Merge branch 'jaffal_crash' into 'dev'

Fix for when JAFFAL returns exit status 1 when no fusions are found

See merge request epi2melabs/workflows/wf-transcriptomes!80

.gitlab-ci.yml

@@ -12,9 +12,6 @@ variables:
 
 docker-run:
 
-    # Remove this directive in downstream templates
-    tags: [large_ram] # no need for big ram
-
     # Define a 1D job matrix to inject a variable named MATRIX_NAME into
     #   the CI environment, we can use the value of MATRIX_NAME to determine
     #   which options to apply as part of the rules block below

CHANGELOG.md

@@ -11,6 +11,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Demo differential expression data in repository.
 - Improved DE explanation in docs
 ### Fixed
+- Fix JAFFAL terminating workflow when no fusions found.
 - Error if condition sheet and sample sheet don't match.
 
 ## [v0.1.5]

bin/report.py

@@ -791,31 +791,43 @@ def jaffal_table(report, result_csv):
         'sample_id', 'fusion genes', 'chrom1', 'chrom2', 'spanning reads',
         'classification', 'known']
 
-    df = pd.read_csv(result_csv)
-    sid_col = df.pop('sample_id')
-    df.insert(0, 'sample_id', sid_col)
-
-    df = df[cols]
-    df['chroms'] = df.chrom1.astype(str) + ':' + df.chrom2.astype(str)
-    df.rename(columns={
-        'spanning reads': 'nreads',
-        'fusion genes': 'genes'}, inplace=True)
-    df.drop(columns=['chrom1', 'chrom2'], inplace=True)
     section = report.add_section()
-    section.markdown("""
-    ### JAFFAL fusion transcript summary
+    try:
+        df = pd.read_csv(result_csv)
+    except pd.errors.EmptyDataError:
+        section.markdown("""
+        ### JAFFAL fusion transcript summary
 
-    This table summarizes putative fusion transcripts identified
-    by [JAFFAL](https://github.com/Oshlack/JAFFA/).
+        This section summarizes putative fusion transcripts identified
+        by [JAFFAL](https://github.com/Oshlack/JAFFA/).
 
-    * genes: the gene symbols of the fusion partners
-    * nreads: The number of reads supporting the fusion
-    * classification: JAFFAL's classification
-    * known: whether this fusion is in the given set of known gene fusions
-    * chroms: the respective, original chromosome location of the two partner
-    genes
-    """)
-    section.table(df)
+        No fusion transcripts detected for current sample.
+        """)
+    else:
+        sid_col = df.pop('sample_id')
+        df.insert(0, 'sample_id', sid_col)
+
+        df = df[cols]
+        df['chroms'] = df.chrom1.astype(str) + ':' + df.chrom2.astype(str)
+        df.rename(columns={
+            'spanning reads': 'nreads',
+            'fusion genes': 'genes'}, inplace=True)
+        df.drop(columns=['chrom1', 'chrom2'], inplace=True)
+
+        section.markdown("""
+        ### JAFFAL fusion transcript summary
+
+        This table summarizes putative fusion transcripts identified
+        by [JAFFAL](https://github.com/Oshlack/JAFFA/).
+
+        * genes: the gene symbols of the fusion partners
+        * nreads: The number of reads supporting the fusion
+        * classification: JAFFAL's classification
+        * known: whether this fusion is in the given set of known gene fusions
+        * chroms: the respective, original chromosome location of the
+        two partner genes
+        """)
+        section.table(df)
 
 
 def de_section(report):

main.nf

@@ -454,7 +454,6 @@ workflow pipeline {
             jaffal_out = file("$projectDir/data/OPTIONAL_FILE_1")
         }
 
-        
         get_transcriptome(
             merge_gff_bundles.out.gff
             .join(run_gffcompare.out.gffcmp_dir)

subworkflows/JAFFAL/gene_fusions.nf

@@ -12,6 +12,8 @@ process jaffal{
     script:
     """
     JAFFAOUT=jaffal_output_$sample_id
+
+    # JAFFAL exists with status code 1 when there's 0 fusion hits. Prevent this with '||:'
     $params.jaffal_dir/tools/bin/bpipe run \
         -n $params.threads \
         -p jaffa_output="\$JAFFAOUT/" \
@@ -20,14 +22,26 @@ process jaffal{
         -p annotation=$annotation \
         -p fastqInputFormat="*.fastq" \
         $params.jaffal_dir/JAFFAL.groovy \
-        $fastq
-    mv "\$JAFFAOUT/jaffa_results.csv" "\$JAFFAOUT/${sample_id}_jaffa_results.csv"
+        $fastq || :
 
-    # Add sample id column
-    sed "s/\$/,${sample_id}/" \$JAFFAOUT/${sample_id}_jaffa_results.csv > tmp1
-    # Add header
-    sed "1 s/${sample_id}/sample_id/" tmp1 > tmp2
-    mv tmp2 \$JAFFAOUT/${sample_id}_jaffa_results.csv
+    summary="\$JAFFAOUT/all/all.summary"
+
+    if [ -f \$summary ]; then
+        # The summary is writtten so assume JAFFAL completed.
+        if [ ! -s \$summary ]; then
+            echo "JAFFAL failed to find any fusion transcripts for ${sample_id}"
+            touch "\$JAFFAOUT/${sample_id}_jaffa_results.csv"
+        else
+            echo JAFFAL found fusion transcripts for ${sample_id}
+            mv "\$JAFFAOUT/jaffa_results.csv" "\$JAFFAOUT/${sample_id}_jaffa_results.csv"
+            # Add sample id column and header
+            sed "s/\$/,${sample_id}/" \$JAFFAOUT/${sample_id}_jaffa_results.csv \
+                | sed "1 s/${sample_id}/sample_id/" > tmp
+            mv tmp \$JAFFAOUT/${sample_id}_jaffa_results.csv
+        fi
+    else
+        echo JAFFAL encountered an error while prosessing ${sample_id}
+    fi
     """
 }