -f refGenome.fna -A 1 -r reads1.fastq -n 8 > tmp.bmf

I want to try it on a virtual machine running Fedora 13. Would be happy about any further suggestions... Thanks in advance!

[sammy07]

Hi Nils,

you are right, this seems to be a problem with the data integrity. The problem looked like that:

@160_774_71
T022221123122231102
+
BBBBBBABBABBABBBBABB+'))?@;55<<6@/);/1<5?91)-/5@9+

If these four lines are removed from the file, it works. Thank you!

[nilshomer]

If you run the "solid2fastq" step again, do you get the same problem read? Could you pull out the same read in the CSFAST/QUAL files so I can see what is going on?

[sammy07]

Ok, it seems that one pair of my .csfasta / .qual files was broken. Both just stop in the middle of a line without a newline character, which produces the short read in the solid2fastq output.

Another strange thing is that the .csfasta and the corresponding .qual file even have a different number of entries (the .qual file is longer), and solid2fastq did not produce any error message concerning this; possibly it just discarded the rest.

[nilshomer]

Sounds like your .csfasta / .qual files are corrupt. They should (IRC) have the same # of lines. It might be a good time to call up some support,

Nils