op_at_rpos is unable to detect the deletion in the transcript even though it is there.

[kirasato0211]

Hi,

I am using BIO-CIGAR module to identify insertions, deletions and polyA tail in the given transcript.

The example transcript is NM_001143805.1. This is a negative strand transcript and CIGAR string for this transcript is as below.

106M1D494M1X1969M1D650M1X470M62058N136M

As this transcript mapped on negative strand, i have reversed the CIGAR string as below.

136M62058N470M1X650M1D1969M1X494M1D106M

Bait position is NM_001143805.1:3708-3828. There is a deletion in this bait, but op_at_rpos is unable to detect it.

Could you please help me in this case? Also, is there a way to reverse the CIGAR string for negative strand.?

Thank You.

[tsibley]

Bait position is NM_001143805.1:3708-3828. There is a deletion in this bait, but op_at_rpos is unable to detect it.

Hmm, what do you mean by "unable to detect it"? The reversed CIGAR string starts as:

136M62058N…

which puts the 3708-3828 range squarely within the ~62kb long N operation there. From the Bio::Cigar docs (quoting from the SAM v1 spec), an N means "skipped region from the reference" and notes that:

For mRNA-to-genome alignment, an N operation represents an intron. For other types of alignments, the interpretation of N is not defined.

This suggests to me that your alignment/mapping tool that produced this CIGAR string decided the 62k region is something other than a "deletion" (D) in the sense of a CIGAR string. I don't think there's any issue with Bio::Cigar here.

Also, is there a way to reverse the CIGAR string for negative strand.?

This function should do the trick. It takes a Bio::Cigar object, reverses it, and returns a new Bio::Cigar object.

sub reverse_cigar {
    my $fwd = shift;
    my $rev = Bio::Cigar->new(join "", map { join "", @$_ } reverse @{$fwd->ops});
    return $rev;
}

Might be worth including a ->reverse method in Bio::Cigar itself!

[kirasato0211]

I should have been more clear about how i am using BIO::CIGAR module in my script. I am using first qpos_to_rpos to convert the transcript to reference location. Once i received the reference position, i am subtracting relative intron summation from it to get the final relative position on the reference. This final relative position on the reference is given as input to op_at_rpos.

This logic is working correctly for positive strand and reporting insertion, deletion and polyA tail at the correct position. Also it is reporting polyA tail correctly for negative strand.

Hence i believe it should work for deletion for negative strand.

Do you have any immediate plan to add revered method in Bio::Cigar module?

Thank you!

[tsibley]

Thanks for your explanation. I'm not sure I follow it correctly though. Can you produce some minimal example code that I can run to reproduce/demonstrate the issue?

I opened a PR adding a ->reversed method, but it's gotten no interaction from https://github.com/MullinsLab so far. Eventually I'll likely fork, merge that PR, and make a new CPAN release, but right now I don't have an immediate plan to do that.

[kirasato0211]

I have modified my code to use reference position (the position retrieved from method qpos to rpos()) to detect the deletion in case of negative strand and it is working correctly.

I appreciate your assistance in this case.

Thank you.