Molecular epidemiology: New rules for new tools?