mirror of
https://github.com/galaxyproject/galaxy.git
synced 2026-10-24 12:13:33 +08:00
Enhance Bowtie wrapper to accept non-sanger variant FASTQ files.
This commit is contained in:
Daniel Blankenberg
@@ -7,7 +7,7 @@
|
||||
--threads="4"
|
||||
## Outputs
|
||||
--output=$output
|
||||
#if $singlePaired == "single"
|
||||
#if str( $singlePaired.sPaired ) == "single"
|
||||
#if $output_unmapped_reads_l
|
||||
--output_unmapped_reads=$output_unmapped_reads_l
|
||||
#end if
|
||||
|
||||
@@ -67,6 +67,11 @@ usage: bowtie_wrapper.py [options]
|
||||
|
||||
import optparse, os, shutil, subprocess, sys, tempfile
|
||||
|
||||
#Allow more than Sanger encoded variants
|
||||
DEFAULT_ASCII_ENCODING = '--phred33-quals'
|
||||
GALAXY_FORMAT_TO_QUALITY_SCORE_ENCODING_ARG = { 'fastqsanger':'--phred33-quals', 'fastqillumina':'--phred64-quals', 'fastqsolexa':'--solexa-quals' }
|
||||
#FIXME: Integer quality scores are supported only when the '--integer-quals' argument is specified to bowtie; this is not currently able to be set in the tool/wrapper/config
|
||||
|
||||
def stop_err( msg ):
|
||||
sys.stderr.write( '%s\n' % msg )
|
||||
sys.exit()
|
||||
@@ -131,6 +136,7 @@ def __main__():
|
||||
parser.add_option( '-c', '--icutoff', dest='icutoff', help='Number of first bases of the reference sequence to index' )
|
||||
parser.add_option( '-x', '--indexSettings', dest='index_settings', help='Whether or not indexing options are to be set' )
|
||||
parser.add_option( '-H', '--suppressHeader', dest='suppressHeader', help='Suppress header' )
|
||||
parser.add_option( '--galaxy_input_format', dest='galaxy_input_format', default="fastqsanger", help='galaxy input format' )
|
||||
parser.add_option( '--do_not_build_index', dest='do_not_build_index', action="store_true", default=False, help='Flag to specify that provided file is already indexed, use as is' )
|
||||
(options, args) = parser.parse_args()
|
||||
stdout = ''
|
||||
@@ -255,9 +261,10 @@ def __main__():
|
||||
mateOrient = '--%s' % options.mateOrient
|
||||
else:
|
||||
mateOrient = ''
|
||||
quality_score_encoding = GALAXY_FORMAT_TO_QUALITY_SCORE_ENCODING_ARG.get( options.galaxy_input_format, DEFAULT_ASCII_ENCODING )
|
||||
if options.params == 'preSet':
|
||||
aligning_cmds = '-q %s %s -p %s -S %s %s ' % \
|
||||
( maxInsert, mateOrient, options.threads, suppressHeader, colorspace )
|
||||
aligning_cmds = '-q %s %s -p %s -S %s %s %s ' % \
|
||||
( maxInsert, mateOrient, options.threads, suppressHeader, colorspace, quality_score_encoding )
|
||||
else:
|
||||
try:
|
||||
if options.skip and int( options.skip ) > 0:
|
||||
@@ -386,14 +393,15 @@ def __main__():
|
||||
else:
|
||||
keepends = ''
|
||||
aligning_cmds = '-q %s %s -p %s -S %s %s %s %s %s %s %s %s %s %s %s %s ' \
|
||||
'%s %s %s %s %s %s %s %s %s %s %s %s %s %s %s %s %s' % \
|
||||
'%s %s %s %s %s %s %s %s %s %s %s %s %s %s %s %s %s %s ' % \
|
||||
( maxInsert, mateOrient, options.threads, suppressHeader,
|
||||
colorspace, skip, alignLimit, trimH, trimL, maqSoapAlign,
|
||||
mismatchSeed, mismatchQual, seedLen, rounding, minInsert,
|
||||
maxAlignAttempt, forwardAlign, reverseAlign, maxBacktracks,
|
||||
tryHard, valAlign, allValAligns, suppressAlign, best,
|
||||
strata, offrate, seed, snpphred, snpfrac, keepends,
|
||||
output_unmapped_reads, output_suppressed_reads )
|
||||
output_unmapped_reads, output_suppressed_reads,
|
||||
quality_score_encoding )
|
||||
except ValueError, e:
|
||||
# clean up temp dir
|
||||
if os.path.exists( tmp_index_dir ):
|
||||
|
||||
@@ -8,13 +8,14 @@
|
||||
--threads="4"
|
||||
## Outputs
|
||||
--output=$output
|
||||
#if $singlePaired == "single"
|
||||
#if str( $singlePaired.sPaired ) == "single"
|
||||
#if $output_unmapped_reads_l
|
||||
--output_unmapped_reads=$output_unmapped_reads_l
|
||||
#end if
|
||||
#if $output_suppressed_reads_l
|
||||
--output_suppressed_reads=$output_suppressed_reads_l
|
||||
#end if
|
||||
--galaxy_input_format="${singlePaired.sInput1.ext}"
|
||||
#else
|
||||
#if $output_unmapped_reads_l and $output_unmapped_reads_r
|
||||
--output_unmapped_reads_l=$output_unmapped_reads_l
|
||||
@@ -24,9 +25,10 @@
|
||||
--output_suppressed_reads_l=$output_suppressed_reads_l
|
||||
--output_suppressed_reads_r=$output_suppressed_reads_r
|
||||
#end if
|
||||
--galaxy_input_format="${singlePaired.pInput1.ext}"
|
||||
#end if
|
||||
## Inputs
|
||||
--dataType="solexa"
|
||||
--dataType="solexa" ##this indicates that nucleotide base space is used in the wrapper
|
||||
--suppressHeader=$suppressHeader
|
||||
--genomeSource=$refGenomeSource.genomeSource
|
||||
#if $refGenomeSource.genomeSource == "history":
|
||||
@@ -59,7 +61,8 @@
|
||||
#end if
|
||||
#else
|
||||
##use pre-built index
|
||||
--ref="${ filter( lambda x: str( x[0] ) == str( $refGenomeSource.index ), $__app__.tool_data_tables[ 'bowtie_indexes' ].get_fields() )[0][-1] }"
|
||||
##--ref="${ filter( lambda x: str( x[0] ) == str( $refGenomeSource.index ), $__app__.tool_data_tables[ 'bowtie_indexes' ].get_fields() )[0][-1] }"
|
||||
--ref="${ refGenomeSource.index.fields.path }"
|
||||
#end if
|
||||
--paired=$singlePaired.sPaired
|
||||
#if $singlePaired.sPaired == "single":
|
||||
@@ -191,7 +194,7 @@
|
||||
<option value="paired">Paired-end</option>
|
||||
</param>
|
||||
<when value="single">
|
||||
<param name="sInput1" type="data" format="fastqsanger" label="FASTQ file" help="Must have Sanger-scaled quality values with ASCII offset 33"/>
|
||||
<param name="sInput1" type="data" format="fastqsanger,fastqillumina,fastqsolexa" label="FASTQ file" help="Must have ASCII encoded quality scores"/>
|
||||
<conditional name="sParams">
|
||||
<param name="sSettingsType" type="select" label="Bowtie settings to use" help="For most mapping needs use Commonly used settings. If you want full control use Full parameter list">
|
||||
<option value="preSet">Commonly used</option>
|
||||
@@ -245,8 +248,14 @@
|
||||
</conditional> <!-- sParams -->
|
||||
</when> <!-- single -->
|
||||
<when value="paired">
|
||||
<param name="pInput1" type="data" format="fastqsanger" label="Forward FASTQ file" help="Must have Sanger-scaled quality values with ASCII offset 33"/>
|
||||
<param name="pInput2" type="data" format="fastqsanger" label="Reverse FASTQ file" help="Must have Sanger-scaled quality values with ASCII offset 33"/>
|
||||
<param name="pInput1" type="data" format="fastqsanger,fastqillumina,fastqsolexa" label="Forward FASTQ file" help="Must have ASCII encoded quality scores"/>
|
||||
<param name="pInput2" type="data" format="fastqsanger,fastqillumina,fastqsolexa" label="Reverse FASTQ file" help="File format must match the Forward FASTQ file">
|
||||
<options options_filter_attribute="ext" from_parameter="tool.app.datatypes_registry.datatypes_by_extension" transform_lines="obj.keys()">>
|
||||
<column name="name" index="0"/>
|
||||
<column name="value" index="0"/>
|
||||
<filter type="param_value" ref="pInput1" ref_attribute="ext" column="0"/>
|
||||
</options>
|
||||
</param>
|
||||
<param name="pMaxInsert" type="integer" value="1000" label="Maximum insert size for valid paired-end alignments (-X)" />
|
||||
<param name="pMateOrient" type="select" label="The upstream/downstream mate orientation for valid paired-end alignment against the forward reference strand (--fr/--rf/--ff)">
|
||||
<option value="fr">FR (for Illumina)</option>
|
||||
@@ -333,7 +342,7 @@
|
||||
</conditional>
|
||||
</actions>
|
||||
</data>
|
||||
<data format="fastqsanger" name="output_suppressed_reads_l" label="${tool.name} on ${on_string}: suppressed reads (L)">
|
||||
<data format="fastq" name="output_suppressed_reads_l" label="${tool.name} on ${on_string}: suppressed reads (L)">
|
||||
<filter>((
|
||||
singlePaired['sPaired'] == "single" and
|
||||
singlePaired['sParams']['sSettingsType'] == "full" and
|
||||
@@ -344,13 +353,41 @@
|
||||
singlePaired['pParams']['pMaxFile'] is True
|
||||
))
|
||||
</filter>
|
||||
<actions>
|
||||
<conditional name="singlePaired.sPaired">
|
||||
<when value="single">
|
||||
<action type="format">
|
||||
<option type="from_param" name="singlePaired.sInput1" param_attribute="ext" />
|
||||
</action>
|
||||
</when>
|
||||
<when value="paired">
|
||||
<action type="format">
|
||||
<option type="from_param" name="singlePaired.pInput1" param_attribute="ext" />
|
||||
</action>
|
||||
</when>
|
||||
</conditional>
|
||||
</actions>
|
||||
</data>
|
||||
<data format="fastqsanger" name="output_suppressed_reads_r" label="${tool.name} on ${on_string}: suppressed reads (R)">
|
||||
<data format="fastq" name="output_suppressed_reads_r" label="${tool.name} on ${on_string}: suppressed reads (R)">
|
||||
<filter>singlePaired['sPaired'] == "paired"</filter>
|
||||
<filter>singlePaired['pParams']['pSettingsType'] == "full"</filter>
|
||||
<filter>singlePaired['pParams']['pMaxFile'] is True</filter>
|
||||
<actions>
|
||||
<conditional name="singlePaired.sPaired">
|
||||
<when value="single">
|
||||
<action type="format">
|
||||
<option type="from_param" name="singlePaired.sInput1" param_attribute="ext" />
|
||||
</action>
|
||||
</when>
|
||||
<when value="paired">
|
||||
<action type="format">
|
||||
<option type="from_param" name="singlePaired.pInput1" param_attribute="ext" />
|
||||
</action>
|
||||
</when>
|
||||
</conditional>
|
||||
</actions>
|
||||
</data>
|
||||
<data format="fastqsanger" name="output_unmapped_reads_l" label="${tool.name} on ${on_string}: unmapped reads (L)">
|
||||
<data format="fastq" name="output_unmapped_reads_l" label="${tool.name} on ${on_string}: unmapped reads (L)">
|
||||
<filter>
|
||||
((
|
||||
singlePaired['sPaired'] == "single" and
|
||||
@@ -362,11 +399,39 @@
|
||||
singlePaired['pParams']['pUnmappedFile'] is True
|
||||
))
|
||||
</filter>
|
||||
<actions>
|
||||
<conditional name="singlePaired.sPaired">
|
||||
<when value="single">
|
||||
<action type="format">
|
||||
<option type="from_param" name="singlePaired.sInput1" param_attribute="ext" />
|
||||
</action>
|
||||
</when>
|
||||
<when value="paired">
|
||||
<action type="format">
|
||||
<option type="from_param" name="singlePaired.pInput1" param_attribute="ext" />
|
||||
</action>
|
||||
</when>
|
||||
</conditional>
|
||||
</actions>
|
||||
</data>
|
||||
<data format="fastqsanger" name="output_unmapped_reads_r" label="${tool.name} on ${on_string}: unmapped reads (R)">
|
||||
<data format="fastq" name="output_unmapped_reads_r" label="${tool.name} on ${on_string}: unmapped reads (R)">
|
||||
<filter>singlePaired['sPaired'] == "paired"</filter>
|
||||
<filter>singlePaired['pParams']['pSettingsType'] == "full"</filter>
|
||||
<filter>singlePaired['pParams']['pUnmappedFile'] is True</filter>
|
||||
<actions>
|
||||
<conditional name="singlePaired.sPaired">
|
||||
<when value="single">
|
||||
<action type="format">
|
||||
<option type="from_param" name="singlePaired.sInput1" param_attribute="ext" />
|
||||
</action>
|
||||
</when>
|
||||
<when value="paired">
|
||||
<action type="format">
|
||||
<option type="from_param" name="singlePaired.pInput1" param_attribute="ext" />
|
||||
</action>
|
||||
</when>
|
||||
</conditional>
|
||||
</actions>
|
||||
</data>
|
||||
</outputs>
|
||||
<tests>
|
||||
@@ -379,6 +444,7 @@
|
||||
chrM_base needs to be the base location/name of the index files.
|
||||
-->
|
||||
<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" />
|
||||
@@ -437,6 +503,86 @@
|
||||
<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="pMismatchSeed" value="2" />
|
||||
<param name="pMismatchQual" value="70" />
|
||||
<param name="pSeedLen" value="28" />
|
||||
<param name="pRounding" value="round" />
|
||||
<param name="pMaqSoapAlign" value="-1" />
|
||||
<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="fastqillumna" file="bowtie_out8_1.fastqillumina.sorted" sort="True" />
|
||||
<output name="output_unmapped_reads_r" ftype="fastqillumna" 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="pMismatchSeed" value="2" />
|
||||
<param name="pMismatchQual" value="70" />
|
||||
<param name="pSeedLen" value="28" />
|
||||
<param name="pRounding" value="round" />
|
||||
<param name="pMaqSoapAlign" value="-1" />
|
||||
<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:
|
||||
|
||||
Reference in New Issue
Block a user