diff --git a/CHANGELOG.md b/CHANGELOG.md
index 8611899..3321011 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -5,6 +5,8 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
 ## [Unreleased]
+### Fixed
+- More informative error message upon read duplicate detection.
 ### Updated
 - Remove duplicate fastcat call.
 
diff --git a/bin/workflow_glue/cluster_umis.py b/bin/workflow_glue/cluster_umis.py
index c879209..619fdd0 100755
--- a/bin/workflow_glue/cluster_umis.py
+++ b/bin/workflow_glue/cluster_umis.py
@@ -212,7 +212,14 @@ def process_records(df, args):
 
 def main(args):
     """Run entry point."""
-    df_tags = pd.read_csv(args.read_tags, sep='\t', index_col=0)
+    df_tags = pd.read_csv(args.read_tags, sep='\t', index_col='read_id')
+
+    dups = df_tags[df_tags.index.duplicated(keep='first')]
+    if not dups.empty:
+        raise ValueError(
+            f"One or more input reads are duplicated, please rectify.\n"
+            f"Duplicated reads: {list(set(dups.index))[:20]}")
+
     df_features = pd.read_csv(
         args.feature_assigns, sep='\t', index_col=0)
     # Merge genes and transcripts onto tags.