Benchmark Data Sets for Assessment of Docking Performance
The performance of docking programs and scoring functions is often evaluated using benchmark data sets for redocking of co-crystallised molecules or ligand enrichment against decoy compounds. For drug-like small molecules several data sets are available. However, for assessing docking performance for less conventional compounds high quality benchmark data sets are missing: no ready-to-use benchmark data set for peptides, an emerging class of therapeutics, has been available. Therefore, we generated LEADS-PEP, the first of several “Lessons for Efficiency Assessment of Docking and Scoring” (LEADS) which is publicly available at www.leads-x.org. LEADS-PEP enables the evaluation of docking programs for their potential to reproduce peptide binding modes and to compare different methods and parameters. The collection consists of 53 protein−peptide complexes that have been prepared using a rational and unbiased workflow. The talk will cover the details on the generation of LEADS-PEP and other data sets of LEADS and their utilization to assess the redocking performance of popular molecular docking programs.