[Bioperl-guts-l] [Bug 2436] Inconsistencies in function HSPI::seq_inds when encountering gaps

bugzilla-daemon at portal.open-bio.org bugzilla-daemon at portal.open-bio.org
Thu Jan 31 11:50:24 EST 2008


http://bugzilla.open-bio.org/show_bug.cgi?id=2436





------- Comment #3 from cjfields at uiuc.edu  2008-01-31 11:50 EST -------
The positions correspond to the sequence, not the alignment.  In the first
BLAST report the first HSP is:

Query: 1  acgtacgtacgtacc-acgtacgtacgtacgtacgtacgtacgtacgtacgtacgtacgt 59
          ||||||||||||||  ||||||||||||||||||||||||||||||||||||||||||||
Sbjct: 1  acgtacgtacgtacgtacgtacgtacgtacgtacgtacgtacgtacgtacgtacgtacgt 60


Query: 60 acgtacgtacgtacgtacgtacgt 83
          ||||||||||||||||||||||||
Sbjct: 61 acgtacgtacgtacgtacgtacgt 84

Here is the script output:

Case one: Seq1 <=> Seq2
We have mismatch+gap, but the gap is not reported in the seq_inds list of the
query:
query no-match indices: 15 
subject no-match indices: 15 16

which makes sense when you realize that gap positions are called at the
sequence position prior to the gap insert (in the query, pos 15).  Since both
mismatch and gap would be 15 for the query, only one position is returned. 
This is exempilfied by the following HSP (just an edit of the one above):

Query: 1  acgtacgtacgtac--acgtacgtacgtacg-acgtacgt-cgtacgtacgtacgtacgt 56
          ||||||||||||||  ||||||||||||||| |||||||| |||||||||||||||||||
Sbjct: 1  acgtacgtacgtacgtacgtacgtacgtacgtacgtacgtacgtacgtacgtacgtacgt 60


Query: 57 acgtacgtacgtacgtacgtacgt 80
          ||||||||||||||||||||||||
Sbjct: 61 acgtacgtacgtacgtacgtacgt 84

which reports back:

Case one: Seq1 <=> Seq2
We have mismatch+gap, but the gap is not reported in the seq_inds list of the
query:
query no-match indices: 14 29 37 
subject no-match indices: 15 16 32 41 

Note the double gap is only reported once at query pos 14.

Make sense?


-- 
Configure bugmail: http://bugzilla.open-bio.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.


More information about the Bioperl-guts-l mailing list