Predicting β-Turns in Protein Using Kernel Logistic Regression

Table 2

Comparison of KLR with other recent 𝛽-turns prediction methods on BT426 dataset.

Method	𝑄total	𝑄pred	𝑄obs	Specificity	MCC
KLR	80.7	58.98	65.25	85.34	0.50
BTNpred [6]	80.9	62.7	55.6	N/A	0.47
NetTurnP [13]	78.2	54.4	75.6	79.1	0.50
BetaTPred2 [10]	75.5	49.8	72.3	N/A	0.43
BTPRED [9]	74.9	55.3	48.0	N/A	0.35
DEBT [8]	79.2	54.8	70.1	N/A	0.48
SVM [26]	79.8	55.6	68.9	N/A	0.47
BTSVM [27]	78.7	56.0	62.0	N/A	0.45
E-SSpred [28]	80.9	63.6	49.2	N/A	0.44
1–4 & 2-3 correlation model [29]	59.1	32.4	61.9	N/A	0.17