compare_gtf_to_ER_01ksb.py crashes when gene_id contains substring "ER"

[netanyak]

Bug: compare_gtf_to_ER_01ksb.py crashes when gene_id contains substring "ER"

Summary

utilities/compare_gtf_to_ER_01ksb.py raises a ValueError during ER sorting if any gene_id includes the substring "ER" (e.g., CHAER1). The script constructs ER labels as "{gene_id}:ER{n}", but the sorting key uses split("ER"), which breaks when "ER" appears earlier in the gene_id.

Steps to reproduce

1. Ensure the ER GTF contains a gene whose gene_id includes "ER", e.g.:

chr4	TranD	exon	56697216	56697886	.	-	.	transcript_id "CHAER1"; gene_id "CHAER1";

2. Run the script:

python utilities/compare_gtf_to_ER_01ksb.py \
  -i <input.gtf> \
  -er <er.gtf> \
  -o <outdir>

Expected behavior

Script completes successfully and writes the output CSVs:

• *_infoERP.csv
• *_flagER.csv

Actual behavior

Script crashes while building geneDct:

ValueError: invalid literal for int() with base 10: '1:'

Where it fails

The error occurs at the ER sorting step:

geneDct = dict(geneERDf.groupby('gene_id').apply(
    lambda x: sorted(set(x['ER']), key=lambda x: int(x.split("ER")[1]))))

When gene_id = "CHAER1", the script generates ER IDs like CHAER1:ER1.
Then "CHAER1:ER1".split("ER") returns ["CHA", "1:", "1"], so split("ER")[1] == "1:", and int("1:") raises ValueError.

Cause

The ER parsing assumes the first "ER" in the string is the ER suffix delimiter, but "ER" can appear inside gene_id. Using split("ER") is not robust for the constructed ER label format.