fix remaining snpEff ANN parsing bugs

broadinstitute · Mar 12, 2015 · 3f83a45 · 3f83a45
1 parent 72f37ce
commit 3f83a45
Show file tree

Hide file tree

Showing 2 changed files with 14 additions and 7 deletions.
diff --git a/intrahost.py b/intrahost.py
@@ -678,6 +678,12 @@ def parse_ann(ann_field, alleles, transcript_blacklist=set(('GP.2','GP.3'))):
         ziplong = itertools.zip_longest if 'zip_longest' in dir(itertools) else itertools.izip_longest
         outalleles = []
         for a in alleles[1:]:
+            if len(a)==len(alleles[0]):
+                out_a = []
+                for ref, alt in reversed(list(zip(alleles[0], a))):
+                    if ref!=alt or out_a:
+                        out_a.append(alt)
+                a = ''.join(reversed(out_a))
             out_a = []
             for ref, alt in ziplong(alleles[0], a, fillvalue=''):
                 if ref!=alt or out_a:
@@ -698,12 +704,13 @@ def parse_ann(ann_field, alleles, transcript_blacklist=set(('GP.2','GP.3'))):
     if not effs:
         return {}
     if len(effs) != len(effs_dict):
-        raise Exception()
-    if len(effs) != len(set(alleles)):
-        raise Exception()
+        raise Exception("ANN field has non-unique alleles")
     for a in alleles:
         if a not in effs_dict:
-           raise Exception("missing allele: " + a)
+           raise Exception("ANN field is missing ALT allele: " + a)
+    if len(effs) != len(set(alleles)):
+        raise Exception("ANN field has %s entries, but ALT field has %s unique alleles: %s" % (
+            len(effs), len(set(alleles)), ','.join(alleles)))
 
     out = {}
     for k in ('eff_type', 'eff_codon_dna', 'eff_aa', 'eff_aa_pos', 'eff_prot_len', 'eff_gene', 'eff_protein'):

diff --git a/pipes/rules/intrahost.rules b/pipes/rules/intrahost.rules
@@ -35,8 +35,8 @@ rule isnvs_vcf:
                 sample=read_samples_file(config["samples_assembly"]))
     output:
             config["dataDir"]+'/'+config["subdirs"]["intrahost"] +'/isnvs.vcf.gz',
-            config["dataDir"]+'/'+config["subdirs"]["intrahost"] +'/isnvs.annot.vcf.gz'
-            #config["dataDir"]+'/'+config["subdirs"]["intrahost"] +'/isnvs.annot.txt.gz'
+            config["dataDir"]+'/'+config["subdirs"]["intrahost"] +'/isnvs.annot.vcf.gz',
+            config["dataDir"]+'/'+config["subdirs"]["intrahost"] +'/isnvs.annot.txt.gz'
     resources: mem=4
     params: LSF=config.get('LSF_queues', {}).get('short', '-W 4:00'),
             logid="all",
@@ -51,5 +51,5 @@ rule isnvs_vcf:
                 + " --strip_chr_version"
                 )
             shell("{config[binDir]}/interhost.py snpEff {output[0]} {params.snpEff_ref} {output[1]}")
-            #shell("{config[binDir]}/intrahost.py iSNV_table {output[1]} {output[2]}")
+            shell("{config[binDir]}/intrahost.py iSNV_table {output[1]} {output[2]}")