In the Statistics table for any ontology, the prefix for one ontology may appear in upper or lower case, and sometimes a lower-case "prefix" entry has an extension or otherwise differs from the usual prefix.
For example, on the FYPO statistics page (where we first noticed the issue), the Class column has non-zero entries for prefixes including 'CHEBI' and 'SO', whereas the same ontologies have ObjectProperty or AnnotationProperty entries using 'chebi' and 'so-xp.obo' respectively. Similarly, 'PATO' is used for Class entries, but 'pato' for ObjectProperty and AnnotationProperty entries.
Furthermore, in rows with lower-case prefix entries, non-zero numbers link to pages that say "0 terms(s) returned", e.g. the '20' in the 'go' row, AnnotationProperty cell. (I didn't test all non-zero links, but all of the ones I tested in rows with lower-case prefixes give 0 results.)
In the Statistics table for any ontology, the prefix for one ontology may appear in upper or lower case, and sometimes a lower-case "prefix" entry has an extension or otherwise differs from the usual prefix.
For example, on the FYPO statistics page (where we first noticed the issue), the Class column has non-zero entries for prefixes including 'CHEBI' and 'SO', whereas the same ontologies have ObjectProperty or AnnotationProperty entries using 'chebi' and 'so-xp.obo' respectively. Similarly, 'PATO' is used for Class entries, but 'pato' for ObjectProperty and AnnotationProperty entries.
Furthermore, in rows with lower-case prefix entries, non-zero numbers link to pages that say "0 terms(s) returned", e.g. the '20' in the 'go' row, AnnotationProperty cell. (I didn't test all non-zero links, but all of the ones I tested in rows with lower-case prefixes give 0 results.)
@cmungall @ValWood