Skip to main content
. 2019 Oct 25;17:1245–1254. doi: 10.1016/j.csbj.2019.09.005

Table 5.

Comparative performance results among different protein function prediction methods.

Techniques Sensitivity Precision Specificity Accuracy MCC AUC
Traditional GRU* 70.8 44 81 79.2 0.44 0.848
BLSTM 54.2 55.8 90.9 84.6 0.46 0.846
BLAST 54.1 52.8 89.8 83.6 0.43 0.82
New GRU** 79.2 48.7 82.9 82.3 0.52 0.861

(* traditional PSSM profiles + GRU, ** our GRU architecture).