Various methods to compare given biological networks have been proposed to date. For an instance, MI-GRAAL[8] is one of such popular methods. However, the method uses only local structural information to calculate a similarity among nodes. Owing to this limitation, the resulted alignment may not reflect the global features of the given networks. In social network analysis certain measurements, so-called network characteristics are used to capture some features of nodes in graphs. And some of these reflect global features of nodes in networks. In this paper, we proposed a network alignment method using a node similarity based on network characteristics so that resulted alignment would reflect the global structural features more than the traditional method. We compared our proposed method with traditional network alignment method, MI-GRAAL, to demonstrate the effectiveness of our proposal. The experiment was carried out through protein-protein interactions (PPI) networks of yeast and human. The results showed that proposed method led to better alignment in view of topological quality than MI-GRAAL.

View full abstract

We developed a new information management system, Protein Experimental Information Management System (PREIMS), which has the ontology-based functions for quality control, validation, scalability, and information sharing. Its contents are mainly experimental protocols for the analyses of protein structures and functions, and their results. They are stored separately in the PREIMS database (DB), as the ontology based protocol data and the result data. The synchrotron experimental information was stored as the latter result data in Extensible Markup Language (XML). Furthermore we converted those protocols in the format of Resource Description Framework (RDF) for integration with other biological information resources.

View full abstract

Protein-ligand interaction prediction plays an important role in drug design and discovery. However, wet lab procedures are inherently time consuming and expensive due to the vast number of candidate compounds and target genes. Hence, computational approaches became imperative and have become popular due to their promising results and practicality. Such methods require high accuracy and precision outputs for them to be useful, thus, the problem of devising such an algorithm remains very challenging. In this paper we propose an algorithm employing both support vector machines (SVM) and an extension of canonical correlation analysis (CCA). Following assumptions of recent chemogenomic approaches, we explore the effects of incorporating bias on similarity of compounds. We introduce kernel weighted CCA as a means of uncovering any underlying relationship between similarity of ligands and known ligands of target proteins. Experimental results indicate statistically significant improvement in the area under the ROC curve (AUC) and F-measure values obtained as opposed to those gathered when only SVM, or SVM with kernel CCA is employed, which translates to better quality of prediction.

View full abstract

More and more biomedical documents are digitally written and stored. To make the most of the rich resources, it is crucial to precisely locate the information pertinent to user's interests. An obstacle in finding information in natural language text is negations, which deny or reverse the meaning of a sentence. This is especially problematic in the biomedical domain since scientific findings and clinical records often contain negated expressions to state negative effects or the absence of symptoms. This paper reports on our work on a hybrid approach to negation identification combining statistical and heuristic approaches and describes an implementation of the approach, named NegFinder, as a Web service.

View full abstract

In this paper, we propose a novel method, named SCPSSMpred (Smoothed and Condensed PSSM based prediction), which uses a simplified position-specific scoring matrix (PSSM) for predicting ligand-binding sites. Although the simplified PSSM has only ten dimensions, it combines abundant features, such as amino acid arrangement, information of neighboring residues, physicochemical properties, and evolutionary information. Our method employs no predicted results from other classifiers as input, i.e., all features used in this method are extracted from the sequences only. Three ligands (FAD, NAD and ATP) were used to verify the versatility of our method, and three alternative traditional methods were also analyzed for comparison. All the methods were tested at both the residue level and the protein sequence level. Experimental results showed that the SCPSSMpred method achieved the best performance besides reducing 50% of redundant features in PSSM. In addition, it showed a remarkable adaptability in dealing with unbalanced data compared to other methods when tested on the protein sequence level. This study not only demonstrates the importance of reducing redundant features in PSSM, but also identifies sequence-derived hallmarks of ligand-binding sites, such that both the arrangements and physicochemical properties of neighboring residues significantly impact ligand-binding behavior.

View full abstract