rbbt-sources 2.0.0 → 2.0.1
Sign up to get free protection for your applications and to get access to all the features.
- data/lib/rbbt/sources/InterPro.rb +5 -5
- data/lib/rbbt/sources/genomes1000.rb +17 -10
- data/lib/rbbt/sources/pfam.rb +2 -2
- metadata +8 -8
@@ -7,16 +7,16 @@ module InterPro
|
|
7
7
|
extend Resource
|
8
8
|
self.subdir = "share/databases/InterPro"
|
9
9
|
|
10
|
-
InterPro.claim InterPro.source.protein2ipr
|
10
|
+
InterPro.claim InterPro.source.protein2ipr, :url, "ftp://ftp.ebi.ac.uk/pub/databases/interpro/protein2ipr.dat.gz"
|
11
11
|
|
12
|
-
InterPro.claim InterPro.protein_domains
|
12
|
+
InterPro.claim InterPro.protein_domains, :proc do
|
13
13
|
organism = "Hsa"
|
14
14
|
uniprot_colum = TSV::Parser.new(Organism.protein_identifiers(organism).open).all_fields.index("UniProt/SwissProt Accession")
|
15
15
|
uniprots = CMD.cmd("grep -v '^#'|cut -f #{uniprot_colum+1}", :in => Organism.protein_identifiers(organism).open).read.split("\n").collect{|l| l.split("|")}.flatten.uniq.reject{|l| l.empty?}
|
16
16
|
|
17
17
|
tsv = nil
|
18
18
|
TmpFile.with_file(uniprots * "\n") do |tmpfile|
|
19
|
-
tsv = TSV.open(CMD.cmd("cut -f 1,2,5,6 | sort -u |grep -w -f #{ tmpfile }", :in => InterPro.source.protein2ipr.open, :pipe => true), :merge => true, :type => :double)
|
19
|
+
tsv = TSV.open(CMD.cmd("cut -f 1,2,5,6 | sort -u |grep -w -F -f #{ tmpfile }", :in => InterPro.source.protein2ipr.open, :pipe => true), :merge => true, :type => :double)
|
20
20
|
end
|
21
21
|
|
22
22
|
tsv.key_field = "UniProt/SwissProt Accession"
|
@@ -24,7 +24,7 @@ module InterPro
|
|
24
24
|
tsv.to_s
|
25
25
|
end
|
26
26
|
|
27
|
-
InterPro.claim InterPro.domain_names
|
27
|
+
InterPro.claim InterPro.domain_names, :proc do
|
28
28
|
#tsv = InterPro.source.protein2ipr.tsv :key_field => 1, :fields => [2], :type => :single
|
29
29
|
tsv = TSV.open(CMD.cmd("cut -f 2,3 | sort -u", :in => InterPro.source.protein2ipr.open, :pipe => true), :merge => true, :type => :single)
|
30
30
|
|
@@ -77,7 +77,7 @@ if defined? Entity
|
|
77
77
|
|
78
78
|
property :genes => :array2single do
|
79
79
|
InterPro.gene_index.values_at(*self).
|
80
|
-
collect{|genes| genes = genes.uniq; genes.organism = organism if genes.respond_to? :organism; genes }.tap{|o| Gene.setup(o, "UniProt/SwissProt Accession", organism)}
|
80
|
+
collect{|genes| genes = [] if genes.nil?; genes = genes.uniq; genes.organism = organism if genes.respond_to? :organism; genes }.tap{|o| Gene.setup(o, "UniProt/SwissProt Accession", organism)}
|
81
81
|
end
|
82
82
|
end
|
83
83
|
|
@@ -8,22 +8,29 @@ module Genomes1000
|
|
8
8
|
|
9
9
|
RELEASE_URL = "ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/20110521/ALL.wgs.phase1_release_v3.20101123.snps_indels_sv.sites.vcf.gz"
|
10
10
|
|
11
|
-
Genomes1000.claim Genomes1000.mutations, :proc do
|
11
|
+
Genomes1000.claim Genomes1000.mutations, :proc do |filename|
|
12
12
|
|
13
|
-
|
14
|
-
|
15
|
-
|
13
|
+
begin
|
14
|
+
Open.write(filename) do |file|
|
15
|
+
file.puts "#: :type=:single#:namespace=Hsa"
|
16
|
+
file.puts "#Variant ID\tGenomic Mutation"
|
16
17
|
|
17
|
-
|
18
|
+
Open.read(RELEASE_URL) do |line|
|
19
|
+
next if line[0] == "#"[0]
|
18
20
|
|
19
|
-
|
20
|
-
end
|
21
|
-
|
22
|
-
tsv.namespace = "Hsa"
|
21
|
+
chromosome, position, id, references, alternative, quality, filter, info = line.split("\t")
|
23
22
|
|
24
|
-
|
23
|
+
file.puts [id, [chromosome, position, alternative] * ":"] * "\t"
|
24
|
+
end
|
25
|
+
end
|
26
|
+
rescue
|
27
|
+
FileUtils.rm filename if File.exists? filename
|
28
|
+
raise $!
|
29
|
+
end
|
30
|
+
nil
|
25
31
|
end
|
26
32
|
|
33
|
+
|
27
34
|
Genomes1000.claim Genomes1000.mutations_hg18, :proc do
|
28
35
|
require 'rbbt/sources/organism'
|
29
36
|
|
data/lib/rbbt/sources/pfam.rb
CHANGED
@@ -31,7 +31,7 @@ module InterPro
|
|
31
31
|
end
|
32
32
|
end
|
33
33
|
|
34
|
-
InterPro.claim InterPro.pfam_names
|
34
|
+
InterPro.claim InterPro.pfam_names, :proc do
|
35
35
|
pfam_domains = Pfam.domains.read.split("\n").collect{|l| l.split("\t").first}.compact.flatten
|
36
36
|
tsv = nil
|
37
37
|
TmpFile.with_file(pfam_domains * "\n") do |tmpfile|
|
@@ -42,7 +42,7 @@ InterPro.claim InterPro.pfam_names.find, :proc do
|
|
42
42
|
tsv.to_s
|
43
43
|
end
|
44
44
|
|
45
|
-
InterPro.claim InterPro.pfam_equivalences
|
45
|
+
InterPro.claim InterPro.pfam_equivalences, :proc do
|
46
46
|
pfam_domains = Pfam.domains.read.split("\n").collect{|l| l.split("\t").first}.compact.flatten
|
47
47
|
tsv = nil
|
48
48
|
TmpFile.with_file(pfam_domains * "\n") do |tmpfile|
|
metadata
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: rbbt-sources
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 2.0.
|
4
|
+
version: 2.0.1
|
5
5
|
prerelease:
|
6
6
|
platform: ruby
|
7
7
|
authors:
|
@@ -9,7 +9,7 @@ authors:
|
|
9
9
|
autorequire:
|
10
10
|
bindir: bin
|
11
11
|
cert_chain: []
|
12
|
-
date:
|
12
|
+
date: 2013-01-07 00:00:00.000000000 Z
|
13
13
|
dependencies:
|
14
14
|
- !ruby/object:Gem::Dependency
|
15
15
|
name: rbbt-util
|
@@ -136,12 +136,12 @@ files:
|
|
136
136
|
- share/install/Organism/organism_helpers.rb
|
137
137
|
- share/install/STITCH/Rakefile
|
138
138
|
- share/install/lib/helpers.rb
|
139
|
-
- test/test_helper.rb
|
140
|
-
- test/rbbt/sources/test_entrez.rb
|
141
|
-
- test/rbbt/sources/test_pubmed.rb
|
142
139
|
- test/rbbt/sources/test_go.rb
|
140
|
+
- test/rbbt/sources/test_entrez.rb
|
143
141
|
- test/rbbt/sources/test_biomart.rb
|
144
142
|
- test/rbbt/sources/test_organism.rb
|
143
|
+
- test/rbbt/sources/test_pubmed.rb
|
144
|
+
- test/test_helper.rb
|
145
145
|
homepage: http://github.com/mikisvaz/rbbt-sources
|
146
146
|
licenses: []
|
147
147
|
post_install_message:
|
@@ -167,9 +167,9 @@ signing_key:
|
|
167
167
|
specification_version: 3
|
168
168
|
summary: Data sources for the Ruby Bioinformatics Toolkit (rbbt)
|
169
169
|
test_files:
|
170
|
-
- test/test_helper.rb
|
171
|
-
- test/rbbt/sources/test_entrez.rb
|
172
|
-
- test/rbbt/sources/test_pubmed.rb
|
173
170
|
- test/rbbt/sources/test_go.rb
|
171
|
+
- test/rbbt/sources/test_entrez.rb
|
174
172
|
- test/rbbt/sources/test_biomart.rb
|
175
173
|
- test/rbbt/sources/test_organism.rb
|
174
|
+
- test/rbbt/sources/test_pubmed.rb
|
175
|
+
- test/test_helper.rb
|