Mercurial > repos > matthias > stacks2_populations

diff stacks_populations.xml @ 6:1e260cccfafb draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/stacks2 commit f507c4eaca0fe04e730d7c3bbb9e9d2488239a9f
author matthias
date Thu, 20 Jun 2019 08:17:27 -0400
parents 35e673d79262
children
line wrap: on
line diff
--- a/stacks_populations.xml	Wed Feb 27 10:01:16 2019 -0500
+++ b/stacks_populations.xml	Thu Jun 20 08:17:27 2019 -0400
@@ -49,7 +49,7 @@
 -r $options_filtering.minperc
 -R $options_filtering.min_samples_overall
 $options_filtering.filter_haplotype_wise
---min_maf $options_filtering.minminor
+--min_maf $options_filtering.min_maf
 --min_mac $options_filtering.min_mac
 #if str($options_filtering.max_obs_het)
     --max_obs_het $options_filtering.max_obs_het
@@ -79,9 +79,9 @@
 ## Fstats
 #if str($fstats_conditional.fstats) != 'no':
     --fstats
-    #if str( $fstats_conditional.correction_select.correction ) != "no_corr"
-        --fst_correction $fstats_conditional.correction_select.correction
-        --p_value_cutoff $fstats_conditional.correction_select.pcutoff
+    #if str( $fstats_conditional.correction_select.fst_correction ) != "no_corr"
+        --fst_correction $fstats_conditional.correction_select.fst_correction
+        --p_value_cutoff $fstats_conditional.correction_select.p_value_cutoff
     #end if
 #end if
 
@@ -90,7 +90,7 @@
 #if str($kernel_smoothing.options_kernel.kernel)!='':
     --sigma $kernel_smoothing.options_kernel.sigma
 #end if
-#if $kernel_smoothing.bootstrap_resampling_mode.bootstrap_all
+#if $kernel_smoothing.bootstrap_resampling_mode.bootstrap
     --bootstrap
 #else
     $kernel_smoothing.bootstrap_resampling_mode.bootstrap_pifis
@@ -99,7 +99,7 @@
     $kernel_smoothing.bootstrap_resampling_mode.bootstrap_phist
 #end if
 
-#if $kernel_smoothing.bootstrap_resampling_mode.bootstrap_all or $kernel_smoothing.bootstrap_resampling_mode.bootstrap_pifis or $kernel_smoothing.bootstrap_resampling_mode.bootstrap_fst or $kernel_smoothing.bootstrap_resampling_mode.bootstrap_div or $kernel_smoothing.bootstrap_resampling_mode.bootstrap_phist:
+#if $kernel_smoothing.bootstrap_resampling_mode.bootstrap or $kernel_smoothing.bootstrap_resampling_mode.bootstrap_pifis or $kernel_smoothing.bootstrap_resampling_mode.bootstrap_fst or $kernel_smoothing.bootstrap_resampling_mode.bootstrap_div or $kernel_smoothing.bootstrap_resampling_mode.bootstrap_phist:
     #if str($kernel_smoothing.bootstrap_reps)
         --bootstrap_reps $kernel_smoothing.bootstrap_reps
     #end if
@@ -120,7 +120,7 @@
 ##$populations_output.phase
 ##$populations_output.fastphase
 $populations_output.plink
-##$populations_output.hzar
+$populations_output.hzar
 $populations_output.phylip
 $populations_output.phylip_var
 ##$populations_output.phylip_var_all
@@ -173,11 +173,11 @@
             <param name="minpop" argument="-p" type="integer" value="1" label="Minimum number of populations a locus must be present in to process a locus" />
             <param name="min_samples_overall" argument="--min-samples-overall" type="float" value="0.0" min="0" max="1" label="minimum percentage of individuals across populations required to process a locus" />
             <param name="filter_haplotype_wise" argument="--filter-haplotype-wise" type="boolean" truevalue="--filter-haplotype-wise" falsevalue="" checked="false" label="apply the above filters haplotype wise" help="unshared SNPs will be pruned to reduce haplotype-wise missing data." />
-            <param name="minminor" argument="--min_maf" type="float" value="0" min="0" max="0.5" label="Minimum minor allele frequency" help="specify a minimum minor allele frequency required to process a nucleotide site at a locus" />
+            <param argument="--min_maf" type="float" value="0" min="0" max="0.5" label="Minimum minor allele frequency" help="specify a minimum minor allele frequency required to process a nucleotide site at a locus" />
             <param argument="--min_mac" type="integer" value="0" min="0" label="Minimum minor allele count" help="specify a minimum minor allele count required to process a nucleotide site at a locus" />
             <param argument="--max_obs_het" type="float" value="" min="0" max="1" optional="true" label="Maximum observed heterozygosity required to process a nucleotide site at a locus." />
             <!-- deprecated https://groups.google.com/forum/#!msg/stacks-users/jdC7gw_MuK4/1v0FG3u7AwAJ
-            <param name="lnl" type="float" value="" optional="true" argument="\-\-lnl_lim" label="Filter loci with log likelihood values below this threshold" />-->
+            <param type="float" value="" optional="true" argument="\-\-lnl_lim" label="Filter loci with log likelihood values below this threshold" />-->
             <param name="filter_single_snp" type="select" label="Analyse all SNPs per locus" >
                 <option value="">yes</option>
                 <option value="--write_single_snp">No: Analyse only the first SNP (--write_single_snp)</option>
@@ -203,14 +203,14 @@
 
         <!-- Fstats -->
         <conditional name="fstats_conditional">
-            <param name="fstats" argument="--fstats" type="select" label="Enable SNP and haplotype-based F statistics" >
+            <param argument="--fstats" type="select" label="Enable SNP and haplotype-based F statistics" >
                 <option value="no">No</option>
                 <option value="yes">Yes</option>
             </param>
             <when value="no"/>
             <when value="yes">
                 <conditional name="correction_select">
-                    <param name="correction" type="select" label="Correction type" help="specify a correction to be applied to Fst values" >
+                    <param argument="--fst_correction" type="select" label="Correction type" help="specify a correction to be applied to Fst values" >
                         <option value="no_corr">No correction</option>
                         <option value="p_value">p_value</option>
                         <option value="bonferroni_win">bonferroni_win</option>
@@ -218,13 +218,13 @@
                     </param>
                     <when value="no_corr"/>
                     <when value="p_value">
-                        <param name="pcutoff" type="float" value="0.05" label="P-value cutoff" help="required p-value to keep an Fst measurement (0.05 by default). Also used as base for Bonferroni correction" />
+                        <param argument="--p_value_cutoff" type="float" value="0.05" label="P-value cutoff" help="required p-value to keep an Fst measurement (0.05 by default). Also used as base for Bonferroni correction" />
                     </when>
                     <when value="bonferroni_win">
-                        <param name="pcutoff" type="float" value="0.05" label="P-value cutoff" help="required p-value to keep an Fst measurement (0.05 by default). Also used as base for Bonferroni correction" />
+                        <param argument="--p_value_cutoff" type="float" value="0.05" label="P-value cutoff" help="required p-value to keep an Fst measurement (0.05 by default). Also used as base for Bonferroni correction" />
                     </when>
                     <when value="bonferroni_gen">
-                        <param name="pcutoff" type="float" value="0.05" label="P-value cutoff" help="required p-value to keep an Fst measurement (0.05 by default). Also used as base for Bonferroni correction" />
+                        <param argument="--p_value_cutoff" type="float" value="0.05" label="P-value cutoff" help="required p-value to keep an Fst measurement (0.05 by default). Also used as base for Bonferroni correction" />
                     </when>
                 </conditional>
             </when>
@@ -232,21 +232,21 @@
 
         <!-- File output options -->
         <section name="populations_output" title="Output options" expanded="true">
-            <param name="ordered_export" argument="--ordered_export" truevalue="--ordered_export" falsevalue="" type="boolean" checked="false" label="If data is reference aligned, exports will be ordered; only a single representative of each overlapping site." />
-            <param name="fasta_loci" argument="--fasta_loci" truevalue="--fasta_loci" falsevalue="" type="boolean" checked="false" label="Output locus consensus sequences in FASTA format" />
-            <param name="fasta_samples" argument="--fasta_samples" truevalue="--fasta_samples" falsevalue="" type="boolean" checked="false" label="Output the sequences of the two haplotypes of each (diploid) sample, for each locus, in FASTA format" />
-            <param name="vcf" argument="--vcf" truevalue="--vcf" falsevalue="" type="boolean" checked="false" label="Output results in Variant Call Format (VCF)" />
-            <param name="genepop" argument="--genepop" truevalue="--genepop" falsevalue="" type="boolean" checked="false" label="Output results in GenePop Format" />
-            <param name="structure" argument="--structure" truevalue="--structure" falsevalue="" type="boolean" checked="false" label="Output results in Structure Format" />
+            <param argument="--ordered_export" truevalue="--ordered_export" falsevalue="" type="boolean" checked="false" label="If data is reference aligned, exports will be ordered; only a single representative of each overlapping site." />
+            <param argument="--fasta_loci" truevalue="--fasta_loci" falsevalue="" type="boolean" checked="false" label="Output locus consensus sequences in FASTA format" />
+            <param argument="--fasta_samples" truevalue="--fasta_samples" falsevalue="" type="boolean" checked="false" label="Output the sequences of the two haplotypes of each (diploid) sample, for each locus, in FASTA format" />
+            <param argument="--vcf" truevalue="--vcf" falsevalue="" type="boolean" checked="false" label="Output results in Variant Call Format (VCF)" />
+            <param argument="--genepop" truevalue="--genepop" falsevalue="" type="boolean" checked="false" label="Output results in GenePop Format" />
+            <param argument="--structure" truevalue="--structure" falsevalue="" type="boolean" checked="false" label="Output results in Structure Format" />
             <param argument="--radpainter" truevalue="--radpainter" falsevalue="" type="boolean" checked="false" label="Output results in fineRADstructure/RADpainter format" />
-            <!--<param name="phase" argument="\-\-phase" truevalue="\-\-phase" falsevalue="" type="boolean" checked="false" label="Output genotypes in PHASE format" />-->
-            <param name="fastphase" argument="--fastphase" truevalue="--fastphase" falsevalue="" type="boolean" checked="false" label="Output genotypes in fastPHASE format" />
-            <param name="plink" argument="--plink" truevalue="--plink" falsevalue="" type="boolean" checked="false" label="Output genotypes in PLINK format" />
-            <!--<param name="hzar" argument="\-\-hzar" truevalue="\-\-hzar" falsevalue="" type="boolean" checked="false" label="Output genotypes in Hybrid Zone Analysis using R (HZAR) format." />-->
-            <param name="phylip" argument="--phylip" truevalue="--phylip" falsevalue="" type="boolean" checked="false" label="Output nucleotides that are fixed-within, and variant among populations in Phylip format for phylogenetic tree construction" />
-            <param name="phylip_var" argument="--phylip_var" truevalue="--phylip_var" falsevalue="" type="boolean" checked="false" label="Include variable sites in the phylip output encoded using IUPAC notation." />
-            <!--<param name="phylip_var_all" argument="\-\-phylip_var_all" truevalue="\-\-phylip_var_all" falsevalue="" type="boolean" checked="false" label="Include all sequence as well as variable sites in the phylip output encoded using IUPAC notation." />-->
-            <param name="treemix" argument="--treemix" truevalue="--treemix" falsevalue="" type="boolean" checked="false" label="Output SNPs in a format useable for the TreeMix program (Pickrell and Pritchard)." />
+            <!--<param argument="\-\-phase" truevalue="\-\-phase" falsevalue="" type="boolean" checked="false" label="Output genotypes in PHASE format" />-->
+            <param argument="--fastphase" truevalue="--fastphase" falsevalue="" type="boolean" checked="false" label="Output genotypes in fastPHASE format" />
+            <param argument="--plink" truevalue="--plink" falsevalue="" type="boolean" checked="false" label="Output genotypes in PLINK format" />
+            <param argument="--hzar" truevalue="--hzar" falsevalue="" type="boolean" checked="false" label="Output genotypes in Hybrid Zone Analysis using R (HZAR) format." />
+            <param argument="--phylip" truevalue="--phylip" falsevalue="" type="boolean" checked="false" label="Output nucleotides that are fixed-within, and variant among populations in Phylip format for phylogenetic tree construction" />
+            <param argument="--phylip_var" truevalue="--phylip_var" falsevalue="" type="boolean" checked="false" label="Include variable sites in the phylip output encoded using IUPAC notation." />
+            <!--<param argument="\-\-phylip_var_all" truevalue="\-\-phylip_var_all" falsevalue="" type="boolean" checked="false" label="Include all sequence as well as variable sites in the phylip output encoded using IUPAC notation." />-->
+            <param argument="--treemix" truevalue="--treemix" falsevalue="" type="boolean" checked="false" label="Output SNPs in a format useable for the TreeMix program (Pickrell and Pritchard)." />
             <param argument="--no_hap_exports" truevalue="--no_hap_exports" falsevalue="" type="boolean" checked="false" label="Omit haplotype outputs" />
             <param argument="--fasta_samples_raw" truevalue="--fasta_samples_raw" falsevalue="" type="boolean" checked="false" label="Output all haplotypes observed in each sample, for each locus, in FASTA format" />
         </section>
@@ -256,9 +256,9 @@
             <conditional name="options_kernel">
                 <param name="kernel" type="select" label="Enable kernel-smoothed calculations" >
                     <option value="">No</option>
-                    <option value="-k">For Pi, Fis, Fst, Fst', and Phi_st</option>
-                    <option value="--smooth_fstats">For Fst, Fst', and Phi_st</option>
-                    <option value="--smooth_popstats">For Pi and Fis</option>
+                    <option value="-k">For Pi, Fis, Fst, Fst', and Phi_st (-k)</option>
+                    <option value="--smooth_fstats">For Fst, Fst', and Phi_st (--smooth_fstats)</option>
+                    <option value="--smooth_popstats">For Pi and Fis (--smooth_popstats)</option>
                 </param>
                 <when value=""/>
                 <when value="-k">
@@ -272,21 +272,21 @@
                 </when>
             </conditional>
             <conditional name="bootstrap_resampling_mode">
-                <param name="bootstrap_all" argument="--bootstrap" type="select" label="Enable bootstrap resampling for all smoothed statistics" >
+                <param argument="--bootstrap" type="select" label="Enable bootstrap resampling for all smoothed statistics" >
                     <option value="">No</option>
                     <option value="--bootstrap">Yes</option>
                 </param>
                 <when value="--bootstrap">
                 </when>
                 <when value="">
-                    <param name="bootstrap_pifis" argument="--bootstrap_pifis" type="boolean" checked="false" truevalue="--bootstrap_pifis" falsevalue="" label="Enable boostrap resampling for smoothed SNP-based Pi and Fis calculations" />
-                    <param name="bootstrap_fst" argument="--bootstrap_fst" type="boolean" checked="false" truevalue="--bootstrap_fst" falsevalue="" label="Enable boostrap resampling for smoothed Fst calculations based on pairwise population comparison of SNPs" />
-                    <param name="bootstrap_div" argument="--bootstrap_div" type="boolean" checked="false" truevalue="--bootstrap_div" falsevalue="" label="Enable boostrap resampling for smoothed haplotype diveristy and gene diversity calculations based on haplotypes" />
-                    <param name="bootstrap_phist" argument="--bootstrap_phist" type="boolean" checked="false" truevalue="--bootstrap_phist" falsevalue="" label="Enable boostrap resampling for smoothed Phi_st calculations based on haplotypes." />
+                    <param argument="--bootstrap_pifis" type="boolean" checked="false" truevalue="--bootstrap_pifis" falsevalue="" label="Enable boostrap resampling for smoothed SNP-based Pi and Fis calculations" />
+                    <param argument="--bootstrap_fst" type="boolean" checked="false" truevalue="--bootstrap_fst" falsevalue="" label="Enable boostrap resampling for smoothed Fst calculations based on pairwise population comparison of SNPs" />
+                    <param argument="--bootstrap_div" type="boolean" checked="false" truevalue="--bootstrap_div" falsevalue="" label="Enable boostrap resampling for smoothed haplotype diveristy and gene diversity calculations based on haplotypes" />
+                    <param argument="--bootstrap_phist" type="boolean" checked="false" truevalue="--bootstrap_phist" falsevalue="" label="Enable boostrap resampling for smoothed Phi_st calculations based on haplotypes." />
                 </when>
             </conditional>
-            <param name="bootstrap_reps" argument="--bootstrap_reps" type="integer" value="100" optional="true" label="Number of bootstrap resamplings to calculate" />
-            <param name="bootstrap_wl" argument="--bootstrap_wl" format="txt,tabular" type="data" optional="true" label="Only bootstrap loci contained in this whitelist" />
+            <param argument="--bootstrap_reps" type="integer" value="100" optional="true" label="Number of bootstrap resamplings to calculate" />
+            <param argument="--bootstrap_wl" format="txt,tabular" type="data" optional="true" label="Only bootstrap loci contained in this whitelist" />
         </section>
 
         <!-- Output options -->
@@ -358,6 +358,7 @@
             <param name="populations_output|phylip_var" value="yes"/>
             <param name="populations_output|genepop" value="yes"/>
             <param name="populations_output|vcf" value="yes"/>
+            <param name="populations_output|hzar" value="yes"/>
             <param name="populations_output|plink" value="yes"/>
             <param name="populations_output|structure" value="yes"/>
             <param name="populations_output|radpainter" value="yes"/>
@@ -376,22 +377,23 @@
             <output ftype="tabular" name="out_phistats" value="populations/populations.phistats.tsv"/>
             <output ftype="tabular" name="out_phistats_sum" value="populations/populations.phistats_summary.tsv"/>
             <output ftype="tabular" name="out_fststats_sum" value="populations/populations.fst_summary.tsv"/>
-            <output ftype="tabular" name="out_fasta_strict" value="populations/populations.loci.fa" lines_diff="2"/>
-            <output ftype="tabular" name="out_fasta" value="populations/populations.samples.fa" lines_diff="2"/>
-            <output ftype="tabular" name="out_fasta_raw" value="populations/populations.samples-raw.fa" lines_diff="2"/>
-            <output ftype="tabular" name="out_phylip_all_pop_fix" value="populations/populations.fixed.phylip" lines_diff="2"/>
+			<output ftype="tabular" name="out_fasta_strict" value="populations/populations.loci.fa" compare="sim_size" delta="50"/><!--there seems to be reordering of the lines, so differences are expected only in the date contained in the Header -->
+			<output ftype="tabular" name="out_fasta" value="populations/populations.samples.fa" compare="sim_size" delta="50"/><!-- " -->
+            <output ftype="tabular" name="out_fasta_raw" value="populations/populations.samples-raw.fa" compare="sim_size" delta="50"/><!-- " -->
+            <output ftype="tabular" name="out_phylip_all_pop_fix" value="populations/populations.fixed.phylip" lines_diff="2"/><!-- " -->
             <output ftype="tabular" name="out_phylip_all_loci_fix" value="populations/populations.fixed.phylip.log" lines_diff="2"/>
-            <output ftype="tabular" name="out_phylip_all_pop_var" value="populations/populations.var.phylip" lines_diff="2"/>
-            <output ftype="tabular" name="out_phylip_all_loci_var" value="populations/populations.var.phylip.log" lines_diff="2"/>
-            <output ftype="tabular" name="out_genepop_snps" value="populations/populations.snps.genepop" lines_diff="2"/>
-            <output ftype="tabular" name="out_genepop_haps" value="populations/populations.haps.genepop" lines_diff="2"/>
-            <output ftype="vcf" name="out_vcf_haplotypes_snps" value="populations/populations.snps.vcf" lines_diff="2"/>
-            <output ftype="vcf" name="out_vcf_haplotypes_haps" value="populations/populations.haps.vcf" lines_diff="2"/>
-            <output ftype="tabular" name="out_plink_markers" value="populations/populations.plink.map" lines_diff="2"/>
-            <output ftype="tabular" name="out_plink_genotypes" value="populations/populations.plink.ped" lines_diff="2"/>
-            <output ftype="tabular" name="out_structure" value="populations/populations.structure" lines_diff="2"/>
-            <output ftype="tabular" name="out_radpainter" value="populations/populations.haps.radpainter"/>
-            <output ftype="tabular" name="out_treemix" value="populations/populations.treemix"/>
+            <output ftype="tabular" name="out_phylip_all_pop_var" value="populations/populations.var.phylip" compare="sim_size" delta="50"/>
+            <output ftype="tabular" name="out_phylip_all_loci_var" value="populations/populations.var.phylip.log" compare="sim_size" delta="50"/><!--there seems to be reordering of the lines, so differences are expected only in the date contained in the Header -->
+            <output ftype="tabular" name="out_genepop_snps" value="populations/populations.snps.genepop" compare="sim_size" delta="50"/><!-- " -->
+            <output ftype="tabular" name="out_genepop_haps" value="populations/populations.haps.genepop" compare="sim_size" delta="50"/><!-- " -->
+            <output ftype="vcf" name="out_vcf_haplotypes_snps" value="populations/populations.snps.vcf" compare="sim_size" delta="50"/><!-- " -->
+            <output ftype="vcf" name="out_vcf_haplotypes_haps" value="populations/populations.haps.vcf" compare="sim_size" delta="50"/><!-- " -->
+            <output ftype="tabular" name="out_plink_markers" value="populations/populations.plink.map" compare="sim_size" delta="50"/><!-- " -->
+            <output ftype="tabular" name="out_plink_genotypes" value="populations/populations.plink.ped" compare="sim_size" delta="50"/><!-- " -->
+            <output ftype="tabular" name="out_hzar" value="populations/populations.hzar.csv" compare="sim_size" delta="50"/><!-- " -->
+            <output ftype="tabular" name="out_structure" value="populations/populations.structure" compare="sim_size" delta="50"/><!-- " -->
+            <output ftype="tabular" name="out_radpainter" value="populations/populations.haps.radpainter" compare="sim_size" delta="50"/><!-- " -->
+            <output ftype="tabular" name="out_treemix" value="populations/populations.treemix" compare="sim_size" delta="50"/><!-- " -->
         </test>
         <!-- test w vcf input and default options, just checking if finished -->
         <test>
@@ -423,7 +425,7 @@
                 <param name="minpop" value="2"/>
                 <param name="min_samples_overall" value="0.01"/>
                 <param name="filter_haplotype_wise" value="--filter-haplotype-wise"/>
-                <param name="minminor" value="0.01" />
+                <param name="min_maf" value="0.01" />
                 <param name="min_mac" value="1"/>
                 <param name="max_obs_het" value="0.8" />
                 <param name="blacklist" value="populations/blacklist.tsv" ftype="tabular" />
@@ -461,7 +463,7 @@
                 <param name="minperc" value="0.1"/>
                 <param name="minpop" value="2"/>
                 <param name="min_samples_overall" value="0.01"/>
-                <param name="minminor" value="0.01" />
+                <param name="min_maf" value="0.01" />
                 <param name="min_mac" value="1"/>
                 <param name="max_obs_het" value="0.8" />
                 <!-- since write_random_snp runs in ifinite loop switch to write_single -->
@@ -533,8 +535,8 @@
             <conditional name="fstats_conditional">
                 <param name="fstats" value="yes" />
                 <conditional name="correction_select">
-                    <param name="correction" value="bonferroni_gen"/>
-                    <param name="pcutoff" value="0.01" />
+                    <param name="fst_correction" value="bonferroni_gen"/>
+                    <param name="p_value_cutoff" value="0.01" />
                 </conditional>
             </conditional>
             <conditional name="options_kernel">
@@ -542,7 +544,7 @@
                 <param name="sigma" value="150001" />
             </conditional>
             <conditional name="bootstrap_resampling_mode">
-                <param name="bootstrap_all" value="--bootstrap" />
+                <param name="bootstrap" value="--bootstrap" />
             </conditional>
             <param name="bootstrap_reps" value="23" />
             <param name="add_log" value="yes" />