Description of the bug

There is a logic error in find_unique_reads.py where duplicate read1s are not filtered correctly.

Expected behaviour:
When multiple alignments share the same read1, only the alignment with the highest MAPQ score should be retained.

Observed behaviour:
In cases of duplicates, the alignment with the lowest MAPQ is kept instead.

Cause:
In lines 91–92:

• alignments[unique_id][1] is the stored MAPQ.
• row[4] is the new alignment’s MAPQ.
• The current comparison replaces the stored alignment if the stored MAPQ is higher, effectively keeping the lower-scoring alignment.

Fix:
The comparison should be reversed ensuring the alignment with the highest MAPQ is kept.

Command used and terminal output

NA

Relevant files

NA

System information

NA