Mercurial > repos > jackcurragh > ribogalaxy_bowtie_genome

--- a/bowtie_genome_wrapper/bowtie_genomic_wrapper.py	Fri Apr 22 11:34:02 2022 +0000
+++ b/bowtie_genome_wrapper/bowtie_genomic_wrapper.py	Fri May 13 09:31:48 2022 +0000
@@ -411,10 +411,16 @@
         # have to nest try-except in try-finally to handle 2.4
         try:
             # prepare actual mapping commands
+
             if options.paired == 'paired':
-                cmd2 = 'bowtie %s %s -1 %s -2 %s > %s' % ( aligning_cmds, ref_file_name, options.input1, options.input2, options.output )
+                # cmd2 = 'bowtie %s %s -1 %s -2 %s > %s | samtools view -bS > %s' % ( aligning_cmds, ref_file_name, options.input1, options.input2, options.output, options.output )
+                cmd2 = 'bowtie %s %s -1 %s -2 %s  > %s' % ( aligning_cmds, ref_file_name, options.input1, options.input2, options.output )
+
             else:
+                print('bowtie %s %s %s | samtools view -bS > %s' % ( aligning_cmds, ref_file_name, options.input1, options.output ))
+                # cmd2 = 'bowtie %s %s %s | samtools view -bS > %s' % ( aligning_cmds, ref_file_name, options.input1, options.output )
                 cmd2 = 'bowtie %s %s %s > %s' % ( aligning_cmds, ref_file_name, options.input1, options.output )
+
             # align
             tmp = tempfile.NamedTemporaryFile( dir=tmp_index_dir ).name
             with open(tmp, 'w') as tmp_stderr:
--- a/bowtie_genome_wrapper/bowtie_genomic_wrapper.xml	Fri Apr 22 11:34:02 2022 +0000
+++ b/bowtie_genome_wrapper/bowtie_genomic_wrapper.xml	Fri May 13 09:31:48 2022 +0000
@@ -1,7 +1,9 @@
-<tool id="bowtie_genomic_wrapper" name="Align to the Genome with Bowtie" version="1.2.0">
-  <description></description>
+<tool id="bowtie_genomic_wrapper" name="Bowtie Genome Alignment" version="1.3.0">
+  <description>Align to the Genome with Bowtie</description>
   <requirements>
     <requirement type="package" version="1.2.0">bowtie</requirement>
+    <requirement type="package" version="1.13">samtools</requirement>
+
   </requirements>
   <version_command>bowtie --version</version_command>
   <command>
@@ -160,7 +162,7 @@
       </param>
       <when value="indexed">
         <param name="index" type="select" label="Select a reference genome" help="if your genome of interest is not listed - contact Galaxy team">
-          <options from_data_table="bowtie_genome_indexes">
+          <options from_data_table="bowtie_indexes">
             <filter type="sort_by" column="2" />
             <validator type="no_options" message="No indexes are available" />
           </options>
@@ -427,7 +429,7 @@
         <conditional name="refGenomeSource.genomeSource">
           <when value="indexed">
             <action type="metadata" name="dbkey">
-              <option type="from_data_table" name="bowtie_genome_indexes" column="1" offset="0">
+              <option type="from_data_table" name="bowtie_indexes" column="1" offset="0">
                 <filter type="param_value" column="0" value="#" compare="startswith" keep="False"/>
                 <filter type="param_value" ref="refGenomeSource.index" column="0"/>
               </option>
@@ -545,246 +547,18 @@
       -p is the number of threads. You need to replace the + with 2 dashes.
       chrM_base needs to be the base location/name of the index files.
       -->
-      <param name="genomeSource" value="indexed" />
+      <param name="genomeSource" value="history" />
       <!-- this is the backwards-compatible "unique value" for this index, not an actual path -->
-      <param name="index" value="equCab2chrM" />
+      <param name="ownFile" value="Homo_sapiens.GRCh38.dna.chromosome.9.fa" />
       <param name="sPaired" value="single" />
-      <param name="sInput1" ftype="fastqsanger" value="bowtie_in2.fastqsanger" />
-      <param name="sSettingsType" value="preSet" />
-      <param name="suppressHeader" value="true" />
-      <output name="output" ftype="sam" file="bowtie_out6.sam" sort="True">
-        <metadata name="dbkey" value="equCab2" />
-      </output>
-    </test>
-    <test>
-      <!--
-      Bowtie command:
-      bowtie-build -f test-data/phiX.fasta phiX_base
-      bowtie -q -X 1000 +ff -p 4 -S +sam-nohead -n 2 -e 70 -l 28 +pairtries 100 +maxbts 800 +best +un bowtie_out8_u.fastq phiX_base -1 test-data/bowtie_in5.fastqsanger -2 test-data/bowtie_in6.fastqsanger > bowtie_out7_u.sam
-      sort bowtie_out7_u.sam > bowtie_out7.sam
-      sort bowtie_out8_u_1.sam > bowtie_out8_1.sam
-      sort bowtie_out8_u_2.sam > bowtie_out8_2.sam
-      Then also need to modify bowtie_out8_1.sam and bowtie_out8_2.sam so that all @ lines come before sequence lines.
-      -p is the number of threads. You need to replace the + with 2 dashes.
-      The two unmapped output files will be named bowtie_out8_1.fastq and bowtie_out8_2.fastq.
-      chrM_base is the index files' location/base name.
-      -->
-      <param name="genomeSource" value="history" />
-      <param name="ownFile" value="phiX.fasta" />
-      <param name="indexSettings" value="indexPreSet" />
-      <param name="sPaired" value="paired" />
-      <param name="pInput1" ftype="fastqsanger" value="bowtie_in5.fastqsanger" />
-      <param name="pInput2" ftype="fastqsanger" value="bowtie_in6.fastqsanger" />
-      <param name="pMaxInsert" value="1000" />
-      <param name="pMateOrient" value="ff" />
-      <param name="pSettingsType" value="full" />
-      <param name="pSkip" value="0" />
-      <param name="pAlignLimit" value="-1" />
-      <param name="pTrimH" value="0" />
-      <param name="pTrimL" value="0" />
-      <param name="alignMode" value="nMode" />
-      <param name="pMismatchSeed" value="2" />
-      <param name="pMismatchQual" value="70" />
-      <param name="pSeedLen" value="28" />
-      <param name="pRounding" value="round" />
-      <param name="pMinInsert" value="0" />
-      <param name="pMaxAlignAttempt" value="100" />
-      <param name="pForwardAlign" value="forward" />
-      <param name="pReverseAlign" value="reverse" />
-      <param name="pTryHard" value="noTryHard" />
-      <param name="pValAlign" value="1" />
-      <param name="pAllValAligns" value="noAllValAligns" />
-      <param name="pSuppressAlign" value="-1" />
-      <param name="pUnmappedFile" value="true" />
-      <param name="pMaxFile" value="false" />
-      <param name="pBest" value="doBest" />
-      <param name="pdMaxBacktracks" value="800" />
-      <param name="pdStrata" value="noStrata" />
-      <param name="pOffrate" value="-1" />
-      <param name="pSeed" value="-1" />
-      <param name="suppressHeader" value="true" />
-      <output name="output" ftype="sam" file="bowtie_out7.sam" sort="True" />
-      <output name="output_unmapped_reads_l" ftype="fastqsanger" file="bowtie_out8_1.fastq" sort="True" />
-      <output name="output_unmapped_reads_r" ftype="fastqsanger" file="bowtie_out8_2.fastq" sort="True" />
-    </test>
-    <!-- start testing of non-sanger variant fastq reads -->
-    <test>
-      <param name="genomeSource" value="history" />
-      <param name="ownFile" value="phiX.fasta" />
-      <param name="indexSettings" value="indexPreSet" />
-      <param name="sPaired" value="paired" />
-      <param name="pInput1" ftype="fastqillumina" value="bowtie_in5.fastqillumina" />
-      <param name="pInput2" ftype="fastqillumina" value="bowtie_in6.fastqillumina" />
-      <param name="pMaxInsert" value="1000" />
-      <param name="pMateOrient" value="ff" />
-      <param name="pSettingsType" value="full" />
-      <param name="pSkip" value="0" />
-      <param name="pAlignLimit" value="-1" />
-      <param name="pTrimH" value="0" />
-      <param name="pTrimL" value="0" />
-      <param name="alignMode" value="nMode" />
-      <param name="pMismatchSeed" value="2" />
-      <param name="pMismatchQual" value="70" />
-      <param name="pSeedLen" value="28" />
-      <param name="pRounding" value="round" />
-      <param name="pMinInsert" value="0" />
-      <param name="pMaxAlignAttempt" value="100" />
-      <param name="pForwardAlign" value="forward" />
-      <param name="pReverseAlign" value="reverse" />
-      <param name="pTryHard" value="noTryHard" />
-      <param name="pValAlign" value="1" />
-      <param name="pAllValAligns" value="noAllValAligns" />
-      <param name="pSuppressAlign" value="-1" />
-      <param name="pUnmappedFile" value="true" />
-      <param name="pMaxFile" value="false" />
-      <param name="pBest" value="doBest" />
-      <param name="pdMaxBacktracks" value="800" />
-      <param name="pdStrata" value="noStrata" />
-      <param name="pOffrate" value="-1" />
-      <param name="pSeed" value="-1" />
-      <param name="suppressHeader" value="true" />
-      <output name="output" ftype="sam" file="bowtie_out7.sam" sort="True" />
-      <output name="output_unmapped_reads_l" ftype="fastqillumina" file="bowtie_out8_1.fastqillumina.sorted" sort="True" />
-      <output name="output_unmapped_reads_r" ftype="fastqillumina" file="bowtie_out8_2.fastqillumina.sorted" sort="True" />
-    </test>
-    <test>
-      <param name="genomeSource" value="history" />
-      <param name="ownFile" value="phiX.fasta" />
-      <param name="indexSettings" value="indexPreSet" />
-      <param name="sPaired" value="paired" />
-      <param name="pInput1" ftype="fastqsolexa" value="bowtie_in5.fastqsolexa" />
-      <param name="pInput2" ftype="fastqsolexa" value="bowtie_in6.fastqsolexa" />
-      <param name="pMaxInsert" value="1000" />
-      <param name="pMateOrient" value="ff" />
-      <param name="pSettingsType" value="full" />
-      <param name="pSkip" value="0" />
-      <param name="pAlignLimit" value="-1" />
-      <param name="pTrimH" value="0" />
-      <param name="pTrimL" value="0" />
-      <param name="alignMode" value="nMode" />
-      <param name="pMismatchSeed" value="2" />
-      <param name="pMismatchQual" value="70" />
-      <param name="pSeedLen" value="28" />
-      <param name="pRounding" value="round" />
-      <param name="pMinInsert" value="0" />
-      <param name="pMaxAlignAttempt" value="100" />
-      <param name="pForwardAlign" value="forward" />
-      <param name="pReverseAlign" value="reverse" />
-      <param name="pTryHard" value="noTryHard" />
-      <param name="pValAlign" value="1" />
-      <param name="pAllValAligns" value="noAllValAligns" />
-      <param name="pSuppressAlign" value="-1" />
-      <param name="pUnmappedFile" value="true" />
-      <param name="pMaxFile" value="false" />
-      <param name="pBest" value="doBest" />
-      <param name="pdMaxBacktracks" value="800" />
-      <param name="pdStrata" value="noStrata" />
-      <param name="pOffrate" value="-1" />
-      <param name="pSeed" value="-1" />
-      <param name="suppressHeader" value="true" />
-      <output name="output" ftype="sam" file="bowtie_out7.sam" sort="True" />
-      <output name="output_unmapped_reads_l" ftype="fastqsolexa" file="bowtie_out8_1.fastqsolexa.sorted" sort="True" />
-      <output name="output_unmapped_reads_r" ftype="fastqsolexa" file="bowtie_out8_2.fastqsolexa.sorted" sort="True" />
-    </test>
-    <!-- end testing of non-sanger variant fastq reads -->
-    <test>
-      <!--
-      Bowtie command:
-      bowtie -q -p 4 -S +sam-nohead -n 2 -e 70 -l 28 -y -k 1 chrM_base test-data/bowtie_in2.fastqsanger > bowtie_out9_u.sam
-      sort bowtie_out9_u.sam > bowtie_out9.sam
-      -p is the number of threads. You need to replace the + with 2 dashes.
-      chrM_base is the index files' location/base name.
-      -->
-      <param name="genomeSource" value="indexed" />
-      <!-- this is the backwards-compatible "unique value" for this index, not an actual path -->
-      <param name="index" value="equCab2chrM" />
-      <param name="sPaired" value="single" />
-      <param name="sInput1" ftype="fastqsanger" value="bowtie_in2.fastqsanger" />
+      <param name="sInput1" ftype="fastqsanger" value="sampled_fq.fastq.fastq" />
       <param name="sSettingsType" value="full" />
-      <param name="sSkip" value="0" />
-      <param name="sAlignLimit" value="-1" />
-      <param name="sTrimH" value="0" />
-      <param name="sTrimL" value="0" />
-      <param name="alignMode" value="nMode" />
-      <param name="sMismatchSeed" value="2" />
-      <param name="sMismatchQual" value="70" />
-      <param name="sSeedLen" value="28" />
-      <param name="sRounding" value="round" />
-      <param name="sForwardAlign" value="forward" />
-      <param name="sReverseAlign" value="reverse" />
-      <param name="sTryHard" value="doTryHard" />
-      <param name="sValAlign" value="1" />
-      <param name="sAllValAligns" value="noAllValAligns" />
-      <param name="sSuppressAlign" value="-1" />
-      <param name="sUnmappedFile" value="false" />
-      <param name="sMaxFile" value="false" />
-      <param name="sBest" value="noBest" />
-      <param name="sOffrate" value="-1" />
-      <param name="sSeed" value="-1" />
-      <param name="suppressHeader" value="true" />
-      <output name="output" ftype="sam" file="bowtie_out9.sam" sort="True">
-        <metadata name="dbkey" value="equCab2" />
-      </output>
-    </test>
-    <test>
-      <!--
-      Bowtie command:
-      bowtie-build +offrate 5 +ftabchars 10 +little -f test-data/phiX.fasta phiX_base
-      bowtie -q -X 1000 +ff -p 4 -S +sam-nohead phiX_base -1 test-data/bowtie_in5.fastqsanger -2 test-data/bowtie_in6.fastqsanger > bowtie_out10_u.sam
-      sort bowtie_out10_u.sam > bowtie_out10.sam
-      -p is the number of threads. You need to replace the + with 2 dashes.
-      chrM_base is the index files' location/base name.
-      -->
-      <param name="genomeSource" value="history" />
-      <param name="ownFile" value="phiX.fasta" />
-      <param name="indexSettings" value="indexFull" />
-      <param name="autoB" value="auto" />
-      <param name="nodc" value="dc" />
-      <param name="noref" value="ref" />
-      <param name="offrate" value="5" />
-      <param name="ftab" value="10" />
-      <param name="ntoa" value="no" />
-      <param name="endian" value="little" />
-      <param name="seed" value="-1" />
-      <param name="sPaired" value="paired" />
-      <param name="pInput1" ftype="fastqsanger" value="bowtie_in5.fastqsanger" />
-      <param name="pInput2" ftype="fastqsanger" value="bowtie_in6.fastqsanger" />
-      <param name="pMaxInsert" value="1000" />
-      <param name="pMateOrient" value="ff" />
-      <param name="pSettingsType" value="preSet" />
-      <param name="suppressHeader" value="true" />
-      <output name="output" ftype="sam" file="bowtie_out10.sam" sort="True" />
-    </test>
-    <test>
-      <!--
-      Bowtie command:
-      bowtie-build +offrate 5 +ftabchars 10 +little -f test-data/phiX.fasta phiX_base
-      bowtie -q -X 1000 +ff -p 4 -S +sam-nohead phiX_base -1 test-data/bowtie_in5.fastqsanger -2 test-data/bowtie_in6.fastqsanger > bowtie_out10_u.sam
-      sort bowtie_out10_u.sam > bowtie_out10.sam
-      -p is the number of threads. You need to replace the + with 2 dashes.
-      chrM_base is the index files' location/base name.
-      -->
-      <param name="genomeSource" value="history" />
-      <param name="ownFile" value="phiX.fasta" />
-      <param name="indexSettings" value="indexFull" />
-      <param name="autoB" value="auto" />
-      <param name="nodc" value="dc" />
-      <param name="noref" value="ref" />
-      <param name="offrate" value="5" />
-      <param name="ftab" value="10" />
-      <param name="ntoa" value="no" />
-      <param name="endian" value="little" />
-      <param name="seed" value="-1" />
-      <param name="sPaired" value="paired" />
-      <param name="pInput1" ftype="fastqsanger" value="bowtie_in5.fastqsanger" />
-      <param name="pInput2" ftype="fastqsanger" value="bowtie_in6.fastqsanger" />
-      <param name="pMaxInsert" value="1000" />
-      <param name="pMateOrient" value="ff" />
-      <param name="pSettingsType" value="preSet" />
-      <param name="suppressHeader" value="true" />
-      <param name="save_mapping_stats" value="true" />
-      <output name="output" ftype="sam" file="bowtie_out10.sam" sort="True" />
-      <output name="mapping_stats" ftype="txt" file="bowtie_out11.txt" sort="True" />
+      <param name="sSuppressAlign" value='1' />
+      <param name="alignMode" value='nMode' />
+      <param name="pMismatchSeed" value='1' />
+
+
+      <output name="output" ftype="bam" file="test.bam"/>
     </test>
   </tests>
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bowtie_genome_wrapper/tool-data/bowtie_indexes.loc.sample	Fri May 13 09:31:48 2022 +0000
@@ -0,0 +1,9 @@
+mus_musculus_m28	m28	Mus musculus (m28) Genome	/data2/indices/bowtie/genome/mus_musculus_m28/mus_musculus_m28_genome
+drosophila_melanogaster_dm6	dm6	Drosophila melanogaster (dm6) Genome	/data2/indices/bowtie/genome/drosophila_melanogater_dm6/drosophila_melanogater_dm6_genome
+drosophila_melanogater_dm3  dm3 Drosophila melanogaster (dm3) Genome    /data2/indices/bowtie/genome/drosophila_melanogaster_dm3/drosophila_melanogaster_dm3_genome
+sacCer_9	sacCer9	Saccharomyces cerevisiae (v9) Genome	/data2/indices/bowtie/genome/sacCer_9/sacCer_9_genome
+sarCer_3	sacCer3	Saccharomyces cerevisiae (v3) Genome	/data2/indices/bowtie/genome/sacCer_3/sacCer_3_genome
+E_coli_K12_MG1655	e_coli	Escherichia coli (K12 MG1655) Genome	/data2/indices/bowtie/genome/E_coli_K12_MG1655/E_coli_K12_MG1655_genome
+Arabidopsis	araThal	Arabidopsis thaliana (TAIR10) Genome	/data2/indices/bowtie/genome/Arabidopsis/Arabidopsis_genome
+homo_sapiens_hg19	hg19	Homo sapiens (hg19) Genome	/data2/indices/bowtie/genome/homo_sapiens_hg19/homo_sapiens_hg19_genome
+homo_sapiens_hg38	hg38	Homo sapiens (hg38) Genome	/data2/indices/bowtie/genome/homo_sapiens_hg38/homo_sapiens_hg38_genome
\ No newline at end of file
--- a/bowtie_genome_wrapper/tool-data/bowtie_indices.loc.sample	Fri Apr 22 11:34:02 2022 +0000
+++ /dev/null	Thu Jan 01 00:00:00 1970 +0000
@@ -1,8 +0,0 @@
-homo_sapiens_hg19	hg19	Homo sapiens (hg19)	/data2/indices/bowtie/genome/homo_sapiens_hg19/homo_sapiens_hg19_genome
-homo_sapiens_hg38	hg38	Homo sapiens (hg38)	/data2/indices/bowtie/genome/homo_sapiens_hg38/homo_sapiens_hg38_genome
-mus_musculus_m28	m28     Mus musculus (m28)	/data2/indices/bowtie/genome/mus_musculus_m28/mus_musculus_m28_genome
-drosophila_melanogaster_dm6	dm6 Drosophila melanogaster (dm6)	/data2/indices/bowtie/genome/drosophila_melanogater_dm6/drosophila_melanogater_dm6_genome
-sacCer_9	sacCer9 Saccharomyces cerevisiae (v9)	/data2/indices/bowtie/genome/sacCer_9/sacCer_9_genome
-sarCer_3	sacCer3 Saccharomyces cerevisiae (v3)	/data2/indices/bowtie/genome/sacCer_3/sacCer_3_genome
-E_coli_K12_MG1655	e_coli	Escherichia coli (K12 MG1655)	/data2/indices/bowtie/genome/E_coli_K12_MG1655/E_coli_K12_MG1655_genome
-Arabidopsis	araThal	Arabidopsis thaliana (TAIR10)	/data2/indices/bowtie/genome/Arabidopsis/Arabidopsis_genome
\ No newline at end of file
--- a/bowtie_genome_wrapper/tool_data_table_conf.xml.sample	Fri Apr 22 11:34:02 2022 +0000
+++ b/bowtie_genome_wrapper/tool_data_table_conf.xml.sample	Fri May 13 09:31:48 2022 +0000
@@ -1,8 +1,8 @@
 <!-- Use the file tool_data_table_conf.xml.oldlocstyle if you don't want to update your loc files as changed in revision 4550:535d276c92bc-->
 <tables>
     <!-- Locations of indexes in the Bowtie mapper format -->
-    <table name="bowtie_genome_indexes" comment_char="#">
+    <table name="bowtie_indexes" comment_char="#">
         <columns>value, dbkey, name, path</columns>
-        <file path="tool-data/bowtie_indices.loc.sample" />
+        <file path="tool-data/bowtie_indexes.loc" />
     </table>
 </tables>