Mercurial > repos > iuc > bbtools_bbduk
changeset 9:7beb76ffbc0f draft
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/bbtools commit cd59ba2c349865259b92302a1d70e103b8a5e3cb
| author | iuc |
|---|---|
| date | Tue, 27 Aug 2024 10:15:37 +0000 |
| parents | d1f62024a08b |
| children | d58c27d2c5a7 |
| files | bbduk.xml macros.xml test-data/cv_output.gff test-data/cv_output.txt test-data/cv_output.vcf test-data/cv_qualityhist_output.tabular test-data/cv_scorehist_output.tabular test-data/cv_zygosityhist_output.tabular |
| diffstat | 8 files changed, 72 insertions(+), 69 deletions(-) [+] |
line wrap: on
line diff
--- a/bbduk.xml Sat May 18 19:13:38 2024 +0000 +++ b/bbduk.xml Tue Aug 27 10:15:37 2024 +0000 @@ -154,11 +154,11 @@ </param> <when value="no_reference"/> <when value="files"> - <param name="reference" type="data" format="fasta,fasta.gz" multiple="true" optional="false" label="Select one or more fasta file"/> + <param name="reference" type="data" format="fasta,fasta.gz" optional="false" label="Select one or more fasta file" multiple="true"/> <expand macro="ktrim_cond"/> </when> <when value="keywords"> - <param name="reference" type="select" multiple="true" optional="false" label="Select one or more keywords"> + <param name="reference" type="select" optional="false" label="Select one or more keywords" multiple="true"> <option value="adapters">adapters</option> <option value="artifacts">artifacts</option> <option value="phix">phix</option> @@ -171,34 +171,34 @@ </when> </conditional> <section name="advanced_options" title="Advanced options" expanded="false"> - <param argument="k" type="integer" value="27" min="1" label="Kmer length used for finding contaminants"/> + <param argument="k" type="integer" min="1" value="27" label="Kmer length used for finding contaminants"/> <param argument="rcomp" type="boolean" truevalue="t" falsevalue="f" checked="true" label="Look for reverse-complements of kmers in addition to forward kmers?"/> <param argument="maskmiddle" type="boolean" truevalue="t" falsevalue="f" checked="true" label="Treat the middle base of a kmer as a wildcard to increase sensitivity in the presence of errors?"/> - <param argument="minkmerhits" type="integer" value="1" min="1" label="Reads need at least this many matching kmers to be considered as matching the reference"/> - <param argument="minkmerfraction" type="float" value="0" min="0" label="A read needs at least this fraction of its total kmers to hit a ref in order to be considered a match"/> - <param argument="mincovfraction" type="float" value="0" min="0" label="A read needs at least this fraction of its total bases to be covered by ref kmers to be considered a match"/> - <param argument="hammingdistance" type="integer" value="0" min="0" label="Maximum Hamming distance for ref kmers (subs only)"/> - <param argument="qhdist" type="integer" value="0" min="0" label="Hamming distance for query kmers"/> - <param argument="editdistance" type="integer" value="0" min="0" label="Maximum edit distance from ref kmers (subs and indels)"/> + <param argument="minkmerhits" type="integer" min="1" value="1" label="Reads need at least this many matching kmers to be considered as matching the reference"/> + <param argument="minkmerfraction" type="float" min="0" value="0" label="A read needs at least this fraction of its total kmers to hit a ref in order to be considered a match"/> + <param argument="mincovfraction" type="float" min="0" value="0" label="A read needs at least this fraction of its total bases to be covered by ref kmers to be considered a match"/> + <param argument="hammingdistance" type="integer" min="0" value="0" label="Maximum Hamming distance for ref kmers (subs only)"/> + <param argument="qhdist" type="integer" min="0" value="0" label="Hamming distance for query kmers"/> + <param argument="editdistance" type="integer" min="0" value="0" label="Maximum edit distance from ref kmers (subs and indels)"/> <param argument="forbidn" type="boolean" truevalue="t" falsevalue="f" checked="false" label="Do not match kmers comntaining N?"/> <param argument="trimfailures" type="boolean" truevalue="t" falsevalue="f" checked="false" label="Trim failed reads to 1bp instead of discarding them?"/> <param argument="findbestmatch" type="boolean" truevalue="t" falsevalue="f" checked="false" label="Associate read with sequence sharing most kmers if multiple matches?"/> <param argument="skipr1" type="boolean" truevalue="t" falsevalue="f" checked="false" label="Don't do kmer-based operations on read 1?"/> <param argument="skipr2" type="boolean" truevalue="t" falsevalue="f" checked="false" label="Don't do kmer-based operations on read 2?"/> </section> - <param name="outputs_select" type="select" multiple="true" optional="false" label="Specify outputs"> + <param name="outputs_select" type="select" optional="false" label="Specify outputs" multiple="true"> <option value="outu">Unmatched</option> <option value="outm">Matched</option> <option value="outs">Single</option> </param> <conditional name="output_stats_cond"> - <param name="output_stats" type="select" label="Output statistics?"> + <param name="output_stats" type="select" label="Output statistics?"> <option value="no" selected="true">No</option> <option value="yes">Yes</option> </param> <when value="no"/> <when value="yes"> - <param name="output_stats_select" type="select" multiple="true" optional="false" label="Specify statistics outputs"> + <param name="output_stats_select" type="select" optional="false" label="Specify statistics outputs" multiple="true"> <option value="stats">Statistics about which contamininants were detected</option> <option value="ref">Statistics on a per-reference-file basis</option> <option value="rpkm">RPKM for each reference sequence (for RNA-seq)</option> @@ -207,13 +207,13 @@ </when> </conditional> <conditional name="output_hists_cond"> - <param name="output_hists" type="select" label="Output histograms?"> + <param name="output_hists" type="select" label="Output histograms?"> <option value="no" selected="true">No</option> <option value="yes">Yes</option> </param> <when value="no"/> <when value="yes"> - <param name="output_hists_select" type="select" multiple="true" optional="false" label="Specify statistics outputs"> + <param name="output_hists_select" type="select" optional="false" label="Specify statistics outputs" multiple="true"> <option value="bhist">Base composition histogram by position</option> <option value="quhist">Quality histogram by position</option> <option value="quchist">Count of bases with each quality value</option> @@ -229,7 +229,7 @@ </inputs> <outputs> <data name="outputu" format="fastqsanger" label="${tool.name} on ${on_string} (Forward Unmatched)"> - <filter>str(outputs_select).find('outu') >= 0</filter> + <filter>str(outputs_select).find('outu') >= 0</filter> <filter>'outu' in outputs_select</filter> </data> <data name="outputu2" format="fastqsanger" label="${tool.name} on ${on_string} (Reverse Unmatched)"> @@ -350,4 +350,3 @@ </help> <expand macro="citations"/> </tool> -
--- a/macros.xml Sat May 18 19:13:38 2024 +0000 +++ b/macros.xml Tue Aug 27 10:15:37 2024 +0000 @@ -1,15 +1,19 @@ <macros> - <token name="@TOOL_VERSION@">39.06</token> - <token name="@VERSION_SUFFIX@">2</token> + <token name="@TOOL_VERSION@">39.08</token> + <token name="@VERSION_SUFFIX@">0</token> <token name="@PROFILE@">22.01</token> <xml name="edam_ontology"> <edam_topics> - <edam_topic>topic_0622</edam_topic> <!-- Genomics --> - <edam_topic>topic_0091</edam_topic> <!-- Bioinformatics --> + <edam_topic>topic_0622</edam_topic> + <!-- Genomics --> + <edam_topic>topic_0091</edam_topic> + <!-- Bioinformatics --> </edam_topics> <edam_operations> - <edam_operation>operation_0496</edam_operation> <!-- Global alignment --> - <edam_operation>operation_0491</edam_operation> <!-- Pairwise sequence alignment --> + <edam_operation>operation_0496</edam_operation> + <!-- Global alignment --> + <edam_operation>operation_0491</edam_operation> + <!-- Pairwise sequence alignment --> </edam_operations> </xml> <xml name="bio.tools"> @@ -20,7 +24,8 @@ <xml name="requirements"> <requirements> <requirement type="package" version="@TOOL_VERSION@">bbmap</requirement> - <requirement type="package" version="1.19.2">samtools</requirement> <!-- automatic solving installs 1.6 in some cases, instead --> + <requirement type="package" version="1.20">samtools</requirement> + <!-- automatic solving installs 1.6 in some cases, instead --> </requirements> </xml> <macro name="dbKeyActionsBBMap"> @@ -62,7 +67,7 @@ <param name="read2" type="data" format="fastqsanger.gz,fastqsanger" label="Reverse reads fastq file"/> </when> <when value="paired"> - <param name="reads_collection" type="data_collection" format="fastqsanger,fastqsanger.gz" collection_type="paired" label="Collection of fastqsanger paired read files"/> + <param name="reads_collection" type="data_collection" format="fastqsanger,fastqsanger.gz" label="Collection of fastqsanger paired read files" collection_type="paired"/> </when> </conditional> </macro> @@ -87,18 +92,18 @@ </macro> <macro name="ktrim_cond"> <conditional name="ktrim_cond"> - <param name="ktrim_select" type="select" label="Trim reads to remove bases matching reference kmers?"> + <param name="ktrim_select" type="select" label="Trim reads to remove bases matching reference kmers?"> <option value="no" selected="true">No</option> <option value="yes">Yes</option> </param> <when value="no"/> <when value="yes"> - <param argument="ktrim" type="select" label="Select trimming position"> + <param argument="ktrim" type="select" label="Select trimming position"> <option value="r">Trim to the right</option> <option value="l">Trim to the left</option> </param> <param argument="minlength" type="integer" value="10" label="Minimum read length" help="Trimmed reads shorter than this will be discarded, pairs will be discarded if both are shorter."/> - </when> + </when> </conditional> </macro> <xml name="citations"> @@ -109,4 +114,3 @@ </citations> </xml> </macros> -
--- a/test-data/cv_output.gff Sat May 18 19:13:38 2024 +0000 +++ b/test-data/cv_output.gff Tue Aug 27 10:15:37 2024 +0000 @@ -1,18 +1,18 @@ -##gff-version -#BBMapVersion +##gff-version 3 +#BBMapVersion 39.08 #ploidy 2 #rarity 1.00000 #minAlleleFraction 0.1000 #reads 2207 #pairedReads 0 #properlyPairedReads 0 -#readLengthAvg +#readLengthAvg 145.46 #properPairRate 0.0000 -#totalQualityAvg -#mapqAvg -#reference +#totalQualityAvg 36.5800 +#mapqAvg 43.35 +#reference /home/bag/projects/code/tools-iuc/tools/bbtools/test-data/NC_002945v4.fasta #seqid source type start end score strand phase attributes -NC_002945.4 . sequence_variant_obs 976327 976327 -NC_002945.4 . sequence_variant_obs 1501932 1501932 -NC_002945.4 . sequence_variant_obs 3380579 3380579 -NC_002945.4 . sequence_variant_obs 3646293 3646293 +NC_002945.4 . sequence_variant_obs 976327 976327 0.87 + . ID=SUB T +NC_002945.4 . sequence_variant_obs 1501932 1501932 0.92 + . ID=SUB C +NC_002945.4 . sequence_variant_obs 3380579 3380579 0.92 + . ID=SUB G +NC_002945.4 . sequence_variant_obs 3646293 3646293 0.89 + . ID=SUB A
--- a/test-data/cv_output.txt Sat May 18 19:13:38 2024 +0000 +++ b/test-data/cv_output.txt Tue Aug 27 10:15:37 2024 +0000 @@ -1,18 +1,18 @@ -#fileformat -#BBMapVersion +#fileformat Var_1.3 +#BBMapVersion 39.08 #ploidy 2 #rarity 1.00000 #minAlleleFraction 0.1000 #reads 2207 #pairedReads 0 #properlyPairedReads 0 -#readLengthAvg -#properPairRate -#totalQualityAvg -#mapqAvg -#reference +#readLengthAvg 145.46 +#properPairRate 0.0000 +#totalQualityAvg 36.5800 +#mapqAvg 43.35 +#reference /home/bag/projects/code/tools-iuc/tools/bbtools/test-data/NC_002945v4.fasta #scaf start stop type call r1p r1m r2p r2m paired lengthSum mapq mapqMax baseq baseqMax edist edistMax id idMax cov minusCov nearbyVarCount flagged contigEndDist phredScore readCount alleleFraction revisedAF strandRatio baseqAvg mapqAvg edistAvg identityAvg edistScore identityScore qualityScore pairedScore biasScore coverageScore homopolymerScore score -0 976326 976327 SUB T -0 1501931 1501932 SUB C -0 3380578 3380579 SUB G -0 3646292 3646293 SUB A +0 976326 976327 SUB T 2 0 0 0 0 292 77 40 48 24 111 65 1930 972 0 -1 0 0 0 22.21 2 1 1 0.5000 24 38.50 55.50 965 0.9972 0.9932 0.7489 0.9800 0.9737 0.7143 1 0.8725 +0 1501931 1501932 SUB C 0 2 0 0 0 292 88 44 64 39 74 38 1986 993 0 -1 0 0 0 27.20 2 1 1 0.5000 32 44 37 993 0.9851 0.9994 0.9831 0.9800 0.9737 0.7143 1 0.9202 +0 3380578 3380579 SUB G 0 2 0 0 0 289 87 44 75 38 105 63 1986 993 0 -1 0 0 0 27.85 2 1 1 0.5000 37.50 43.50 52.50 993 0.9968 0.9994 0.9972 0.9800 0.9737 0.7143 1 0.9250 +0 3646292 3646293 SUB A 1 1 0 0 0 291 86 44 76 38 44 31 1986 993 0 -1 0 0 0 23.34 2 1 1 1 38 43 22 993 0.7996 0.9994 0.9975 0.9800 0.9747 0.7143 1 0.8853
--- a/test-data/cv_output.vcf Sat May 18 19:13:38 2024 +0000 +++ b/test-data/cv_output.vcf Tue Aug 27 10:15:37 2024 +0000 @@ -1,16 +1,16 @@ ##fileformat=VCFv4.2 -##BBMapVersion= +##BBMapVersion=39.08 ##ploidy=2 ##rarity=1.00000 ##minallelefraction=0.10000 ##reads=2207 ##pairedReads=0 ##properlyPairedReads=0 -##readLengthAvg= +##readLengthAvg=145.462 ##properPairRate=0.00000 -##totalQualityAvg= +##totalQualityAvg=36.580 ##mapqAvg=43.348 -##reference= +##reference=/home/bag/projects/code/tools-iuc/tools/bbtools/test-data/NC_002945v4.fasta ##contig=<ID=NC_002945.4,length=7000> ##FILTER=<ID=FAIL,Description="Fail"> ##FILTER=<ID=PASS,Description="Pass"> @@ -53,8 +53,8 @@ ##FORMAT=<ID=SB,Number=1,Type=Float,Description="Strand Bias"> ##FORMAT=<ID=SC,Number=1,Type=Float,Description="Score"> ##FORMAT=<ID=PF,Number=1,Type=String,Description="Pass Filter"> -#CHROM POS ID REF ALT QUAL FILTER INFO FORMAT -NC_002945.4 976327 . N T 22.21 PASS SN=0;STA=976326;STO=976327;TYP=SUB;R1P=2;R1M=0;R2P=0;R2M=0;AD=2;DP=2;MCOV=-1;PPC=0;AF=1.0000;RAF=1.0000;LS=292;MQS=77;MQM=40;BQS=48;BQM=24;EDS=111;EDM=65;IDS=1930;IDM=972;NVC=0;FLG=0;CED=0;HMP=0;SB=0.9980 GT:DP:AD:AF:RAF:NVC:FLG:SB:SC:PF 1/1:2:2:1.0000:1.0000:0:0:0.9980:22.21:PASS -NC_002945.4 1501932 . N C 27.20 PASS SN=0;STA=1501931;STO=1501932;TYP=SUB;R1P=0;R1M=2;R2P=0;R2M=0;AD=2;DP=2;MCOV=-1;PPC=0;AF=1.0000;RAF=1.0000;LS=292;MQS=88;MQM=44;BQS=64;BQM=39;EDS=74;EDM=38;IDS=1986;IDM=993;NVC=0;FLG=0;CED=0;HMP=0;SB=0.9980 GT:DP:AD:AF:RAF:NVC:FLG:SB:SC:PF 1/1:2:2:1.0000:1.0000:0:0:0.9980:27.20:PASS -NC_002945.4 3380579 . N G 27.85 PASS SN=0;STA=3380578;STO=3380579;TYP=SUB;R1P=0;R1M=2;R2P=0;R2M=0;AD=2;DP=2;MCOV=-1;PPC=0;AF=1.0000;RAF=1.0000;LS=289;MQS=87;MQM=44;BQS=75;BQM=38;EDS=105;EDM=63;IDS=1986;IDM=993;NVC=0;FLG=0;CED=0;HMP=0;SB=0.9980 GT:DP:AD:AF:RAF:NVC:FLG:SB:SC:PF 1/1:2:2:1.0000:1.0000:0:0:0.9980:27.85:PASS -NC_002945.4 3646293 . N A 23.34 PASS SN=0;STA=3646292;STO=3646293;TYP=SUB;R1P=1;R1M=1;R2P=0;R2M=0;AD=2;DP=2;MCOV=-1;PPC=0;AF=1.0000;RAF=1.0000;LS=291;MQS=86;MQM=44;BQS=76;BQM=38;EDS=44;EDM=31;IDS=1986;IDM=993;NVC=0;FLG=0;CED=0;HMP=0;SB=1.0000 GT:DP:AD:AF:RAF:NVC:FLG:SB:SC:PF 1/1:2:2:1.0000:1.0000:0:0:1.0000:23.34:PASS +#CHROM POS ID REF ALT QUAL FILTER INFO FORMAT dataset_4f1db038-76fe-46bc-a222-dfc4894a8ff2.dat +NC_002945.4 976327 . N T 22.21 PASS SN=0;STA=976326;STO=976327;TYP=SUB;R1P=2;R1M=0;R2P=0;R2M=0;AD=2;DP=2;MCOV=-1;PPC=0;AF=1;RAF=1;LS=292;MQS=77;MQM=40;BQS=48;BQM=24;EDS=111;EDM=65;IDS=1930;IDM=972;NVC=0;FLG=0;CED=0;HMP=0;SB=0.9980 GT:DP:AD:AF:RAF:NVC:FLG:SB:SC:PF 1/1:2:2:1:1:0:0:0.9980:22.21:PASS +NC_002945.4 1501932 . N C 27.20 PASS SN=0;STA=1501931;STO=1501932;TYP=SUB;R1P=0;R1M=2;R2P=0;R2M=0;AD=2;DP=2;MCOV=-1;PPC=0;AF=1;RAF=1;LS=292;MQS=88;MQM=44;BQS=64;BQM=39;EDS=74;EDM=38;IDS=1986;IDM=993;NVC=0;FLG=0;CED=0;HMP=0;SB=0.9980 GT:DP:AD:AF:RAF:NVC:FLG:SB:SC:PF 1/1:2:2:1:1:0:0:0.9980:27.20:PASS +NC_002945.4 3380579 . N G 27.85 PASS SN=0;STA=3380578;STO=3380579;TYP=SUB;R1P=0;R1M=2;R2P=0;R2M=0;AD=2;DP=2;MCOV=-1;PPC=0;AF=1;RAF=1;LS=289;MQS=87;MQM=44;BQS=75;BQM=38;EDS=105;EDM=63;IDS=1986;IDM=993;NVC=0;FLG=0;CED=0;HMP=0;SB=0.9980 GT:DP:AD:AF:RAF:NVC:FLG:SB:SC:PF 1/1:2:2:1:1:0:0:0.9980:27.85:PASS +NC_002945.4 3646293 . N A 23.34 PASS SN=0;STA=3646292;STO=3646293;TYP=SUB;R1P=1;R1M=1;R2P=0;R2M=0;AD=2;DP=2;MCOV=-1;PPC=0;AF=1;RAF=1;LS=291;MQS=86;MQM=44;BQS=76;BQM=38;EDS=44;EDM=31;IDS=1986;IDM=993;NVC=0;FLG=0;CED=0;HMP=0;SB=1 GT:DP:AD:AF:RAF:NVC:FLG:SB:SC:PF 1/1:2:2:1:1:0:0:1:23.34:PASS
--- a/test-data/cv_qualityhist_output.tabular Sat May 18 19:13:38 2024 +0000 +++ b/test-data/cv_qualityhist_output.tabular Tue Aug 27 10:15:37 2024 +0000 @@ -1,8 +1,8 @@ #BaseQualityHist -#Vars -#Mean -#Median -#Mode +#Vars 4 +#Mean 32.75 +#Median 32 +#Mode 32 #Quality AvgCount MaxCount 0 0 0 1 0 0