This fixes the detection of probe_type of UNIPROT and ACCNUM. Most UNIPROT ids will match both UNIPROT and ACCNUM keys in the orgdb table. Instead of the number of probe matches, we count the number of not-NA SYMBOLS/GENENAME of the annotation to determine the optimal key match (which seems to be ACCNUM because e.g. immunoglobuline genes did not get properly annotation with UNIPROT but did correctly get annotated with ACCNUM).
This fixes the detection of probe_type of UNIPROT and ACCNUM. Most UNIPROT ids will match both UNIPROT and ACCNUM keys in the orgdb table. Instead of the number of probe matches, we count the number of not-NA SYMBOLS/GENENAME of the annotation to determine the optimal key match (which seems to be ACCNUM because e.g. immunoglobuline genes did not get properly annotation with UNIPROT but did correctly get annotated with ACCNUM).