Figure 6: Ranking and retrieval performance for the proposed methods. (a) Enrichment plot for screening protein structures using the proposed descriptors. Results are average for 5 query proteins from cell division protein kinase 2 family (Group 3), using DATASET-A (416 protein chains). (b) Average precision and recall for three queries, one for each group in DATASET-A. DD1 corresponds to the distance distribution proposed in [11], as described in Section 2 (see Section 4.5 on comparison with related methods).