diff --git a/CGAT/IndexedFasta.py b/CGAT/IndexedFasta.py
index 647cd7b0..33648f7a 100644
--- a/CGAT/IndexedFasta.py
+++ b/CGAT/IndexedFasta.py
@@ -1051,7 +1051,6 @@ def getSequence(self,
                     last_pos, lsequence - first_pos
 
         sequence = self.mDatabaseFile.fetch(contig, first_pos, last_pos)
-
         if str(strand) in ("-", "0", "-1"):
             try:
                 # works in py2 only
diff --git a/CGAT/scripts/bam_vs_bed.py b/CGAT/scripts/bam_vs_bed.py
index ce4d2e43..1b1753e7 100644
--- a/CGAT/scripts/bam_vs_bed.py
+++ b/CGAT/scripts/bam_vs_bed.py
@@ -1,5 +1,4 @@
-'''
-bam_vs_bed.py - count context that reads map to
+'''bam_vs_bed.py - count context that reads map to
 ======================================================
 
 :Tags: Genomics NGS Intervals BAM BED Counting
@@ -17,6 +16,19 @@
 in the :term:`bed` file can be overlapping - they are counted
 independently.
 
+Note that duplicate intervals will be counted multiple times. This
+situation can easily arise when building a set of genomic annotations
+based on a geneset with alternative transcripts. For example::
+
+   chr1     10000     20000     protein_coding            # gene1, transrcipt1
+   chr1     10000     20000     protein_coding            # gene1, transcript2
+
+Any reads overlapping the interval chr1:10000-20000 will be counted
+twice into the protein_coding bin by bedtools. To avoid this, remove any
+duplicates from the :term:`bed` file::
+
+   zcat input_with_duplicates.bed.gz | cgat bed2bed --merge-by-name | bgzip > input_without_duplicates.bed.gz
+
 This scripts requires bedtools_ to be installed.
 
 Options