Mercurial > repos > jjohnson > gmap

comparison gmap/gmap_build.xml @ 2:52da588232b0

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Add datatypes for maps and snpindex, add iit_store and snpindex tools, update GMAP and GSNAP to use these.
author Jim Johnson <jj@umn.edu>
date Fri, 21 Oct 2011 11:38:55 -0500
parents d58d272914e7
children
comparison
equal deleted inserted replaced
1:30d42bb409b8 2:52da588232b0
1 <tool id="gmap_build" name="GMAP Build" version="2.0.0"> 1 <tool id="gmap_build" name="GMAP Build" version="2.0.0">
2 <description>a GMAP DB Index</description> 2 <description>a database genome index for GMAP and GSNAP</description>
3 <requirements> 3 <requirements>
4 <requirement type="binary">gmap_build</requirement> 4 <requirement type="binary">gmap_build</requirement>
5 <!-- proposed tag for added datatype dependencies -->
6 <requirement type="datatype">gmapdb</requirement>
7 <requirement type="datatype">gmap_snps</requirement>
5 </requirements> 8 </requirements>
6 <version_string>gmap --version</version_string> 9 <version_string>gmap --version</version_string>
7 <command interpreter="command"> /bin/bash $shscript 2>1 1> $output </command> 10 <command interpreter="command"> /bin/bash $shscript 2>1 1> $output </command>
8 <inputs> 11 <inputs>
9 <!-- Input data --> 12 <!-- Name for this gmapdb -->
10 <param name="input" type="data" format="fasta" label="reference sequence fasta" /> 13 <param name="refname" type="text" label="Name you want to give this gmap database" help="">
11 <param name="refname" type="text" label="reference name" help="">
12 <validator type="empty_field" message="A database name is required."/> 14 <validator type="empty_field" message="A database name is required."/>
13 </param> 15 </param>
16 <!-- Input data -->
17 <repeat name="inputs" title="Reference Sequence" min="1">
18 <param name="input" type="data" format="fasta" label="reference sequence fasta" />
19 </repeat>
20
14 <param name="kmer" type="select" multiple="true" force_select="true" label="kmer size" help=""> 21 <param name="kmer" type="select" multiple="true" force_select="true" label="kmer size" help="">
15 <option value="12">12</option> 22 <option value="12">12</option>
16 <option value="13">13</option> 23 <option value="13">13</option>
17 <option value="14">14</option> 24 <option value="14">14</option>
18 <option value="15" selected="true">15</option> 25 <option value="15" selected="true">15</option>
77 </inputs> 84 </inputs>
78 <outputs> 85 <outputs>
79 <!-- 86 <!--
80 <data format="txt" name="log" label="${tool.name} on ${on_string}: log"/> 87 <data format="txt" name="log" label="${tool.name} on ${on_string}: log"/>
81 --> 88 -->
82 <!-- this should have its own datatype: gmapdb -->
83 <data format="gmapdb" name="output" label="${tool.name} on ${on_string} gmapdb ${refname}" /> 89 <data format="gmapdb" name="output" label="${tool.name} on ${on_string} gmapdb ${refname}" />
84 </outputs> 90 </outputs>
85 <configfiles> 91 <configfiles>
86 <configfile name="shscript"> 92 <configfile name="shscript">
87 #!/bin/bash 93 #!/bin/bash
88 #set $ds = chr(36) 94 #set $ds = chr(36)
89 #set $gt = chr(62) 95 #set $gt = chr(62)
90 #set $lt = chr(60) 96 #set $lt = chr(60)
91 #set $ad = chr(38) 97 #set $ad = chr(38)
98 ## #set $ref_files = ''
99 ## #for $i in $inputs:
100 ## #set $ref_files = $ref_files $i.input
101 ## #end for
102 ## echo $ref_files
92 #import os.path 103 #import os.path
93 #set $gmapdb = $output.extra_files_path 104 #set $gmapdb = $output.extra_files_path
94 #set $mapsdir = $os.path.join($os.path.join($gmapdb,str($refname)), str($refname) + '.maps') 105 #set $mapsdir = $os.path.join($os.path.join($gmapdb,str($refname)), str($refname) + '.maps')
95 mkdir -p $gmapdb 106 mkdir -p $gmapdb
96 ## export GMAPDB required for cmetindex and atoiindex 107 ## export GMAPDB required for cmetindex and atoiindex
97 export GMAPDB=$gmapdb 108 export GMAPDB=$gmapdb
98 #for $k in $kmer.__str__.split(','): 109 #for $k in $kmer.__str__.split(','):
99 gmap_build -D $gmapdb -d $refname -s numeric-alpha -k $k $input 110 gmap_build -D $gmapdb -d $refname -s numeric-alpha -k $k #for i in $inputs# ${i.input}#end for#
100 #end for 111 #end for
101 get-genome -D $gmapdb -d '?' | sed 's/^Available .*/gmap db: /' 112 get-genome -D $gmapdb -d '?' | sed 's/^Available .*/gmap db: /'
102 echo "kmers: " $kmer 113 echo "kmers: " $kmer
103 #if $splicesite.splice_source == 'refGeneTable': 114 #if $splicesite.splice_source == 'refGeneTable':
104 #if $splicesite.refGenes.__str__ != 'None': 115 #if $splicesite.refGenes.__str__ != 'None':
114 #if $splicesite.gff3Genes.__str__ != 'None': 125 #if $splicesite.gff3Genes.__str__ != 'None':
115 cat $splicesite.gff3Genes | gff3_splicesites | iit_store -o $os.path.join($mapsdir,'splicesites') 126 cat $splicesite.gff3Genes | gff3_splicesites | iit_store -o $os.path.join($mapsdir,'splicesites')
116 cat $splicesite.gff3Genes | gff3_introns | iit_store -o $os.path.join($mapsdir,'introns') 127 cat $splicesite.gff3Genes | gff3_introns | iit_store -o $os.path.join($mapsdir,'introns')
117 #end if 128 #end if
118 #end if 129 #end if
130 #if $dbsnp.snp_source != 'none' and $dbsnp.snps.__str__ != 'None':
119 #if $dbsnp.snp_source == 'snpTable': 131 #if $dbsnp.snp_source == 'snpTable':
120 #if $dbsnp.snps.__str__ != 'None':
121 #if $dbsnp.snpsex.__str__ != 'None': 132 #if $dbsnp.snpsex.__str__ != 'None':
122 cat $dbsnp.snps | dbsnp_iit -w $dbsnp.weight -e $dbsnp.snpsex | iit_store -o $os.path.join($mapsdir,'snps') 133 cat $dbsnp.snps | dbsnp_iit -w $dbsnp.weight -e $dbsnp.snpsex | iit_store -o $os.path.join($mapsdir,'snps')
123 #else: 134 #else:
124 cat $dbsnp.snps | dbsnp_iit -w $dbsnp.weight | iit_store -o $os.path.join($mapsdir,'snps') 135 cat $dbsnp.snps | dbsnp_iit -w $dbsnp.weight | iit_store -o $os.path.join($mapsdir,'snps')
125 #end if 136 #end if
137 #else:
138 cat $dbsnp.snps | iit_store -o $os.path.join($mapsdir,'snps')
139 #end if
126 snpindex -d $refname -v snps 140 snpindex -d $refname -v snps
127 #end if 141 echo "snpindex" -d $refname -v snps
128 #end if 142 #end if
129 #if $cmetindex.__str__ == 'yes': 143 #if $cmetindex.__str__ == 'yes':
130 cmetindex -d $refname 144 cmetindex -d $refname
131 echo "cmetindex" 145 echo "cmetindex" -d $refname
132 #end if 146 #end if
133 #if $atoiindex.__str__ == 'yes': 147 #if $atoiindex.__str__ == 'yes':
134 atoiindex -d $refname 148 atoiindex -d $refname
135 echo "atoiindex" 149 echo "atoiindex" -d $refname
136 #end if 150 #end if
137 get-genome -D $gmapdb -d $refname -m '?' | sed 's/^Available maps .*/maps: /' 151 get-genome -D $gmapdb -d $refname -m '?' | sed 's/^Available maps .*/maps: /'
138 </configfile> 152 </configfile>
139 </configfiles> 153 </configfiles>
140 154
144 <help> 158 <help>
145 159
146 160
147 **GMAP Build** 161 **GMAP Build**
148 162
149 GMAP Build creates an index of a genomic sequence for alignments using GMAP_ (Genomic Mapping and Alignment Program for mRNA and EST sequences) and GSNAP_ (Genomic Short-read Nucleotide Alignment Program). 163 GMAP Build creates an index of a genomic sequence for mapping and alignment using GMAP_ (Genomic Mapping and Alignment Program for mRNA and EST sequences) and GSNAP_ (Genomic Short-read Nucleotide Alignment Program). (GMAP Build uses GMSP commands: gmap_build, iit_store, psl_splicesites, psl_introns, gtf_splicesites, gtf_introns, gff3_splicesites, gff3_introns, dbsnp_iit, snpindex, cmetindex, and atoiindex.)
150 164
151 You will want to read the README_ 165 You will want to read the README_
152 166
153 Publication_ citation: Thomas D. Wu, Colin K. Watanabe Bioinformatics 2005 21(9):1859-1875; doi:10.1093/bioinformatics/bti310 167 Publication_ citation: Thomas D. Wu, Colin K. Watanabe Bioinformatics 2005 21(9):1859-1875; doi:10.1093/bioinformatics/bti310
154 168