A new decoding algorithm for hidden Markov models improves the prediction of the topology of all-beta membrane proteins

Table 2 PV accuracy compared with other algorithms and HMM models

Method	Q ₂	SOV	SOV(BetaTM)	SOV(Loop)	Q _ok
cross-validation
Posterior-Viterbi^l	0.82	0.87	0.92	0.81	0.80
Viterbi¹	0.63	0.33	0.27	0.35	0.0
1-best¹	0.65	0.41	0.37	0.41	0.07
HMMB2HTMR²	0.83	0.87	0.88	0.84	0.73
PROFTmb³	0.83	0.87	0.88	0.84	0.73
pred-tmbb⁴ (Viterbi)	0.78	0.83	0.81	0.82	0.60
pred-tmbb⁴ (1-best)	0.78	0.83	0.81	0.82	0.60
pred-tmbb⁴ (posterior)	0.78	0.82	0.80	0.82	0.60
blind-test
Posterior-Viterbi¹	0.80	0.81	0.84	0.74	0.60
Viterbi¹	0.62	0.38	0.35	0.40	0.20
1-best¹	0.63	0.38	0.36	0.40	0.20
HMMB2HTMR²	0.80	0.81	0.84	0.74	0.60
PROFTmb³	0.72	0.65	0.72	0.58	0.40
pred-tmbb⁴ (Viterbi)	0.71	0.73	0.79	0.71	0.20
pred-tmbb⁴ (1-best)	0.71	0.73	0.79	0.71	0.20
pred-tmbb⁴ (posterior)	0.72	0.75	0.81	0.71	0.20

ISSN: 1471-2105