Gene Prediction in Metagenomic Fragments with Deep Learning

Table 3

The overall accuracy (%) of Orphelia, FragGeneScan, MGC, MetaGUN, Meta-MFSVM, and Meta-MFDL on the 120 independent testing dataset.

Species	Orphelia	Meta-MFSVM	MGC	FragGeneScan	MetaGUN	Meta-MFDL
A. fulgidus	84.05	85.22	85.67	85.92	86.49	87.38
N. pharaonis	82.43	82.79	82.96	82.94	83.12	83.03
M. jannaschii	82.11	84.25	84.69	84.80	85.23	85.59
B. aphidicola	85.91	87.50	88.21	88.96	89.96	90.06
B. pseudomallei	84.28	84.73	84.93	85.10	85.37	85.60
B. subtilis	88.13	88.36	88.46	88.51	88.60	87.85
C. jeikeium	80.95	83.19	84.64	85.96	87.38	91.04
C. tepidum	79.17	79.67	80.02	80.16	80.52	79.22
E. coli	85.98	86.31	86.56	86.70	86.89	86.91
H. pylori	88.09	89.37	90.66	91.42	92.51	92.81
P. aeruginosa	83.71	84.28	84.57	84.84	85.15	84.67
P. marinus	88.26	88.76	88.98	89.11	89.71	88.73
W. endosymbiont	74.75	76.50	77.72	78.51	79.23	79.55
Average	83.68	84.69	85.24	85.61	86.17	86.34