Mercurial > repos > iuc > busco

changeset 12:1278204240d8 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload commit d00431b92294849bd41d369d8a699e4669e5fc19"
author iuc
date Tue, 02 Nov 2021 09:40:29 +0000
parents 0d243f458b53
children 46ae58b1d792
files busco.xml macros.xml test-data/genome_results_metaeuk_auto/full_table test-data/genome_results_metaeuk_auto/missing_buscos_list test-data/genome_results_metaeuk_auto/short_summary test-data/genome_results_metaeuk_auto/summary.png
diffstat 6 files changed, 611 insertions(+), 25 deletions(-) [+]
line wrap: on
line diff
--- a/busco.xml	Mon Jul 26 21:29:40 2021 +0000
+++ b/busco.xml	Tue Nov 02 09:40:29 2021 +0000
@@ -1,5 +1,8 @@
 <tool id="busco" name="Busco" profile="20.01" version="@TOOL_VERSION@+galaxy@VERSION_SUFFIX@">
     <description>assess genome assembly and annotation completeness</description>
+    <xrefs>
+        <xref type="bio.tools">busco</xref>
+    </xrefs>
     <macros>
         <import>macros.xml</import>
     </macros>
@@ -22,7 +25,6 @@
 
 busco
 --in '${input}'
---lineage_dataset '${lineage_dataset}'
 --update-data
 --mode '${busco_mode.mode}'
 --out busco_galaxy
@@ -30,8 +32,10 @@
 --evalue ${adv.evalue}
 --limit ${adv.limit}
 
-#if $adv.auto_lineage:
-    $adv.auto_lineage
+#if $lineage.lineage_mode == "auto_detect":
+    $lineage.auto_lineage
+#else if $lineage.lineage_mode == "select_lineage":
+    --lineage_dataset '${lineage.lineage_dataset}'
 #end if
 #if $busco_mode.mode == 'geno' and $busco_mode.use_augustus.use_augustus_selector == 'yes':
 
@@ -96,18 +100,28 @@
             <when value="prot" />
         </conditional>
 
-        <param argument="--lineage_dataset" type="select" label="Lineage">
-            <expand macro="lineages" />
-        </param>
+        <conditional name="lineage">
+            <param name="lineage_mode" type="select" label="Auto-detect or select lineage?" help="Let BUSCO decide the best lineage automatically, or select from known lineage">
+                <option value="auto_detect">Auto-detect</option>
+                <option value="select_lineage">Select lineage</option>
+            </param>
+            <when value="auto_detect">
+                <param name="auto_lineage" type="select" label="auto-lineage group" help="Taxonomic group to run with auto-lineage.">
+                    <option value="--auto-lineage">All taxonomic groups (--auto-lineage)</option>
+                    <option value="--auto-lineage-prok">Prokaryotes (--auto-lineage-prok)</option>
+                    <option value="--auto-lineage-euk">Eukaryotes (--auto-lineage-euk)</option>
+                </param>
+            </when>
+            <when value="select_lineage">
+                <param argument="--lineage_dataset" type="select" label="Lineage">
+                    <expand macro="lineages" />
+                </param>
+            </when>
+        </conditional>
 
         <section name="adv" title="Advanced Options" expanded="False">
             <param argument="--evalue" type="float" value="0.001" min="0" max="1" label="E-value cutoff for BLAST searches." />
             <param argument="--limit" type="integer" value="3" label="How many candidate regions to consider" />
-            <param name="auto_lineage" type="select" optional="true" label="Run auto-lineage to find optimal lineage path">
-                <option value="--auto-lineage">Run auto-lineage to find optimum lineage path</option>
-                <option value="--auto-lineage-prok">Run auto-lineage just on non-eukaryote trees to find optimum lineage path</option>
-                <option value="--auto-lineage-euk">Run auto-placement just on eukaryote tree to find optimum lineage path</option>
-            </param>
             <param name="outputs" type="select" optional="true" multiple="true" label="Which outputs should be generated">
                 <option value="short_summary">short summary text</option>
                 <option value="missing">list with missing IDs</option>
@@ -130,7 +144,10 @@
     <tests>
         <test expect_num_outputs="3">
             <param name="input" value="genome.fa" />
-            <param name="lineage_dataset" value="arthropoda_odb10" />
+            <conditional name="lineage">
+                <param name="lineage_mode" value="select_lineage" />
+                <param name="lineage_dataset" value="arthropoda_odb10" />
+            </conditional>
             <conditional name="busco_mode">
                 <param name="mode" value="geno" />
                 <conditional name="use_augustus">
@@ -150,7 +167,10 @@
         </test>
         <test expect_num_outputs="4">
             <param name="input" value="proteome.fa" />
-            <param name="lineage_dataset" value="arthropoda_odb10" />
+            <conditional name="lineage">
+                <param name="lineage_mode" value="select_lineage" />
+                <param name="lineage_dataset" value="arthropoda_odb10" />
+            </conditional>
             <conditional name="busco_mode">
                 <param name="mode" value="prot" />
             </conditional>
@@ -164,22 +184,27 @@
         </test>
         <test expect_num_outputs="4">
             <param name="input" value="transcriptome.fa" />
-            <param name="lineage_dataset" value="arthropoda_odb10" />
+            <conditional name="lineage">
+                <param name="lineage_mode" value="select_lineage" />
+                <param name="lineage_dataset" value="arthropoda_odb10" />
+            </conditional>
             <conditional name="busco_mode">
                 <param name="mode" value="tran" />
             </conditional>
             <section name="adv">
-                <param name="auto_lineage" value="--auto-lineage" />
                 <param name="outputs" value="short_summary,missing,image" />
             </section>
-            <output name="busco_sum" file="transcriptome_results/short_summary" compare="diff" lines_diff="4" />
-            <output name="busco_table" file="transcriptome_results/full_table" compare="diff" lines_diff="4" />
-            <output name="busco_missing" file="transcriptome_results/missing_buscos_list" compare="diff" lines_diff="4" />
+            <output name="busco_sum" file="transcriptome_results/short_summary" compare="diff" lines_diff="6" />
+            <output name="busco_table" file="transcriptome_results/full_table" compare="diff" lines_diff="6" />
+            <output name="busco_missing" file="transcriptome_results/missing_buscos_list" compare="diff" lines_diff="6" />
             <output name="summary_image" file="transcriptome_results/summary.png" compare="sim_size" />
         </test>
         <test expect_num_outputs="2">
             <param name="input" value="genome.fa" />
-            <param name="lineage_dataset" value="arthropoda_odb10" />
+            <conditional name="lineage">
+                <param name="lineage_mode" value="select_lineage" />
+                <param name="lineage_dataset" value="arthropoda_odb10" />
+            </conditional>
             <conditional name="busco_mode">
                 <param name="mode" value="geno" />
                 <conditional name="use_augustus">
@@ -198,7 +223,10 @@
         </test>
         <test expect_num_outputs="3">
             <param name="input" value="genome.fa" />
-            <param name="lineage_dataset" value="arthropoda_odb10" />
+            <conditional name="lineage">
+                <param name="lineage_mode" value="select_lineage" />
+                <param name="lineage_dataset" value="arthropoda_odb10" />
+            </conditional>
             <conditional name="busco_mode">
                 <param name="mode" value="geno" />
                 <conditional name="use_augustus">
@@ -218,7 +246,10 @@
         </test>
         <test expect_num_outputs="4">
             <param name="input" value="genome.fa" />
-            <param name="lineage_dataset" value="arthropoda_odb10" />
+            <conditional name="lineage">
+                <param name="lineage_mode" value="select_lineage" />
+                <param name="lineage_dataset" value="arthropoda_odb10" />
+            </conditional>
             <conditional name="busco_mode">
                 <param name="mode" value="geno" />
                 <conditional name="use_augustus">
@@ -228,11 +259,31 @@
             <section name="adv">
                 <param name="outputs" value="short_summary,missing,image" />
             </section>
-            <output name="busco_sum" file="genome_results_metaeuk/short_summary" compare="diff" lines_diff="4" />
-            <output name="busco_table" file="genome_results_metaeuk/full_table" compare="diff" lines_diff="4" />
-            <output name="busco_missing" file="genome_results_metaeuk/missing_buscos_list" compare="diff" lines_diff="4" />
+            <output name="busco_sum" file="genome_results_metaeuk/short_summary" compare="diff" lines_diff="6" />
+            <output name="busco_table" file="genome_results_metaeuk/full_table" compare="diff" lines_diff="6" />
+            <output name="busco_missing" file="genome_results_metaeuk/missing_buscos_list" compare="diff" lines_diff="6" />
             <output name="summary_image" file="genome_results_metaeuk/summary.png" compare="sim_size" />
         </test>
+        <test expect_num_outputs="4">
+            <param name="input" value="genome.fa" />
+            <conditional name="lineage">
+                <param name="lineage_mode" value="auto_detect" />
+                <param name="auto_lineage" value="--auto-lineage" />
+            </conditional>
+            <conditional name="busco_mode">
+                <param name="mode" value="geno" />
+                <conditional name="use_augustus">
+                    <param name="use_augustus_selector" value="no" />
+                </conditional>
+            </conditional>
+            <section name="adv">
+                <param name="outputs" value="short_summary,missing,image" />
+            </section>
+            <output name="busco_sum" file="genome_results_metaeuk_auto/short_summary" compare="diff" lines_diff="4" />
+            <output name="busco_table" file="genome_results_metaeuk_auto/full_table" compare="diff" lines_diff="0" />
+            <output name="busco_missing" file="genome_results_metaeuk_auto/missing_buscos_list" compare="diff" lines_diff="0" />
+            <output name="summary_image" file="genome_results_metaeuk_auto/summary.png" compare="sim_size" />
+        </test>
     </tests>
     <help><![CDATA[
 
--- a/macros.xml	Mon Jul 26 21:29:40 2021 +0000
+++ b/macros.xml	Tue Nov 02 09:40:29 2021 +0000
@@ -1,7 +1,7 @@
 <?xml version="1.0"?>
 <macros>
     <token name="@TOOL_VERSION@">5.2.2</token>
-    <token name="@VERSION_SUFFIX@">0</token>
+    <token name="@VERSION_SUFFIX@">1</token>
 
     <xml name="citations">
         <citations>
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/genome_results_metaeuk_auto/full_table	Tue Nov 02 09:40:29 2021 +0000
@@ -0,0 +1,258 @@
+# BUSCO version is: 5.2.2 
+# The lineage dataset is: eukaryota_odb10 (Creation date: 2020-09-10, number of genomes: 70, number of BUSCOs: 255)
+# Busco id	Status	Sequence	Gene Start	Gene End	Strand	Score	Length
+39650at2759	Missing
+83779at2759	Missing
+87842at2759	Missing
+97116at2759	Missing
+100698at2759	Missing
+142542at2759	Missing
+156083at2759	Missing
+160593at2759	Missing
+166920at2759	Missing
+176625at2759	Missing
+179362at2759	Missing
+245208at2759	Missing
+257318at2759	Missing
+261328at2759	Missing
+261419at2759	Missing
+270107at2759	Missing
+271586at2759	Missing
+290630at2759	Missing
+293315at2759	Missing
+296129at2759	Missing
+299766at2759	Missing
+306227at2759	Missing
+320059at2759	Missing
+324863at2759	Missing
+325552at2759	Missing
+330169at2759	Missing
+331411at2759	Missing
+341721at2759	Missing
+345441at2759	Missing
+355408at2759	Missing
+369837at2759	Missing
+375960at2759	Missing
+383503at2759	Missing
+388820at2759	Missing
+390348at2759	Missing
+392369at2759	Missing
+396755at2759	Missing
+418107at2759	Missing
+426305at2759	Missing
+450058at2759	Missing
+453044at2759	Missing
+457861at2759	Missing
+464990at2759	Missing
+491869at2759	Missing
+513979at2759	Missing
+543764at2759	Missing
+549762at2759	Missing
+551907at2759	Missing
+570797at2759	Missing
+582756at2759	Missing
+598949at2759	Missing
+604979at2759	Missing
+621827at2759	Missing
+625387at2759	Missing
+655400at2759	Missing
+664730at2759	Missing
+671536at2759	Missing
+671846at2759	Missing
+673132at2759	Missing
+674160at2759	Missing
+674169at2759	Missing
+679187at2759	Missing
+679771at2759	Missing
+687505at2759	Missing
+692986at2759	Missing
+708105at2759	Missing
+719531at2759	Missing
+720340at2759	Missing
+721605at2759	Missing
+722805at2759	Missing
+734341at2759	Missing
+734666at2759	Missing
+736068at2759	Missing
+751335at2759	Missing
+759498at2759	Missing
+761109at2759	Missing
+768809at2759	Missing
+774318at2759	Missing
+777920at2759	Missing
+779909at2759	Missing
+801857at2759	Missing
+814241at2759	Missing
+817008at2759	Missing
+834694at2759	Missing
+836599at2759	Missing
+844748at2759	Missing
+855834at2759	Missing
+858842at2759	Missing
+865202at2759	Missing
+866359at2759	Missing
+869548at2759	Missing
+878143at2759	Missing
+887370at2759	Missing
+891751at2759	Missing
+898782at2759	Missing
+901894at2759	Missing
+905026at2759	Missing
+911863at2759	Missing
+917326at2759	Missing
+918816at2759	Missing
+919955at2759	Missing
+923657at2759	Missing
+924753at2759	Missing
+931188at2759	Missing
+937275at2759	Missing
+937686at2759	Missing
+939345at2759	Missing
+944899at2759	Missing
+946128at2759	Missing
+956854at2759	Missing
+973442at2759	Missing
+974865at2759	Missing
+975158at2759	Missing
+975557at2759	Missing
+976469at2759	Missing
+981902at2759	Missing
+996662at2759	Missing
+1001705at2759	Missing
+1003258at2759	Missing
+1010730at2759	Missing
+1014314at2759	Missing
+1018517at2759	Missing
+1019762at2759	Missing
+1025450at2759	Missing
+1030907at2759	Missing
+1032689at2759	Missing
+1038775at2759	Missing
+1041560at2759	Missing
+1049599at2759	Missing
+1051021at2759	Missing
+1053181at2759	Complete	sample	34845	35678	-	45.0	149
+1057950at2759	Missing
+1065019at2759	Missing
+1076134at2759	Missing
+1079130at2759	Missing
+1079827at2759	Missing
+1085752at2759	Missing
+1087488at2759	Missing
+1090038at2759	Missing
+1094121at2759	Missing
+1096688at2759	Missing
+1106766at2759	Missing
+1107630at2759	Missing
+1108845at2759	Missing
+1111142at2759	Missing
+1112002at2759	Missing
+1115196at2759	Missing
+1128607at2759	Missing
+1129824at2759	Missing
+1138059at2759	Missing
+1157302at2759	Missing
+1161199at2759	Missing
+1173229at2759	Missing
+1178688at2759	Missing
+1182451at2759	Missing
+1193442at2759	Missing
+1194691at2759	Missing
+1194797at2759	Missing
+1197019at2759	Missing
+1200489at2759	Missing
+1217666at2759	Missing
+1220881at2759	Missing
+1222562at2759	Missing
+1223488at2759	Missing
+1228942at2759	Missing
+1233814at2759	Missing
+1236198at2759	Missing
+1247641at2759	Missing
+1248958at2759	Missing
+1249159at2759	Missing
+1251252at2759	Missing
+1257440at2759	Missing
+1258856at2759	Missing
+1259741at2759	Missing
+1260807at2759	Missing
+1264469at2759	Missing
+1266231at2759	Missing
+1269244at2759	Missing
+1269649at2759	Missing
+1275837at2759	Missing
+1284731at2759	Missing
+1287094at2759	Missing
+1287401at2759	Missing
+1291729at2759	Missing
+1304061at2759	Missing
+1309031at2759	Missing
+1312453at2759	Missing
+1314980at2759	Missing
+1322299at2759	Missing
+1322642at2759	Missing
+1323575at2759	Missing
+1324510at2759	Missing
+1338131at2759	Missing
+1339553at2759	Missing
+1342242at2759	Missing
+1346165at2759	Missing
+1346432at2759	Missing
+1348942at2759	Missing
+1355894at2759	Missing
+1358374at2759	Missing
+1364586at2759	Missing
+1370285at2759	Missing
+1370304at2759	Missing
+1377237at2759	Missing
+1379600at2759	Missing
+1380409at2759	Missing
+1382842at2759	Missing
+1395649at2759	Missing
+1398309at2759	Missing
+1404162at2759	Missing
+1405073at2759	Missing
+1405146at2759	Missing
+1407446at2759	Missing
+1421503at2759	Missing
+1423847at2759	Missing
+1426075at2759	Missing
+1428265at2759	Missing
+1428854at2759	Missing
+1430056at2759	Missing
+1432328at2759	Missing
+1434266at2759	Missing
+1440961at2759	Missing
+1441854at2759	Missing
+1442062at2759	Missing
+1445102at2759	Missing
+1450538at2759	Missing
+1454155at2759	Missing
+1455730at2759	Missing
+1459797at2759	Missing
+1474000at2759	Missing
+1479417at2759	Missing
+1488235at2759	Missing
+1488436at2759	Missing
+1490022at2759	Missing
+1504863at2759	Missing
+1513531at2759	Missing
+1525971at2759	Missing
+1526152at2759	Missing
+1530008at2759	Missing
+1538526at2759	Missing
+1545004at2759	Missing
+1552706at2759	Missing
+1558822at2759	Missing
+1563319at2759	Missing
+1567796at2759	Missing
+1575179at2759	Missing
+1576404at2759	Missing
+1588798at2759	Missing
+1593937at2759	Missing
+1617752at2759	Missing
+1620056at2759	Missing
+1623701at2759	Missing
+1626636at2759	Missing
+1633672at2759	Missing
+1645187at2759	Missing
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/genome_results_metaeuk_auto/missing_buscos_list	Tue Nov 02 09:40:29 2021 +0000
@@ -0,0 +1,257 @@
+# BUSCO version is: 5.2.2 
+# The lineage dataset is: eukaryota_odb10 (Creation date: 2020-09-10, number of genomes: 70, number of BUSCOs: 255)
+# Busco id
+1001705at2759
+1003258at2759
+100698at2759
+1010730at2759
+1014314at2759
+1018517at2759
+1019762at2759
+1025450at2759
+1030907at2759
+1032689at2759
+1038775at2759
+1041560at2759
+1049599at2759
+1051021at2759
+1057950at2759
+1065019at2759
+1076134at2759
+1079130at2759
+1079827at2759
+1085752at2759
+1087488at2759
+1090038at2759
+1094121at2759
+1096688at2759
+1106766at2759
+1107630at2759
+1108845at2759
+1111142at2759
+1112002at2759
+1115196at2759
+1128607at2759
+1129824at2759
+1138059at2759
+1157302at2759
+1161199at2759
+1173229at2759
+1178688at2759
+1182451at2759
+1193442at2759
+1194691at2759
+1194797at2759
+1197019at2759
+1200489at2759
+1217666at2759
+1220881at2759
+1222562at2759
+1223488at2759
+1228942at2759
+1233814at2759
+1236198at2759
+1247641at2759
+1248958at2759
+1249159at2759
+1251252at2759
+1257440at2759
+1258856at2759
+1259741at2759
+1260807at2759
+1264469at2759
+1266231at2759
+1269244at2759
+1269649at2759
+1275837at2759
+1284731at2759
+1287094at2759
+1287401at2759
+1291729at2759
+1304061at2759
+1309031at2759
+1312453at2759
+1314980at2759
+1322299at2759
+1322642at2759
+1323575at2759
+1324510at2759
+1338131at2759
+1339553at2759
+1342242at2759
+1346165at2759
+1346432at2759
+1348942at2759
+1355894at2759
+1358374at2759
+1364586at2759
+1370285at2759
+1370304at2759
+1377237at2759
+1379600at2759
+1380409at2759
+1382842at2759
+1395649at2759
+1398309at2759
+1404162at2759
+1405073at2759
+1405146at2759
+1407446at2759
+1421503at2759
+1423847at2759
+142542at2759
+1426075at2759
+1428265at2759
+1428854at2759
+1430056at2759
+1432328at2759
+1434266at2759
+1440961at2759
+1441854at2759
+1442062at2759
+1445102at2759
+1450538at2759
+1454155at2759
+1455730at2759
+1459797at2759
+1474000at2759
+1479417at2759
+1488235at2759
+1488436at2759
+1490022at2759
+1504863at2759
+1513531at2759
+1525971at2759
+1526152at2759
+1530008at2759
+1538526at2759
+1545004at2759
+1552706at2759
+1558822at2759
+156083at2759
+1563319at2759
+1567796at2759
+1575179at2759
+1576404at2759
+1588798at2759
+1593937at2759
+160593at2759
+1617752at2759
+1620056at2759
+1623701at2759
+1626636at2759
+1633672at2759
+1645187at2759
+166920at2759
+176625at2759
+179362at2759
+245208at2759
+257318at2759
+261328at2759
+261419at2759
+270107at2759
+271586at2759
+290630at2759
+293315at2759
+296129at2759
+299766at2759
+306227at2759
+320059at2759
+324863at2759
+325552at2759
+330169at2759
+331411at2759
+341721at2759
+345441at2759
+355408at2759
+369837at2759
+375960at2759
+383503at2759
+388820at2759
+390348at2759
+392369at2759
+39650at2759
+396755at2759
+418107at2759
+426305at2759
+450058at2759
+453044at2759
+457861at2759
+464990at2759
+491869at2759
+513979at2759
+543764at2759
+549762at2759
+551907at2759
+570797at2759
+582756at2759
+598949at2759
+604979at2759
+621827at2759
+625387at2759
+655400at2759
+664730at2759
+671536at2759
+671846at2759
+673132at2759
+674160at2759
+674169at2759
+679187at2759
+679771at2759
+687505at2759
+692986at2759
+708105at2759
+719531at2759
+720340at2759
+721605at2759
+722805at2759
+734341at2759
+734666at2759
+736068at2759
+751335at2759
+759498at2759
+761109at2759
+768809at2759
+774318at2759
+777920at2759
+779909at2759
+801857at2759
+814241at2759
+817008at2759
+834694at2759
+836599at2759
+83779at2759
+844748at2759
+855834at2759
+858842at2759
+865202at2759
+866359at2759
+869548at2759
+878143at2759
+87842at2759
+887370at2759
+891751at2759
+898782at2759
+901894at2759
+905026at2759
+911863at2759
+917326at2759
+918816at2759
+919955at2759
+923657at2759
+924753at2759
+931188at2759
+937275at2759
+937686at2759
+939345at2759
+944899at2759
+946128at2759
+956854at2759
+97116at2759
+973442at2759
+974865at2759
+975158at2759
+975557at2759
+976469at2759
+981902at2759
+996662at2759
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/genome_results_metaeuk_auto/short_summary	Tue Nov 02 09:40:29 2021 +0000
@@ -0,0 +1,20 @@
+# BUSCO version is: 5.2.2 
+# The lineage dataset is: eukaryota_odb10 (Creation date: 2020-09-10, number of genomes: 70, number of BUSCOs: 255)
+# Summarized benchmarking in BUSCO notation for file /tmp/tmp5e3tdhzs/files/c/9/b/dataset_c9b61090-5a2d-4dd0-8fd1-d147a4e1573c.dat
+# BUSCO was run in mode: genome
+# Gene predictor used: metaeuk
+
+	***** Results: *****
+
+	C:0.4%[S:0.4%,D:0.0%],F:0.0%,M:99.6%,n:255	   
+	1	Complete BUSCOs (C)			   
+	1	Complete and single-copy BUSCOs (S)	   
+	0	Complete and duplicated BUSCOs (D)	   
+	0	Fragmented BUSCOs (F)			   
+	254	Missing BUSCOs (M)			   
+	255	Total BUSCO groups searched		   
+
+Dependencies and versions:
+	hmmsearch: 3.1
+	prodigal: 2.6.3
+	metaeuk: 5.34c21f2
Binary file test-data/genome_results_metaeuk_auto/summary.png has changed