Functional annotation returned by CELLO2GO for the Gram-positive bacteria dataset PS30GP.

1The percentage of homologous sequences for which GO functional annotations were not found by a BLAST search of the in-house database derived from the UniProtKB/SwissProt database for bacteria. 2The percentage of homologous sequences for which GO functional annotations were not found by a BLAST search of the in-house database derived from the UniProtKB/TrEMBL database for bacteria. 3The percentage of entries for which GO annotations for cellular components were missing or homologs were not retrieved by BLAST searching of the UniProtKB/TrEMBL databases, but for which CELLO accurately predicted the subcellular localization(s). 4The Gram-positive bacterial benchmark dataset found in PSORTb3.0 [23], denoted PS30GP, includes 2576 protein sequences in four subcellular categories: extracellular, cell wall, membrane, and cytoplasmic.