KeyError 'W' during ADRSM simulation; inconsistent behavior across FASTA vs FASTA.GZ

[ryfitzwater]

Hello, I am using ADRSM (v0.9.5) to simulate ancient metagenomic datasets, and I’ve encountered a reproducible error:
KeyError: 'W'
This occurs in:
sequencefunctions.py -> add_error_fwd() / add_error_rev() -> quality.qdict[phred[j]]

I’ve tested across multiple genomes and parameters, including both uncompressed and gzipped FASTA inputs.

• FASTA inputs were cleaned (ACGTN only) and confirmed as ASCII
• Issue persists with --threads 1
• Failure is not explained solely by read count or parameters
• Gzipping rescues some runs, but not others

Do you know what might cause a 'W' character to appear in the phred string during error simulation, or if this is a known issue?

Happy to share configs if helpful.

Thanks!