Is next-generation sequencing a realistic approach for inferring an antibiogram from genomic data today? Tomorrow? for all bacteria/antibiotics?