PanSN naming issue in pggb fastas

Question

[Whirlingdaf]

I am having an issue with the PanSN naming for pggb fastas and would love thoughts!

I am running:

fastix -p "NP117#hap#" 117_Assembly.fasta > 117_shasta_fastix_output.fasta

But run into naming issues with:

pggb -i combined_aligned_contigs.fasta.gz -o pangenome

[pggb] warning: there are sequence names (like 'NP117#hap#178') that do not match the Pangenome Sequence Naming (PanSN). [pggb] ERROR: -n/--n-haplotypes must be greater than or equal to 1 when the Pangenome Sequence Naming (PanSN) is not respected.

Unsure what the issue is as the fastix example is:

fastix -p "gen#1#" genome.fa > genome_prefixed.fa

Any thoughts?

Answer 1

[cmdcolin]

looks like the naming scheme pggb wants has an integer in between those two #, so 1 is allowed but hap is not

the regex it checks appears to be ^([^#]+#)[0-9]+#[^#]+$ (from https://github.com/pangenome/pggb/blob/cc332526727a9cc99f4194ef47212e7c06175106/pggb#L312C20-L312C42)

gemini explains the regex as follows

This regex aims to match strings that:

• Start with any sequence of characters followed by a single "#".
• Continue with one or more digits.
• Followed by another single "#".
• End with any sequence of characters.

this confirms what https://github.com/pangenome/PanSN-spec says: the haplotype id is the middle part which is a number

Answer 2

[colindaven]

Exactly what cmdcolin said.

You can check your fasta headers using

samtools faidx 117_Assembly.fasta

cat 117_Assembly.fasta.fai