The paper by AAi’s Qian Liu, Jinyan Li and Zhenhua Li on Using B-factor-related features for accurate classification between protein binding interfaces and crystal packing contacts won the Best Paper Award (BMC Bioinformatics) at the recent International Conference on Bioinformatics. The rapidly expanding volume of protein 3D structure data is valuable in understanding biological functions. But there are many unannotated crystal contacts and false annotations, and previous tools have low performance. The authors propose to use four features related to B factor for the classification between biological interfaces and crystal packing contacts. We investigate the distribution properties of these features on five datasets of biological binding and/or crystal packing. They also compare the cross-dataset classification performance of these features with existing methods and with a widely-used and the most effective feature interface area. The results demonstrate that their features significantly outperform the interface area approach and the existing prediction methods in many tests on all of these datasets. The computational methods thus have a potential for large-scale and accurate identification of biological interactions from the experimentally determined structural data stored at the Protein DataBank which may have diverse interface sizes. For more information, email Jinyan.li@uts.edu.au
