Merge pull request #98 from cidgoh/MZA_dev

Minor updates
cidgoh · Apr 15, 2022 · f18b6d0 · f18b6d0
2 parents 69ddff1 + a287ea0
commit f18b6d0
Show file tree

Hide file tree

Showing 3 changed files with 9 additions and 6 deletions.
diff --git a/bin/merge_pangolin_metadata.py b/bin/merge_pangolin_metadata.py
@@ -42,9 +42,12 @@ def write_metadata(dataframe):
     metadata_df = pd.read_csv(args.metadata, sep="\t")
     pangolin_df = pd.read_csv(args.pangolin)
 
-    merged_df = pd.merge(metadata_df, pangolin_df, left_on='isolate',
+    metadata_df['fasta header name'] = metadata_df['fasta header name'].str.strip()
+    pangolin_df['taxon'] = pangolin_df['taxon'].str.strip()
+
+    merged_df = pd.merge(metadata_df, pangolin_df, left_on='fasta header name',
                          right_on='taxon')
     merged_df = merged_df.rename(columns={"lineage": "pango_lineage",
-                                          "isolate": "strain"})
+                                          "fasta header name": "strain"})
 
     write_metadata(dataframe=merged_df)
diff --git a/main.nf b/main.nf
@@ -151,10 +151,11 @@ workflow {
               Channel.fromPath( "$params.userfile", checkIfExists: true)
                 .set{ ch_vcf }
             }
-            else if (input_file.getExtension() == "tsv"){
+            //else if (input_file.getExtension() == "tsv"){
               //add module to change tsv to vcf
-              ch_vcf=Channel.empty()
-            }
+            //  Channel.fromPath( "$params.userfile", checkIfExists: true)
+            //    .set{ ch_vcf }
+            //}
             if (!params.skip_annotation) {
 
               ch_stats=ch_refgff

diff --git a/workflows/covidmvp_annotation.nf b/workflows/covidmvp_annotation.nf
@@ -32,7 +32,6 @@ workflow annotation {
       if(params.mode == 'reference'){
         vcfTogvf.out.gvf
               .collect()
-              .unique()
               .set{ ch_gvf_surv }
       }
       else{