A conformation ensemble approach to protein residue-residue contact

BMC Structural Biology

Table 6 Representation of predicted contact clusters in an ensemble

Target	Num. Clusters	Cluster Coverage (percentage of models from ensemble with stated coverage)
T0534[31-80,257-384]	5	5(.05)	4(.14)	3(.16)	2(.17)
T0534[81-255]	3	3(.11)	2(.24)	1(.23)	0(.40)
T0544[1-135]	7	7(.07)	6(.12)	5(.13)	4(.14)
T0550[178-339]	13	11(.01)	10(.01)	9(.07)	8(.09)
T0561[1-109,112-161]	5	5(.12)	4(.12)	3(.20)	2(.16)
T0571[197-331]	6	6(.03)	5(.08)	4(.10)	3(.16)
T0608[29-117]	5	5(.06)	4(.08)	3(.11)	2(.17)
T0621[2-170]	6	6(.03)	5(.10)	4(.19)	3(.22)

Long range contacts predicted by the conformation ensemble are clustered for a number of CASP9 FM targets. The cluster coverage is the percentage of models in the ensemble that cover a given number of clusters recovered by the ensemble method. A cluster is considered covered by a model if the model contains a contact within 4 residues of the cluster's representative contact. For each target, the cluster coverages are calculated for the top four cluster counts. Num. Clusters (column 2) is the total number of true contact clusters recovered by the ensemble method for the target. Other columns followed list the percent of models in the pool containing a specific number of the true clusters. The results show that the ensemble method can recover more contact clusters even though the proportion of models in the pool having high cluster coverage is very low.

ISSN: 1472-6807