Asia-Pacific Biotech News

The article is an overview of bioinformatics in Asia Pacific.

Bioinformatics is the intersection of computer science, statistics, molecular biology and genetics. It is one of the most important emerging research areas of the 21st century and has already attracted worldwide interest. It is clear that major initiatives are being undertaken which will establish Australia both as a vital link in the international bioinformatics community for research and development and also as an Asia-Pacific service for bioinformatics. This article briefly notes some groups carrying out bioinformatics research in Australia.

This review describes the current state of bioinformatics education in Australia, outlining the award and non-award courses offered by Australian Universities and centers.

Singapore seeks to be an international center for the biomedical sciences and its related industries. Bioinformatics is seen as an important ingredient in this ambition. We provide in this short report a brief overview of bioinformatics in Singapore. We cover aspects such as training (Section 3), research (Section 4), and commercialization (Section 5). We also introduce some of the main centers of activities, as well as some of the bioinformaticists in these centers.

Bioinformatics plays an important role for in the research and development of the life science and biotechnology. This paper intends to give an overview of the activities of bioinformatics service, research and education at the Center of Bioinformatics, Peking University; the national node of the European Molecular Biology Network and the Asia Pacific Bioinformatics Network.

This article starts off by defining bioinformatics, then it goes on to describes the bioinformatics industry in Australia.

The Oracle relational database management system, with object-oriented extensions and numerous application-driven enhancements, plays a critical role worldwide in managing the exploding volumes of bioinformatics data. There are many features of the Oracle product which support the bioinformatics community directly already and there are several features which could be exploited more thoroughly by users, service vendors, and Oracle itself to extend that level of support. This paper will present an overview of Oracle features which support storage of bioinformatics data and will discuss extensibility features which give the product room to grow. Some attention will be given to Oracle's own efforts to use that extensibility to exploit emerging standardization of many of the complex data and computation requirements of the life sciences.

The article describes the market for bioinformatics, its regional attractiveness index and the market segments of bioinformatics. It also talks about bioinformatics in Asia Pacific: Japan, Australia, India, Singapore and South Korea.

This paper considers the problem of inferring an original sequence from a number of erroneous copies. The problem arises in DNA sequencing, particularly in the context of emerging technologies that provide high throughput or other advantages at the cost of an increased number of errors. We describe and compare two approaches that have recently been developed by the authors. The first approach searches for a sequence known as a Steiner string; the second searches for the most probable original sequence with respect to a simple Bayesian model of sequencing errors. We present the results of extensive tests in which erroneous copies of real DNA sequences were simulated and the algorithms were used to infer the original sequences. The results are used to compare the two approaches to each other and to a third, more conventional, approach based on multiple sequence alignment. We find that the Bayesian approach is superior to the Steiner approach, which in turn is superior to the alignment approach.

The two new algorithms can also be used to construct multiple sequence alignments. We show that the two methods produce alignments of approximately equal quality, and conclude that the Steiner approach is better for this purpose because it is faster. Both methods produce better alignments than a well-known multiple sequence alignment package, for the cases tested.

The interactions of large molecules with surfaces and with each other are strongly dependent upon their surface, rather than their bulk properties. In addition, the local properties of biomolecular surfaces are very important in their own right in biomedicine and other areas, for example for locating binding sites.

Following to previous work, we have developed a program to compute to compute amino acid and atom-based surface descriptors, and used it to generate a small database of charge and hydrophobicity-related surface properties for a set of proteins. The program requires the user to input two text files: one assigning a real number to each atom of each amino acid, and one assigning a real number to each amino acid. Although we have so far only computed surface charge (atom-based) and surface hydrophobicity (amino acid-based), we note that this program could be used to compute any surface parameter whatsoever, since the user can assign arbitrary atom-by-atom and amino acid properties.We discuss possible applications of this program and describe one current application, the Biomolecular Adsorption Database.

Bioinformatics techniques to protein secondary structure prediction, such as Support Vector Machine (SVM) and GOR approaches, are mostly single-stage approaches; they predict secondary structures of the protein by taking into account only the information available in amino acid sequences. On the other hand, PHD (Profile network from HeiDelberg) method is a two-stage technique where two Multi-Layer Perceptrons (MLPs) are cascaded; the second neural network receives the output of the first neural network captures any contextual relationships among the secondary structure elements predicted by the first neural network. In this paper, we argue that it is feasible to extend the current single-stage approaches by adding a second-stage prediction scheme to capture the contextual information among secondary structural elements and thereby improving their accuracies. We demonstrate that two-stage SVMs perform better than present techniques for protein secondary structure prediction.