question "assembly track"

[evezeyl]

If I understand well your SeqSero2 article,
You use k-mers identification for the draft assemblies input.
Would it not provide better accuracy to use the kmers to locate the candidate alleles on the draft assembly and then run the same pipeline as the micro-assembly track with blast for serotyping purpose. ?

I am a bit wondering why (Is there a rational I do not see?) you did not choose this solution, and therefore wondering if I did not understand the assembly method correctly.

All the best
Eve