From communities to protein complexes: A local community detection algorithm on PPI networks

Identifying protein complexes in protein-protein interaction (ppi) networks is often handled as a community detection problem, with algorithms generally relying exclusively on the network topology for discovering a solution. The advancement of experimental techniques on ppi has motivated the generation of many Gene Ontology (go) databases. Incorporating the functionality extracted from go with the topological properties from the underlying ppi network yield a novel approach to identify protein complexes. Additionally, most of the existing algorithms use global measures that operate on the entire network to identify communities. The result of using global metrics are large communities that are often not correlated with the functionality of the proteins. Moreover, ppi network analysis shows that most of the biological functions possibly lie between local neighbours in ppi networks, which are not identifiable with global metrics. In this paper, we propose a local community detection algorithm, (lcda-go), that uniquely exploits information of functionality from go combined with the network topology. lcda-go identifies the community of each protein based on the topological and functional knowledge acquired solely from the local neighbour proteins within the ppi network. Experimental results using the Krogan dataset demonstrate that our algorithm outperforms in most cases state-of-the-art approaches in assessment based on Precision, Sensitivity, and particularly Composite Score. We also deployed lcda, the local-topology based precursor of lcda-go, to compare with a similar state-of-the-art approach that exclusively incorporates topological information of ppi networks for community detection. In addition to the high quality of the results, one main advantage of lcda-go is its low computational time complexity.

UAV-UGV-UMV Multi-Swarms for Cooperative Surveillance

In this paper we present a surveillance system for early detection of escapers from a restricted area based on a new swarming mobility model called CROMM-MS (Chaotic Rössler Mobility Model for Multi-Swarms). CROMM-MS is designed for controlling the trajectories of heterogeneous multi-swarms of aerial, ground and marine unmanned vehicles with important features such as prioritising early detections and success rate. A new Competitive Coevolutionary Genetic Algorithm (CompCGA) is proposed to optimise the vehicles’ parameters and escapers’ evasion ability using a predator-prey approach. Our results show that CROMM-MS is not only viable for surveillance tasks but also that its results are competitive in regard to the state-of-the-art approaches.

Interplay between success and patterns of human collaboration: case study of a Thai Research Institute

Networks of collaboration are notoriously complex and the mechanisms underlying their evolution, although of high interest, are still not fully understood. In particular, collaboration networks can be used to model the interactions between scientists and analyze the circumstances that lead to successful research. This task is not trivial and conventional metrics, based on number of publications and number of citations of individual authors, may not be sufficient to provide a deep insight into the factors driving scientific success. However, network analysis techniques based on centrality measures and measures of the structural properties of the network are promising to that effect. In recent years, it has become evident that most successful research works are achieved by teams rather than individual researchers. Therefore, researchers have developed a keen interest in the dynamics of social groups. In this study, we use real world data from a Thai computer technology research center, where researchers collaborate on different projects and team up to produce a range of artifacts. For each artifact, a score that measures quality of research is available and shared between the researchers that contributed to its creation, according to their percentage of contribution. We identify several measures to quantify productivity and quality of work, as well as centrality measures and structural measures. We find that, at individual level, centrality metrics are linked to high productivity and quality of work, suggesting that researchers who cover strategic positions in the network of collaboration are more successful. At the team level, we show that the evolution in time of structural measures are also linked to high productivity and quality of work. This result suggests that variables such as team size, turnover rate, team compactness and team openness are critical factors that must be taken into account for the success of a team. The key findings of this study indicate that the success of a research institute needs to be assessed in the context of not just researcher or team level, but also on how the researchers engage in collaboration as well as on how teams evolve.

Link Definition Ameliorating Community Detection in Collaboration Networks

Collaboration networks are defined as a set of individuals who come together and collaborate on particular tasks such as publishing a paper. The analysis of such networks permits to extract knowledge on the structure and patterns of communities. The link definition and network extraction have a high impact on the analysis of collaboration networks. Previous studies model the connectivity in a network considering it as a binomial problem with respect to the existence of a collaboration between individuals. However, such a data consists of a high diversity of features that describe the quality of the interaction such as the contribution amount of each individual. In this paper, we have determined a solution to extract collaboration networks using corresponding features in a dataset. We define collaboration score to quantify the collaboration between collaborators. In order to validate our proposed method, we benefit from a scientific research institute dataset in which researchers are co–authors who are involved in the production of papers, prototypes, and intellectual properties (IP). We evaluated the generated networks, produced through different thresholds of collaboration score, by employing a set of network analysis metrics such as clustering coefficient, network density, and centrality measures. We investigated more the obtained networks using a community detection algorithm to further discuss the impact of our model on community detection. The outcome shows that the quality of resulted communities on the extracted collaboration networks can differ significantly based on the choice of the linkage threshold.