Fig 1.
Evaluation of the derived prediction models.
A) T-statistics calculated for the derived prediction models with 1, 2, …, 100 estimators w.r.t. the prediction model with 50 estimators. B) Performance metrics on the test set for the best 25 prediction models in terms of t-statistics.
Table 1.
Benchmark results of the prediction model.
Fig 2.
Distribution of point mutations across proteins.
Black and light gray bars correspond to the total number of point mutations and the number of point mutations correctly predicted by the optimal prediction model, respectively. Dark grey bars correspond to the number of correctly predicted point mutations in the leave-one-out validation (LOO).
Table 2.
Comparison of BorodaTM and Enterprise predictions for point mutations in the transmembrane and soluble regions.