Module: GEO::SOFT
- Defined in:
- lib/MARQ/GEO.rb
Overview
Parse information in .soft files
Constant Summary collapse
- GEO_SOFT =
"http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?targ=self&view=full&form=text&acc="
- ID_FIX =
{ :mgi_unigene => proc{|gene| if gene then gene.match(/^Mm./) ? gene : "Mm." + gene end}, :human_unigene => proc{|gene| if gene then gene.match(/^Hs./) ? gene : "Hs." + gene end}, }
- @@nice =
1
- @@formats =
{{{ Guess the format of the IDS
{}
Class Method Summary collapse
-
.consecutive?(ids) ⇒ Boolean
Id list is in sequence.
-
.dna_sequence?(ids) ⇒ Boolean
ID are DNA bases.
-
.get_soft(item) ⇒ Object
Download a soft file.
- .GPL(platform) ⇒ Object
-
.GSE(series) ⇒ Object
{{{ Parse soft files for several GEO entities.
- .GSM(array) ⇒ Object
-
.guessIds(genes, org, name = nil) ⇒ Object
Guess the format of the id in the list.
-
.numerical?(ids) ⇒ Boolean
Id list is numerical.
Class Method Details
.consecutive?(ids) ⇒ Boolean
Id list is in sequence
82 83 84 |
# File 'lib/MARQ/GEO.rb', line 82 def self.consecutive?(ids) ids.collect{|id| id.to_i}.sort[0..19] == (1..20).to_a end |
.dna_sequence?(ids) ⇒ Boolean
ID are DNA bases
92 93 94 |
# File 'lib/MARQ/GEO.rb', line 92 def self.dna_sequence?(ids) ids.compact.select{|id| ! id.strip.match(/^[ATCG]+$/i)}.empty? end |
.get_soft(item) ⇒ Object
Download a soft file. Uses cache
57 58 59 60 61 62 63 64 65 66 67 68 69 70 |
# File 'lib/MARQ/GEO.rb', line 57 def self.get_soft(item) item = item.strip cache_file = File.join(CACHE_DIR, item + '.soft') if File.exist?( cache_file ) File.open(cache_file).read else content = Open.read(GEO_SOFT + item, :nocache => true, :nice => @@nice) raise "SOFT file error" if content !~ /!/ fout = File.open(cache_file,'w') fout.write content fout.close content end end |
.GPL(platform) ⇒ Object
182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 |
# File 'lib/MARQ/GEO.rb', line 182 def self.GPL(platform) if !File.exist?(File.join(DATA_DIR, 'platforms',"#{platform}.yaml")) && !File.exist?(File.join(DATA_DIR, 'platforms',"#{platform}.skip")) begin raise "Platform Blacklisted" if PLATFORM_BLACKLIST.include? platform if platform =~ /_/ organism = GPL(platform.match(/(.*?)_/)[1])[:organism] info = { :organism => organism, :title => "Merged platforms #{ platform }", } return info end soft = get_soft(platform) raise "SOFT file error" if soft !~ /!/ organisms = soft.scan(/!Platform_organism\s*=\s*(.*)/).collect{|v| v.first.strip} if organisms.empty? raise "No Organism information" else # This might happen actually GPL2529 organisms.delete('Schizosaccharomyces pombe') if organisms.include?('Saccharomyces cerevisiae') org_name = organisms.first end title = "" if soft.match(/!Platform_title\s*=\s*(.*)/) title = $1 end org = Organism.name2org(org_name) raise "Organism not identified: #{org_name}" if org.nil? if soft.match(/!platform_table_begin/) data = soft.split(/!platform_table_begin/s)[1].collect{|l| l.chomp.split(/\t/)} data.shift names = data.shift total = data.first.length genes = data.sort_by{ rand }[1..1000].collect{|v| v.first} id = guessIds(genes,org, names.first) other = nil other_pos = 0 other_count = 0 other_name = 0 if id.nil? (1..total - 1).to_a.each{|num| genes = data.collect{|v| v[num]} other = guessIds(genes,org, name = names[num]) if other && other[1] > other_count other_pos = num other_count = other[1] other_name = names[num] end } end else raise "Soft file incomplete" end info = {:organism => org, :BioMart_ID => id ? id.first : nil, :title => title } info[:other_ID_field] = [other_pos + 1, other_name] if other_pos > 0 Open.write(File.join(DATA_DIR, 'platforms',"#{platform}.yaml"), info.to_yaml) rescue Exception puts $!. puts $!.backtrace Open.write(File.join(DATA_DIR, 'platforms',"#{platform}.skip"), $!.) end end raise "Platform info for #{ platform } is not available and could not be automatically produced." if File.exist?(File.join(MARQ.datadir, 'GEO', 'platforms',"#{platform}.skip")) YAML::load(File.open(File.join(DATA_DIR, 'platforms',"#{platform}.yaml"))) end |
.GSE(series) ⇒ Object
{{{ Parse soft files for several GEO entities
116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 |
# File 'lib/MARQ/GEO.rb', line 116 def self.GSE(series) soft = get_soft(series) # Find platform if match = soft.scan(/!Series_platform_id\s*=?\s*(.*)/) platform = match.flatten.collect{|p| p.strip}.join("_") else raise "No Platform information" end # Find title if soft.match(/!Series_title \s*=?\s*(.*)/) title = $1 else raise "No Title information" end # Find summary if soft.match(/!Series_summary \s*=?\s*(.*)/) matches = soft.scan(/!Series_summary \s*=?\s*(.*)/).to_a description = matches.collect{|m| m.to_s.strip.sub(/!Series_summary \s*=?\s*/,'')}.join("\n") else raise "No Summary information" end # Find samples if soft.match(/!Series_sample_id \s*=?\s*(.*)/) matches = soft.scan(/!Series_sample_id \s*=?\s*(.*)/).to_a samples = matches.collect{|m| m.to_s.strip.sub(/!Series_sample_id \s*=?\s*/,'')} else raise "No Summary information" end { :platform => platform, :description =>description.strip, :title => title.strip, :samples => samples, } end |
.GSM(array) ⇒ Object
157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 |
# File 'lib/MARQ/GEO.rb', line 157 def self.GSM(array) soft = get_soft(array) # Find title if soft.match(/!Sample_title\s*=?\s*(.*)/) title = $1 else raise "No Title information" end # Find description if soft.match(/!Sample_description \s*=?\s*(.*)/) description = $1 else raise "No Description information" end { :description =>description.strip, :title => title.strip, } end |
.guessIds(genes, org, name = nil) ⇒ Object
Guess the format of the id in the list. The name parameter can be used to identify some exceptions
98 99 100 101 102 103 104 105 106 107 108 109 110 111 |
# File 'lib/MARQ/GEO.rb', line 98 def self.guessIds(genes,org, name = nil) @@formats[org] ||= Organism.id_formats(org) if consecutive?(genes) || dna_sequence?(genes) || (numerical?(genes) && (name.nil? || !name.match(/entrez/i))) id = nil else fix = ID_FIX[(org + "_" + name.downcase).to_sym] if name if fix genes = genes.collect{|gene| fix.call(gene)} end id = Organism.guessIdFormat(@@formats[org], genes) end id end |
.numerical?(ids) ⇒ Boolean
Id list is numerical
87 88 89 |
# File 'lib/MARQ/GEO.rb', line 87 def self.numerical?(ids) ids.compact.select{|id| ! id.match(/^\d+$/)}.uniq.length < ids.length.to_f / 10 end |