Choice of reference genomes

[LudoPoire]

Hi, I'm looking towards using your program for bacterial GWAS for multiple phenotypes. I was wondering if you had some advice regarding the choice of reference strains (for both associations and annotation). Should I have one or multiple strains, and how to choose them ?

I see you chose IAI39 as "enrichment reference" and multiple E. coli strains with the flags --reference and --focus-strain for bootstraping, could you provide some details about what these parameters are about ? I'm working with a collection of A. baumannii strains fyi.

Thanks for your answer and for developing this tool !

[jburgaya]

Hi Ludo,

Thanks for your interest in our tool. Using multiple reference strains can better capture the genetic diversity and improve the accuracy in the associations. However, in some cases a single well-chosen reference strain-often the reference lab strain- might be enough and can simplify the interpretation of the analyses. For populations with high genetic diversity, such as E. coli, multiple strains are recommended. There's no real suggestion for a good reference really. We would suggest to pick more than one and pick whichever makes sense for the phenotype being studied as the "main" reference.

A single reference strain enrichment_reference is used to perform the enrichment analyses for the associated genetic variants. While multiple strains are used to annotated --focus-strain and summarize --reference the hits.

[LudoPoire]

Hi and thanks for your answer ! So you recommend having multiple strains from different origins/lineages to match the diversity present in a collection ? Let's say one or few lab strains, plus some from the environment and/or clinical strains if the species is present in multiple habitats ?

[jburgaya]

Yes 👍🏽