Table 2 Performance of the Multivariate Bayesian annotation method. Re-annotation of the filtered ENZYME database with the multivariate Bayesian method. Since we systematically sample 10 enzymes to calculate the probabilities for a protein to belong to each functional class (See Different strategies of annotation), probabilities can only take one of the following eleven values: 0, 0.1, ..., 0.9, 1. We report for each assignment probability level and globally the number of correct annotations, the number of annotation errors and the corresponding error rate and coverage of the database.
From: Probabilistic annotation of protein sequences based on functional classifications
Multivariate Bayesian method | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
Assignment probability | 0 | 0.1 | 0.2 | 0.3 | 0.4 | 0.5 | 0.6 | 0.7 | 0.8 | 0.9 | 1 | TOT |
Correct annotations | 0 | 0 | 0 | 0 | 9 | 35 | 109 | 116 | 188 | 511 | 26898 | 27866 |
Annotation errors | 0 | 0 | 0 | 5 | 10 | 37 | 34 | 37 | 29 | 17 | 53 | 222 |
Error rate (%) | - | - | - | 100.0 | 52.6 | 51.4 | 23.4 | 24.2 | 13.4 | 3.2 | 0.20 | 0.79 |
Coverage (%) | 0.0 | 0.0 | 0.0 | 0.0 | 0.1 | 0.3 | 0.5 | 0.5 | 0.8 | 1.9 | 96.0 | 100 |