Next‐generation sequencing for molecular ecology: a caveat regarding pooled samples