Fig. 7. Prediction of the oligomeric state.

a The average maximum predicted contact probabilities for n = 50 monomers from the PDBTM database and n = 52 oligomers in the TMP test set. The vertical bars represent the mean, and the lines are error bars (standard deviations). The outliers are plotted as circle. b The average maximum predicted contact probabilities for different oligomeric states for n = 52 transmembrane protein complexes in the test set. The vertical bars represent the mean, and the lines are error bars (1.5 times standard error). c The success rates in predicting the oligomeric states of monomer versus oligomer, and dimer versus oligomer>2 as a function of contact probability threshold. d The precisions for oligomers (solid lines) and the false positive rates for monomers (dashed lines) as a function of contact probability threshold.