Performance optimization for mobile og pl

[DrKaiyangZheng]

Enhanced CLI & Core Performance Optimization

Changes

1. CLI Enhancements (mobileOGs-pl-kyanite.sh & mobileOGs-pl-amethyst.sh)

Feature	Description
-h, --help	Added usage help. Previously -h caused silent exit.
-o, --out	Redirect all outputs (.faa, .tsv, .csv) to a custom directory. Auto-creates directory.
--faa	Accept pre-computed protein FASTA (e.g. from prodigal). Skips gene prediction step when provided.
Input validation	-i/--input is now mandatory; script exits with error message if missing.

2. Core Performance Optimization (mobileOGs-pl-kyanite.py)

Bottleneck	Fix	Impact
Row-wise apply(lambda r: '_'.join(...))	str.rsplit('_', n=1).str[0]	Eliminates Python-level loop; ~10–100× speedup on large datasets
7× redundant str.split('|') on same column	Single split(..., expand=True) + multi-column assignment	Reduces string scanning overhead
ORF counting via copy → drop_duplicates → groupby transform → drop_duplicates	Direct groupby(...).nunique()	Fewer DataFrame copies and passes
Nested for loops doing replace()	Flat dictionary + vectorized replace()	Entirely removes Python iteration over categories
Redundant merge(left_on=Series) syntax	Clean left_on="column"	Avoids spurious key_0 columns

3. Documentation (UsageGuidance.md)

• Added "April 2026 Update" section documenting all new options.
• Included usage examples for standard run, custom output directory, and --faa.

Backward Compatibility

• Default behavior is preserved. Omitting -o writes to the current directory; omitting --faa runs prodigal as before.
• Output values are numerically equivalent for standard prodigal-formatted input. The only intentional CSV-level changes are:
  • Added index=False to to_csv() calls (removes unwanted index column).
  • Percentage columns return 0.0 instead of NaN when Total Number of Hits = 0.

Testing

# Help
./mobileOGs-pl-kyanite.sh -h

# Standard run
./mobileOGs-pl-kyanite.sh -i test.fasta -d db.dmnd -m meta.csv -k 15 -e 1e-20 -p 90 -q 90

# Custom output + skip prodigal
./mobileOGs-pl-kyanite.sh -i test.fasta --faa test.faa -o ./results -d db.dmnd -m meta.csv -k 15 -e 1e-20 -p 90 -q 90