Merge pull request #1150 from nf-core/flexible_tx2gene

Be more flexible on attribute values in GTFs

Author: drpatelh

CHANGELOG.md

@@ -17,6 +17,15 @@ Special thanks to the following for their contributions to the release:
 - [Phil Ewels](https://github.com/ewels)
 - [Vlad Savelyev](https://github.com/vladsavelyev)
 
+### Enhancements & fixes
+
+- [PR #1135](https://github.com/nf-core/rnaseq/pull/1135) - Update [action-tower-launch](https://github.com/marketplace/actions/action-tower-launch) to v2 which supports more variable handling
+- [PR #1141](https://github.com/nf-core/rnaseq/pull/1141) - Important! Template update for nf-core/tools v2.11
+- [PR #1143](https://github.com/nf-core/rnaseq/pull/1143) - Move fasta check back to Groovy ([#1142](https://github.com/nf-core/rnaseq/issues/1142))
+- [PR #1144](https://github.com/nf-core/rnaseq/pull/1144) - Interface to kmer size for pseudoaligners
+- [PR #1149](https://github.com/nf-core/rnaseq/pull/1149) - Fix and patch version commands for Fastp, FastQC and UMI-tools modules ([#1103](https://github.com/nf-core/rnaseq/issues/1103))
+- [PR #1150](https://github.com/nf-core/rnaseq/pull/1150) - Be more flexible on attribute values in GTFs ([#1132](https://github.com/nf-core/rnaseq/issues/1132))
+
 ### Parameters
 
 | Old parameter | New parameter                |
@@ -27,14 +36,6 @@ Special thanks to the following for their contributions to the release:
 > **NB:** Parameter has been **added** if just the new parameter information is present.
 > **NB:** Parameter has been **removed** if new parameter information isn't present.
 
-### Enhancements & fixes
-
-- [PR #1135](https://github.com/nf-core/rnaseq/pull/1135) - Update [action-tower-launch](https://github.com/marketplace/actions/action-tower-launch) to v2 which supports more variable handling
-- [PR #1141](https://github.com/nf-core/rnaseq/pull/1141) - Important! Template update for nf-core/tools v2.11
-- [PR #1149](https://github.com/nf-core/rnaseq/pull/1149) - Fix and patch version commands for Fastp, FastQC and UMI-tools modules ([#1103](https://github.com/nf-core/rnaseq/issues/1103))
-- [PR #1144](https://github.com/nf-core/rnaseq/pull/1144) - Interface to kmer size for pseudoaligners
-- [PR #1143](https://github.com/nf-core/rnaseq/pull/1143) - Move fasta check back to Groovy ([#1142](https://github.com/nf-core/rnaseq/issues/1142))
-
 ## [[3.13.2](https://github.com/nf-core/rnaseq/releases/tag/3.13.2)] - 2023-11-21
 
 ### Credits

bin/tx2gene.py

@@ -6,6 +6,7 @@
 import argparse
 import glob
 import os
+import re
 from collections import Counter, defaultdict, OrderedDict
 from collections.abc import Set
 from typing import Dict
@@ -50,14 +51,18 @@ def discover_transcript_attribute(gtf_file: str, transcripts: Set[str]) -> str:
     Returns:
     str: The attribute name that corresponds to transcripts in the GTF file.
     """
+
     votes = Counter()
     with open(gtf_file) as inh:
         # Read GTF file, skipping header lines
         for line in filter(lambda x: not x.startswith("#"), inh):
             cols = line.split("\t")
-            # Parse attribute column and update votes for each attribute found
-            attributes = dict(item.strip().split(" ", 1) for item in cols[8].split(";") if item.strip())
-            votes.update(key for key, value in attributes.items() if value.strip('"') in transcripts)
+
+            # Use regular expression to correctly split the attributes string
+            attributes_str = cols[8]
+            attributes = dict(re.findall(r'(\S+) "(.*?)(?<!\\)";', attributes_str))
+
+            votes.update(key for key, value in attributes.items())
 
     if not votes:
         # Log a warning if no matching attribute is found