MOTIVATION: Protein-protein interactions (PPIs) are pivotal for many biological processes and similarity in Gene Ontology (GO) annotation has been found to be one of the strongest indicators for PPI. Most GO-driven algorithms for PPI inference combine machine learning and semantic similarity techniques. We introduce the concept of inducers as a method to integrate both approaches more effectively, leading to superior prediction accuracies. RESULTS: An inducer (ULCA) in combination with a Random Forest classifier compares favorably to several sequence-based methods, semantic similarity measures and multi-kernel approaches. On a newly created set of high-quality interaction data, the proposed method achieves high cross-species prediction accuracies (Area under the ROC curve ≤ 0.88), rendering it a valuable companion to sequence-based methods. AVAILABILITY: Software and datasets are available at http://bioinformatics.org.au/go2ppi/ CONTACT: [email protected].

Abstract Motivation: Phylogenetic profiling methods can achieve good accuracy in predicting protein–protein interactions, especially in prokaryotes. Recent studies have shown that the choice of reference taxa (RT) is critical for accurate prediction, but with more than 2500 fully sequenced taxa publicly available, identifying the most-informative RT is becoming increasingly difficult. Previous studies on the selection of RT have provided guidelines for manual taxon selection, and for eliminating closely related taxa. However, no general strategy for automatic selection of RT is currently available. Results: We present three novel methods for automating the selection of RT, using machine learning based on known protein–protein interaction networks. One of these methods in particular, Tree-Based Search, yields greatly improved prediction accuracies. We further show that different methods for constituting phylogenetic profiles often require very different RT sets to support high prediction accuracy. Availability: The datasets and software used in the experiments can be found at http://users-birc.au.dk/zxr/phyloprof/ Contact: [email protected]; [email protected] Supplementary information: Supplementary data are available at Bioinformatics online.

The objective in molecular docking is to determine the best binding mode of two molecules in silico. A common application of molecular docking is in drug discovery where a large number of ligands are docked against a protein to identify potential drug candidates. This is a computationally intensive problem especially if flexibility of the molecules are taken into account. In this paper we show how MolDock, which is a high accuracy method for flexible molecular docking using a variant of differential evolution, can be parallelised on both CPU and GPU. The methods presented for parallelising the workload result in an average speedup of 3.9x on a 4-core CPU and 27.4x on a comparable CUDA enabled GPU when docking 133 ligands of different sizes. Furthermore, the presented parallelisation schemes are generally applicable and can easily be adapted to other common flexible docking methods.

The neighbour-joining method reconstructs phylogenies by iteratively joining pairs of nodes until a single node remains. The criterion for which pair of nodes to merge is based on both the distance between the pair and the average distance to the rest of the nodes. In this paper, we present a new search strategy for the optimisation criteria used for selecting the next pair to merge and we show empirically that the new search strategy is superior to other state-of-the-art neighbour-joining implementations. These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.