diff --git a/CHANGES.md b/CHANGES.md
index 1cfe73a27..54b14f8d2 100644
--- a/CHANGES.md
+++ b/CHANGES.md
@@ -9,6 +9,7 @@
      * Incompatible arguments are now checked, especially related to VCF vs FASTA inputs. 
      * `--vcf-reference` and `--root-sequence` are now mutually exclusive.
 * translate: Tree nodes are checked against the node-data JSON input to ensure sequences are present. [#1348][] (@jameshadfield)
+* translate: The 'source' ID for GFF files is now ignored as a potential gene feature. [#1348][] (@jameshadfield)
 * translate: Improvements to command line arguments.  [#1348][] (@jameshadfield)
     * `--tree` and `--ancestral-sequences` are now required arguments.
     * separate VCF-only arguments into their own group
diff --git a/augur/translate.py b/augur/translate.py
index 1fc73040f..8c7253c39 100644
--- a/augur/translate.py
+++ b/augur/translate.py
@@ -129,7 +129,7 @@ def translate_feature(aln, feature):
     return translations
 
 
-def translate_vcf_feature(sequences, ref, feature):
+def translate_vcf_feature(sequences, ref, feature, feature_name):
     '''Translates a subsequence of input nucleotide sequences.
 
     Parameters
@@ -168,7 +168,7 @@ def str_reverse_comp(str_seq):
     # Need to get ref translation to store. check if multiple of 3 for sanity.
     # will be padded in safe_translate if not
     if len(refNuc)%3:
-        print("Gene length of {} is not a multiple of 3. will pad with N".format(feature.qualifiers['Name'][0]), file=sys.stderr)
+        print(f"Gene length of {feature_name!r} is not a multiple of 3. will pad with N", file=sys.stderr)
 
     ref_aa_seq = safe_translate(refNuc)
     prot['reference'] = ref_aa_seq
@@ -409,13 +409,16 @@ def run(args):
     print("Read in {} features from reference sequence file".format(len(features)))
 
     ## Read in sequences & for each sequence translate each feature _except for_ the source (nuc) feature
+    ## Note that `load_features` _only_ extracts {'gene', 'source'} for GFF files, {'CDS', 'source'} for GenBank.
     translations = {}
     if is_vcf:
         (sequences, ref) = sequences_vcf(args.vcf_reference, args.ancestral_sequences)
         features_without_variation = []
         for fname, feat in features.items():
+            if feat.type=='source':
+                continue
             try:
-                translations[fname] = translate_vcf_feature(sequences, ref, feat)
+                translations[fname] = translate_vcf_feature(sequences, ref, feat, fname)
             except NoVariationError:
                 features_without_variation.append(fname)
         if len(features_without_variation):
diff --git a/tests/functional/translate/cram/translate-with-gff-and-locus-tag.t b/tests/functional/translate/cram/translate-with-gff-and-locus-tag.t
index f0d25bee4..1de2856ee 100644
--- a/tests/functional/translate/cram/translate-with-gff-and-locus-tag.t
+++ b/tests/functional/translate/cram/translate-with-gff-and-locus-tag.t
@@ -15,7 +15,7 @@ Translate amino acids for genes using a GFF3 file where the gene names are store
   >   --output-node-data aa_muts.json \
   >   --alignment-output translations.vcf \
   >   --vcf-reference-output translations_reference.fasta
-  Gene length of rrs_Rvnr01 is not a multiple of 3. will pad with N
+  Gene length of 'rrs' is not a multiple of 3. will pad with N
   Read in 187 specified genes to translate.
   Read in 187 features from reference sequence file
   162 genes had no mutations and so have been be excluded.