Skip to main content

Table 1 Performance of HMM-d, HMM-t, HMM-ModE and HMM-Sub for the sub-family classification of the AGC family of kinases.

From: HMM-ModE – Improved classification using profile hidden Markov models by optimising the discrimination threshold and modifying emission probabilities with negative training sequences

Methods

HMM-d

HMM-T

HMM-ModE

HMM-Sub

Sub-groups of AGC kinases

Sensitivity

Specificity

Sensitivity

Specificity

Sensitivity

Specificity

Sensitivity

Specificity

GRK

1

(0)

0.27

(0.03)

1

(0)

1

(0)

1

(0)

1

(0)

1

(0)

1

(0)

PKA

0.96

(0.1)

0.18

(0.02)

0.89

(0.13)

1

(0)

0.89

(0.13)

1

(0)

0.96

(0.07)

1

(0)

PKC

0.99

(0.05)

0.42

(0.1)

0.95

(0.08)

0.99

(0.2)

0.96

(0.08)

1

(0)

0.97

(0.06)

1

(0)

PVPK

1

(0)

0.17

(0.03)

0.94

(0.1)

0.93

(0.13)

0.94

(0.1)

0.96

(0.12)

0.96

(0.08)

1

(0)

RAC

1

(0)

0.09

(0.00)

0.9

(0.16)

1

(0)

0.93

(0.1)

1

(0)

0.97

(0.1)

1

(0)

S6PK

1

(0)

0.14

(0.01)

0.98

(0.08)

0.98

(0.06)

0.975

(0.08)

0.98

(0.06)

0.93

(0.24)

0.98

(0.06)

  1. The values in parentheses indicate the standard deviation for all 10 samples in a 10-fold cross validation.
  2. HMM-d – HMM profile used with default threshold
  3. HMM-t – HMM profile used with optimised threshold
  4. HMM-ModE – profile with modified emission probabilities
  5. HMM-Sub – Log-difference-of-odds-score method