fix: making compatible with varfish export and mehari

bihealth · Jul 17, 2023 · 67d8009 · 67d8009
1 parent 04551d5
commit 67d8009
Show file tree

Hide file tree

Showing 2 changed files with 22 additions and 3 deletions.
diff --git a/snappy_pipeline/workflows/varfish_export/__init__.py b/snappy_pipeline/workflows/varfish_export/__init__.py
@@ -87,9 +87,6 @@
 #: Extension of files
 EXTS = (".tsv.gz", ".tsv.gz.md5")
 
-#: Infixes to use for file name generation
-INFIXES = ("gts", "db-infos", "bam-qc")
-
 # TODO: the number of restart times is high because tabix in HTSJDK/Jannovar is flaky...
 
 #: Default configuration for the somatic_variant_calling step
@@ -391,6 +388,8 @@ def _get_output_files_annotate_strucvars(self):
         work_paths = {
             "gts": f"{prefix}.gts.tsv.gz",
             "gts_md5": f"{prefix}.gts.tsv.gz.md5",
+            "feature_effects": f"{prefix}.feature-effects.tsv.gz",
+            "feature_effects_md5": f"{prefix}.feature-effects.tsv.gz.md5",
             "db_infos": f"{prefix}.db-infos.tsv.gz",
             "db_infos_md5": f"{prefix}.db-infos.tsv.gz.md5",
         }

diff --git a/snappy_wrappers/wrappers/mehari/annotate_strucvars/wrapper.py b/snappy_wrappers/wrappers/mehari/annotate_strucvars/wrapper.py
@@ -92,6 +92,25 @@
     tabix -s1 -b2 -e2 -f $TMPDIR/final_for_import.$num.vcf.gz
 done
 
+cat <<"EOF" > $TMPDIR/feature-effects.tsv
+case_id
+set_id
+sv_uuid
+refseq_gene_id
+refseq_transcript_id
+refseq_transcript_coding
+refseq_effect
+ensembl_gene_id
+ensembl_transcript_id
+ensembl_transcript_coding
+ensembl_effect
+EOF
+
+cat $TMPDIR/feature-effects.tsv \
+| tr '\n' '\t' \
+| sed -e 's/\t$/\n/g' \
+>{snakemake.output.feature_effects}
+
 # Perform Mehari structural variant annotation.
 mehari \
     annotate \
@@ -113,6 +132,7 @@
 # Compute MD5 sums on output files
 compute-md5 {snakemake.output.db_infos} {snakemake.output.db_infos_md5}
 compute-md5 {snakemake.output.gts} {snakemake.output.gts_md5}
+compute-md5 {snakemake.output.feature_effects} {snakemake.output.feature_effects_md5}
 
 # Create output links -----------------------------------------------------------------------------