Mercurial > repos > cpt > cpt_psm_prep

annotate lib/CPT/Bio/SAR.pm @ 1:d724f34e671d draft default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
author cpt
date Mon, 05 Jun 2023 02:50:07 +0000
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
1
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
1 package CPT::Bio::SAR;
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
2 use strict;
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
3 use warnings;
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
4 use autodie;
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
5 use Moose;
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
6
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
7 sub filter_sar {
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
8 my ($self, @seqs) = @_;
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
9 my @good;
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
10 foreach(@seqs){
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
11 if(has_sar_motif($_)){
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
12 push(@good, $_);
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
13 }
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
14 }
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
15 return @good;
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
16 }
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
17
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
18 sub has_sar_motif {
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
19 my ( $self, $seq ) = @_;
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
20
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
21 return 0 if(length $seq < 40);
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
22
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
23 my $reg_a = qr/([^DEKR]{3}K[^DEKR]{8,}[^DER]{1}[^DEKR]{3})/;
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
24 my $reg_b = qr/([KR]{1,}[^DEKR]{12,}[^DER]{1}[^DEKR]{3})/;
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
25
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
26 my $first40 = substr( $seq, 0, 40 );
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
27
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
28 # there is a transmembrane domain in the first 40 AAs
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
29 # there is at least one positive charged AAs in front of the TMD
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
30 if ( $first40 =~ $reg_a || $first40 =~ $reg_b ) {
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
31 my $modi1st40 = $first40;
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
32 # Cut out the match, and then add the whole thing to the end.
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
33 my $t4homology =
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
34 #substr($seq,0, $-[0] ), # Before the match
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
35 #substr($seq, $-[0], ($+[0] - $-[0])), # the match
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
36 substr($seq, $+[0]). # After the match
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
37 $first40;
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
38 $t4homology = substr($t4homology, 0 , 40);
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
39
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
40 if ( $t4homology =~ qr/E[A-Z]{8}[DC][A-Z]{4,5}T/ ) {
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
41 return 1;
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
42 }
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
43 }
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
44 return 0;
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
45 }
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
46
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
47 no Moose;
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
48 1;
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
49
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
50 __END__
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
51
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
52 =pod
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
53
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
54 =encoding UTF-8
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
55
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
56 =head1 NAME
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
57
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
58 CPT::Bio::SAR
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
59
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
60 =head1 VERSION
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
61
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
62 version 1.99.4
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
63
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
64 =head1 AUTHOR
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
65
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
66 Eric Rasche <rasche.eric@yandex.ru>
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
67
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
68 =head1 COPYRIGHT AND LICENSE
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
69
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
70 This software is Copyright (c) 2014 by Eric Rasche.
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
71
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
72 This is free software, licensed under:
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
73
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
74 The GNU General Public License, Version 3, June 2007
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
75
d724f34e671d planemo upload commit 94b0cd1fff0826c6db3e7dc0c91c0c5a8be8bb0c
cpt
parents:
diff changeset
76 =cut