Update snapatac2 to 2.6.4 (galaxyproject#6157)

* Update snapatac2 to 2.6.4 * add min_len parameter and fix test * add tests back
pavanvidem · Jul 17, 2024 · b37248e · b37248e
1 parent bfc4ff4
commit b37248e
Show file tree

Hide file tree

Showing 3 changed files with 58 additions and 32 deletions.
diff --git a/tools/snapatac2/macros.xml b/tools/snapatac2/macros.xml
@@ -1,16 +1,16 @@
 <macros>
-    <token name="@TOOL_VERSION@">2.5.3</token>
-    <token name="@VERSION_SUFFIX@">2</token>
+    <token name="@TOOL_VERSION@">2.6.4</token>
+    <token name="@VERSION_SUFFIX@">0</token>
     <token name="@PROFILE@">23.0</token>
     <xml name="requirements">
         <requirement type="package" version="@TOOL_VERSION@">snapatac2</requirement>
-        <requirement type="package" version="5.18.0">plotly</requirement>
+        <requirement type="package" version="5.22.0">plotly</requirement>
         <requirement type="package" version="0.2.1">python-kaleido</requirement>
-        <requirement type="package" version="0.19.19">polars</requirement>
-        <requirement type="package" version="14.0.1">pyarrow</requirement>
-        <requirement type="package" version="0.11.3">python-igraph</requirement>
-        <requirement type="package" version="0.8.33">hdbscan</requirement>
-        <requirement type="package" version="0.0.9">harmonypy</requirement>
+        <requirement type="package" version="1.1.0">polars</requirement>
+        <requirement type="package" version="16.1.0">pyarrow</requirement>
+        <requirement type="package" version="0.11.6">python-igraph</requirement>
+        <requirement type="package" version="0.8.37">hdbscan</requirement>
+        <requirement type="package" version="0.0.10">harmonypy</requirement>
         <requirement type="package" version="1.7.4">scanorama</requirement>
         <requirement type="package" version="3.0.1">macs3</requirement>
         <requirement type="package" version="0.70.16">multiprocess</requirement>
@@ -158,10 +158,10 @@ with open('anndata_info.txt','w', encoding='utf-8') as ainfo:
         <has_text_matching expression="height = 450"/>
     </xml>
     <xml name="param_counting_strategy">
-        <param argument="counting_strategy" type="select" label="he strategy to compute feature counts">
-            <option value="fragment">fragment</option>
-            <option value="insertion" selected="true">insertion</option>
-            <option value="paired-insertion">paired-insertion</option>
+        <param argument="counting_strategy" type="select" label="The strategy to compute feature counts">
+            <option value="fragment">"fragment": based on the number of fragments that overlap with a region of interest</option>
+            <option value="insertion" selected="true">"insertion": based on the number of insertions that overlap with a region of interest</option>
+            <option value="paired-insertion">"paired-insertion": similar to "insertion", but it only counts the insertions once if the pair of insertions of a fragment are both within the same region of interest</option>
         </param>
     </xml>
 

diff --git a/tools/snapatac2/peaks_and_motif_analysis.xml b/tools/snapatac2/peaks_and_motif_analysis.xml
@@ -36,6 +36,9 @@ if __name__ == '__main__': ## a temporary fix https://github.com/kaizhang/SnapAT
         nolambda = $method.nolambda,
         shift = $method.shift,
         extsize = $method.extsize,
+        #if $method.min_len
+        min_len = $method.min_len,
+        #end if
         #if $method.blacklist
         blacklist = '$method.blacklist',
         #end if
@@ -62,7 +65,7 @@ peaks.write_csv('merged_peaks.tabular', separator = '\t')
 #else if $method.method == 'pp.make_peak_matrix'
 import polars
 peaks = polars.read_csv('$method.merged_peaks', separator='\t')
-sa.pp.make_peak_matrix(
+adata = sa.pp.make_peak_matrix(
     adata,
     use_rep = peaks['Peaks'],
     chunk_size = $method.chunk_size,
@@ -73,8 +76,7 @@ sa.pp.make_peak_matrix(
     #if $method.max_frag_size
     max_frag_size = $method.max_frag_size,
     #end if
-    ##counting_strategy = '$method.counting_strategy'
-    count_frag_as_reads = $method.count_frag_as_reads
+    counting_strategy = '$method.counting_strategy'
 )
 
 #else if $method.method == 'tl.marker_regions'
@@ -150,7 +152,9 @@ sa.pl.regions(
 )
 
 #end if
+#if $method.method != 'pp.make_peak_matrix'
 @CMD_anndata_write_outputs@
+#end if
     ]]></configfile>
     </configfiles>
     <inputs>
@@ -173,6 +177,7 @@ sa.pl.regions(
                 <param argument="nolambda" type="boolean" checked="false" truevalue="True" falsevalue="False" label="If True, macs3 will use the background lambda as local lambda"/>
                 <param argument="shift" type="integer" value="-100" label="The shift size in MACS"/>
                 <param argument="extsize" type="integer" value="200" label="The extension size in MACS"/>
+                <param argument="min_len" type="integer" optional="true" value="" label="The minimum length of a called peak" help="If not set, it is set to the extension size."/>
                 <param argument="blacklist" type="data" format="bed" optional="true" label="Blacklist file in BED format"
                     help="If provided, regions in the blacklist will be removed."/>
                 <param argument="key_added" type="text" value="macs3"  label="`.uns` key under which to add peak information"/>
@@ -192,8 +197,7 @@ sa.pl.regions(
                 <expand macro="param_chunk_size" size="500"/>
                 <param argument="use_x" type="boolean" truevalue="True" falsevalue="False" checked="false" label="If True, use the matrix stored in .X as raw counts"/>
                 <expand macro="min_max_frag_size"/>
-                <!--expand macro="param_counting_strategy"/-->
-                <param argument="count_frag_as_reads" type="boolean" truevalue="True" falsevalue="False" checked="true" label="Whether to count fragments as reads"/>
+                <expand macro="param_counting_strategy"/>
             </when>
             <when value="tl.marker_regions">
                 <expand macro="inputs_anndata"/>
@@ -264,6 +268,7 @@ sa.pl.regions(
                 <param name="qvalue" value="0.1"/>
                 <param name="shift" value="-100"/>
                 <param name="extsize" value="200"/>
+                <param name="min_len" value="50"/>
                 <param name="key_added" value="macs3"/>
                 </conditional>
             <section name="advanced_common">
@@ -276,6 +281,7 @@ sa.pl.regions(
                     <has_text_matching expression="qvalue = 0.1"/>
                     <has_text_matching expression="shift = -100"/>
                     <has_text_matching expression="extsize = 200"/>
+                    <has_text_matching expression="min_len = 50"/>
                     <has_text_matching expression="key_added = 'macs3'"/>
                 </assert_contents>
             </output>
@@ -317,7 +323,7 @@ sa.pl.regions(
                 <param name="merged_peaks" location="https://zenodo.org/records/11260316/files/merged_peaks.tabular"/>
                 <param name="chunk_size" value="500"/>
                 <param name="use_x" value="False"/>
-                <param name="count_frag_as_reads" value="True"/>
+                <param name="counting_strategy" value="insertion"/>
                 </conditional>
             <section name="advanced_common">
                 <param name="show_log" value="true" />
@@ -327,10 +333,10 @@ sa.pl.regions(
                     <has_text_matching expression="sa.pp.make_peak_matrix"/>
                     <has_text_matching expression="chunk_size = 500"/>
                     <has_text_matching expression="use_x = False"/>
-                    <has_text_matching expression="count_frag_as_reads = True"/>
+                    <has_text_matching expression="counting_strategy = 'insertion'"/>
                 </assert_contents>
             </output>
-            <output name="anndata_out" location="https://zenodo.org/records/11260316/files/pp.make_peak_matrix.pbmc_500_chr21.h5ad" ftype="h5ad" compare="sim_size" delta_frac="0.1" />
+            <output name="anndata_out" location="https://zenodo.org/records/12751925/files/pp.make_peak_matrix.pbmc_500_chr21.h5ad" ftype="h5ad" compare="sim_size" delta_frac="0.1" />
         </test>
         <test expect_num_outputs="3">
             <!-- tl.marker_regions -->

diff --git a/tools/snapatac2/preprocessing.xml b/tools/snapatac2/preprocessing.xml
@@ -85,8 +85,7 @@ sa.pp.add_tile_matrix(
     #if $method.max_frag_size
     max_frag_size = $method.max_frag_size,
     #end if
-    ##counting_strategy = '$method.counting_strategy',
-    count_frag_as_reads = $method.count_frag_as_reads,
+    counting_strategy = '$method.counting_strategy',
     n_jobs = int(os.getenv("GALAXY_SLOTS", 4))
 )
 
@@ -107,8 +106,7 @@ adata = sa.pp.make_gene_matrix(
     #if $method.max_frag_size
     max_frag_size = $method.max_frag_size,
     #end if
-    ##counting_strategy = '$method.counting_strategy'
-    count_frag_as_reads = $method.count_frag_as_reads
+    counting_strategy = '$method.counting_strategy'
 )
 
 #else if $method.method == 'pp.filter_cells'
@@ -285,8 +283,7 @@ sa.metrics.tsse(
                     <expand macro="sanitize_query"/>
                 </param>
                 <expand macro="min_max_frag_size"/>
-                <!--expand macro="param_counting_strategy"/-->
-                <param argument="count_frag_as_reads" type="boolean" truevalue="True" falsevalue="False" checked="true" label="Whether to count fragments as reads"/>
+                <expand macro="param_counting_strategy"/>
             </when>
             <when value="pp.make_gene_matrix">
                 <expand macro="inputs_anndata"/>
@@ -302,8 +299,7 @@ sa.metrics.tsse(
                 <param argument="gene_name_key" type="text" value="gene_name" label="The key of the gene name in the gene annotation file"/>
                 <param argument="gene_id_key" type="text" value="gene_id" label="The key of the gene id in the gene annotation file"/>
                 <expand macro="min_max_frag_size"/>
-                <!--expand macro="param_counting_strategy"/-->
-                <param argument="count_frag_as_reads" type="boolean" truevalue="True" falsevalue="False" checked="true" label="Whether to count fragments as reads"/>
+                <expand macro="param_counting_strategy"/>
             </when>
             <when value="pp.filter_cells">
                 <expand macro="inputs_anndata"/>
@@ -446,7 +442,7 @@ sa.metrics.tsse(
                 <param name="transcript_id_key" value="transcript_id"/>
                 <param name="gene_name_key" value="gene_name"/>
                 <param name="gene_id_key" value="gene_id"/>
-                <param name="count_frag_as_reads" value="True"/>
+                <param name="counting_strategy" value="insertion"/>
                 </conditional>
             <section name="advanced_common">
                 <param name="show_log" value="true" />
@@ -461,7 +457,7 @@ sa.metrics.tsse(
                     <has_text_matching expression="transcript_id_key = 'transcript_id'"/>
                     <has_text_matching expression="gene_name_key = 'gene_name'"/>
                     <has_text_matching expression="gene_id_key = 'gene_id'"/>
-                    <has_text_matching expression="count_frag_as_reads = True"/>
+                    <has_text_matching expression="counting_strategy = 'insertion'"/>
                 </assert_contents>
             </output>
             <output name="anndata_out" location="https://zenodo.org/records/12548681/files/pp.make_gene_matrix.pbmc_500_chr21.h5ad" ftype="h5ad" compare="sim_size" delta_frac="0.1" />
@@ -513,7 +509,7 @@ sa.metrics.tsse(
                 <param name="bin_size" value="5000"/>
                 <param name="chunk_size" value="500"/>
                 <param name="exclude_chroms" value="chr1, chr2, chr3, chr4, chr5, chr6, chr7, chr8, chr9, chr10, chr11, chr12, chr13, chr14, chr15, chr16, chr17, chr18, chr19, chr20, chr22, chrX, chrY"/>
-                <param name="count_frag_as_reads" value="True"/>
+                <param name="counting_strategy" value="insertion"/>
                 </conditional>
             <section name="advanced_common">
                 <param name="show_log" value="true" />
@@ -524,7 +520,31 @@ sa.metrics.tsse(
                     <has_text_matching expression="bin_size = 5000"/>
                     <has_text_matching expression="chunk_size = 500"/>
                     <has_text_matching expression="exclude_chroms = \['chr1', 'chr2', 'chr3', 'chr4', 'chr5', 'chr6', 'chr7', 'chr8', 'chr9', 'chr10', 'chr11', 'chr12', 'chr13', 'chr14', 'chr15', 'chr16', 'chr17', 'chr18', 'chr19', 'chr20', 'chr22', 'chrX', 'chrY'\]"/>
-                    <has_text_matching expression="count_frag_as_reads = True"/>
+                    <has_text_matching expression="counting_strategy = 'insertion'"/>
+                </assert_contents>
+            </output>
+            <output name="anndata_out"  ftype="h5ad" compare="sim_size" delta_frac="0.1" location="https://zenodo.org/records/11260316/files/pp.add_tile_matrix.pbmc_500_chr21.h5ad"/>
+        </test>
+        <test expect_num_outputs="2">
+            <!-- pp.add_tile_matrix counting_strategy fragment -->
+            <conditional name="method">
+                <param name="method" value="pp.add_tile_matrix"/>
+                <param name="adata" location="https://zenodo.org/records/11260316/files/pp.filter_cells.pbmc_500_chr21.h5ad"/>
+                <param name="bin_size" value="5000"/>
+                <param name="chunk_size" value="500"/>
+                <param name="exclude_chroms" value="chr1, chr2, chr3, chr4, chr5, chr6, chr7, chr8, chr9, chr10, chr11, chr12, chr13, chr14, chr15, chr16, chr17, chr18, chr19, chr20, chr22, chrX, chrY"/>
+                <param name="counting_strategy" value="fragment"/>
+                </conditional>
+            <section name="advanced_common">
+                <param name="show_log" value="true" />
+            </section>
+            <output name="hidden_output">
+                <assert_contents>
+                    <has_text_matching expression="sa.pp.add_tile_matrix"/>
+                    <has_text_matching expression="bin_size = 5000"/>
+                    <has_text_matching expression="chunk_size = 500"/>
+                    <has_text_matching expression="exclude_chroms = \['chr1', 'chr2', 'chr3', 'chr4', 'chr5', 'chr6', 'chr7', 'chr8', 'chr9', 'chr10', 'chr11', 'chr12', 'chr13', 'chr14', 'chr15', 'chr16', 'chr17', 'chr18', 'chr19', 'chr20', 'chr22', 'chrX', 'chrY'\]"/>
+                    <has_text_matching expression="counting_strategy = 'fragment'"/>
                 </assert_contents>
             </output>
             <output name="anndata_out"  ftype="h5ad" compare="sim_size" delta_frac="0.1" location="https://zenodo.org/records/11260316/files/pp.add_tile_matrix.pbmc_500_chr21.h5ad"/>