If PLOS ONE were really 101 different specialized journals: A proposed approach to the evaluation of multidisciplinary megajournals