Comments on: Summary of known bugs (so far) in 3.0b2 http://selab.janelia.org/people/eddys/blog/?p=195 The Eddy lab: genome sequence analysis Thu, 12 Aug 2010 11:54:36 -0400 http://wordpress.org/?v=2.8.4 hourly 1 By: Sean Eddy http://selab.janelia.org/people/eddys/blog/?p=195&cpage=1#comment-8785 Sean Eddy Thu, 04 Mar 2010 22:35:49 +0000 http://selab.janelia.org/people/eddys/blog/?p=195#comment-8785 No. HMMER3 currently does only local alignment, so it gives no special consideration to whether an alignment is full length or not. No. HMMER3 currently does only local alignment, so it gives no special consideration to whether an alignment is full length or not.

]]> By: Paul Havlak http://selab.janelia.org/people/eddys/blog/?p=195&cpage=1#comment-8782 Paul Havlak Thu, 04 Mar 2010 20:11:32 +0000 http://selab.janelia.org/people/eddys/blog/?p=195#comment-8782 I'm still coming up to speed on the full subtleties of protein alignments. It seems that having a * codon is at least as meaningful as having the initial M codon. Since that initial methionine is commonly removed post-translationally, the main significance of having both the M and the * is knowing that you've got a full-length protein model. What significance does HMMER3 ascribe to start and stop codons, that would result in scoring any different from that for alignment of a partial protein sequence? That is, given amino-acid sequences of the same length, aligning to the same references with the same BLAST bit scores, does HMMER3 adjust the score of a sequence with start and stop codons, based on their consistency with the reference-sequence's endpoints, compared with an unterminated sequence? I’m still coming up to speed on the full subtleties of protein alignments. It seems that having a * codon is at least as meaningful as having the initial M codon. Since that initial methionine is commonly removed post-translationally, the main significance of having both the M and the * is knowing that you’ve got a full-length protein model.

What significance does HMMER3 ascribe to start and stop codons, that would result in scoring any different from that for alignment of a partial protein sequence? That is, given amino-acid sequences of the same length, aligning to the same references with the same BLAST bit scores, does HMMER3 adjust the score of a sequence with start and stop codons, based on their consistency with the reference-sequence’s endpoints, compared with an unterminated sequence?

]]>