Can gubbins be used as a replacement for hyphy GARD or 3seq?

[phiweger]

When run on a nucleotide (codon) alignment, is gubbins substantially different from GARD and 3seq? I thought not, but then saw that in this PNAS pneumococci paper the authors of gubbins use 3seq for this:

To reduce any biases introduced by recombination, the codon alignments were scanned for breakpoints. The initial scan was performed with 3SEQ ... --

[nickjcroucher]

Gubbins is very different to GARD and 3seq in how it detects recombination - Gubbins is not appropriate for detecting exchanges between donors and recipients within an alignment of species-wide diversity (as in the PNAS paper), but is instead designed for detecting the import of diversity into a strain.

[phiweger]

Thanks for clarifying @nickjcroucher

[phiweger]

@nickjcroucher do I understand correctly that the --pairwise flag does allow the exchanges between donors and recipients?

[nickjcroucher]

You can use --pairwise to search for recombinations where the recipient sequence is known, as in https://doi.org/10.1101/2022.09.21.508813.