Table 4 Centroid prediction.

From: Deep learning for discriminating non-trivial conformational changes in molecular dynamics simulations of SARS-CoV-2 spike-ACE2

System

\(c_0\)

\(c_1\)

\(c_2\)

\(c_3\)

\(c_4\)

\(p_0\)

\(p_1\)

\(p_2\)

\(p_3\)

\(p_4\)

\({\overline{p}}_w\)

\({\hat{y}}\)

Beta

4364

632

2

2

1

1

1

1

0.998

1

1

\(++\)

Omicrona

3312

1043

640

5

1

0.999

1

1

1

1

0.999

\(++\)

Theta

4947

27

16

10

1

0.999

1

0.999

1

0.999

0.999

\(++\)

Mu

2765

1157

872

200

7

0.997

1

1

0

0.782

0.958

\(++\)

Delta

3571

1418

10

1

1

0.999

0.967

0.999

0

0.983

0.990

\(++\)

Gamma

4601

72

15

  

0.831

0.989

0.998

0

0

0.782

\(++\)

S:Q498I

4501

439

51

1

1

0.292

0.995

0.999

0.999

0.999

0.361

\(- -\)

S:N439R

4006

549

441

4

1

0.077

0.068

0.027

0

0.008

0.072

\(- -\)

S:K417Y

3026

1860

107

4

4

7.65E−05

0

0.999

0.965

1.02E-04

0.022

\(- -\)

S:Y489F

3963

675

265

78

22

0.012

0

0

0

0.4

0.011

\(- -\)

WT

3019

1952

24

5

1

0

0

0

0

0

0

\(- -\)

  1. For each system in the database it is possible to observe the number of frames per cluster, \(c_i\), as well as the predicted probability, \(p_i\),for the corresponding centroids. The last column presents the predicted class, \({\hat{y}}\).
  2. a The Omicron variant is the only one with more than 3 (three) point mutations in the RBD52