Fig. 8
From: A benchmark study of ab initio gene prediction methods in diverse eukaryotic organisms
Effect of undetermined sequence regions (UDT) on prediction performance of the five gene prediction programs, using Confirmed benchmark sequences from Metazoa, where 542 sequences have no undetermined regions (−UDT: light colors) and 133 sequences have undetermined regions (+UDT: dark colors). a sensitivity and specificity of nucleotide prediction. b sensitivity and specificity of exon prediction c accuracy of protein sequence prediction (% identity) and number of proteins correctly predicted with 100% identity. Sn = sensitivity; Sp = specificity; F1 = F1 score; ME = Missing Exons; WE = Wrong Exons; 5′ First = percentage of correctly predicted 5′ boundaries of first exons only; 3′ Last = percentage of correctly predicted 3′ boundaries of last exons; 3′ and 5′ internal are the percentage of correctly predicted 3′ and 5′ internal exon boundaries. %Identity indicates the sequence identity observed between the predicted proteins and the Confirmed benchmark sequences