Editing SAM files

[gracezheng]

Dear all,

I want to remove reads that map to ribosomal RNAs. I tried to modify the sam file (output of TopHat), and remove the reads that correspond to rRNAs. However, I could not run cufflinks on the modified sam files. Then I tried to sort them by converting them to bam (using samtools), but even then, it still did not work. I got the following error messages from cuffdiff.

[22:39:41] Inspecting maps and determining fragment length distributions.
SAM error on line 66185: CIGAR op has zero length
SAM error on line 94994: CIGAR op has zero length
SAM error on line 135374: CIGAR op has zero length

Has anyone seen this error? What does it mean?

What would you recommend to remove ribosomal reads?

thanks,
Grace

[zhidkov.ilia]

How did you removed the reads from SAM which correspond to rRNA?

[gracezheng]

I just looked at the coordinates of each SAM output, and removed lines that overlapped with rRNA locations, as I read in each line.

I reasoned that since my SAM file was sorted before editing, it should have remained sorted if I just moved certain lines.

[zhidkov.ilia]

Regarding the error message you got, there is some discussion at:http://seqanswers.com/forums/showthread.php?t=3551
If you moving out the reads (from sam file) corresponding to rRNA, you can remove them from your original fastq file (using read IDs) and then re-run the alignment.