split alleles before merge_duplicates

src/cactus/refmap/cactus_graphmap_join.py

@@ -1130,17 +1130,21 @@ def vcfwave(job, config, out_name, vcf_ref, vcf_id, tbi_id, max_ref_allele, fast
                    ['bcftools', 'annotate', '-x', 'INFO/AT'],                   
                    ['vcfwave'] + wave_opts.split(' ') + ['-t', str(job.cores)]]
     run_norm = getOptionalAttrib(findRequiredNode(config.xmlRoot, "graphmap_join"), "vcfwaveNorm", typeFn=bool, default=True)
+    merge_duplicates_opts = getOptionalAttrib(findRequiredNode(config.xmlRoot, "graphmap_join"), "mergeDuplicatesOptions", typeFn=str, default=None)
+    run_merge = merge_duplicates_opts is not None and merge_duplicates_opts != "0" and run_norm    
     if run_norm:
         fa_ref_path = os.path.join(work_dir, tag + 'fa.gz')
         job.fileStore.readGlobalFile(fasta_ref_dict[vcf_ref], fa_ref_path)
         bubwave_cmd.append(['bcftools', 'norm', '-f', fa_ref_path])
         bubwave_cmd.append(['bcftools', 'sort', '-T', os.path.join(work_dir, 'bcftools.XXXXXX')])
+        if run_merge:
+            #merge_duplicates.py want biallelic
+            bubwave_cmd.append(['bcftools', 'norm', '-m', '-any'])
     bubwave_cmd.append(['bgzip'])
 
     cactus_call(parameters=bubwave_cmd, outfile=vcfwave_path)
 
-    merge_duplicates_opts = getOptionalAttrib(findRequiredNode(config.xmlRoot, "graphmap_join"), "mergeDuplicatesOptions", typeFn=str, default=None)
-    if merge_duplicates_opts and merge_duplicates_opts != "0" and run_norm:
+    if run_merge:
         #note: merge_duplcates complains about not having a .tbi but I don't think it actually affects anything
         merge_path = os.path.join(work_dir, os.path.basename(out_name) + '.' + vcf_ref + tag + 'wave.merge.vcf.gz')
         cactus_call(parameters=['merge_duplicates.py', '-i', vcfwave_path, '-o', merge_path] + merge_duplicates_opts.split(' '))