Table 3.

Comparison of contact predictors on top L/5 predictions for CASP9 FM targets

Prediction Methods	Medium range contacts	Long range contacts
Conformation ensemble	.34	.30
SVMcon	.19	.19
BAKER-ROSETTASERVER ensemble	.27	.20
Zhang-Server ensemble	.28	.23

The precision of predicted contacts obtained by various contact prediction methods. For our conformation ensemble, we used sets of server submitted tertiary structure predictions for each FM target during CASP9. SVMcon is a machine learning, sequence based contact prediction methods. BAKER-ROSETTASERVER ensemble and Zhang-Server ensemble were made by applying the conformation ensemble approach to the structural predictions made by each predictor during CASP9. L is the sequence length of each target domain.