Fix UMI barcode extraction to handle FASTQ/FASTA descriptions

[Stefan-JLU]

FASTQ format allows descriptions in the identifier line, separated by whitespace (e.g., @READID_UMI 1:N:0:1). The solexa2fasta.awk script converts the entire identifier line to FASTA format, preserving any descriptions (>@READID_UMI 1:N:0:1).

Previously, make_comp_fasta.pl had two issues when descriptions were present:

1. UMI Detection Regex: The pattern /^[^\t]*_[A-Z]*\t/ expected the UMI barcode to appear immediately before the tab character. When descriptions were present, the line looked like "READID_UMI description\tseq", causing the regex to fail since the description appeared before the tab instead of the UMI pattern.

2. Barcode Extraction: The script split the entire first field (including any description) by '_' to extract the UMI. With descriptions containing underscores or other text, this extracted the wrong value as the barcode.

This fix addresses both issues:

• Updated the UMI detection regex to /^\S*_[A-Z]+\s/, which matches the UMI pattern before any whitespace (space or tab)
• Modified barcode extraction to first strip the description by taking only the portion before the first whitespace, then splitting by '_'

Tested with multiple cases:

• UMI with/without descriptions
• Descriptions with/without underscores
• Read IDs with/without additional underscores
• No UMI cases (to ensure no false positives)

All of my test cases correctly detect UMI barcodes regardless of whether FASTQ/FASTA descriptions are present.