|
International Journal of Information Technology Vol. 11 No. 8 2005 (Special Issue) Guest Editorial Special Issue on Bioinformatics and Biomedical Systems With rapid advances in technology and the development of ultra high-throughput research, the field of biotechnology is now considered to be suffering from data overload. This has led to the development of a broadening field of science, termed bioinformatics, which has evolved from the convergence of biology and information technology. Bioinformatics facilitates genomic and post-genomic data analysis and the integration of data from the related fields of genomics, transcriptomics, proteomics, metabolomics and phenomics. This data integration enables the identification of genes and gene products and characterises the functional relationships between the genotype and the observed phenotype, to enable a systems biology analysis of information from the genome to the phenome, and eventually, the biome. The application of bioinformatics has expanded with the ‘omics’ technologies, and bioinformatics is required to integrate the diverse data from these technologies. Biotechnology demands the integration, intelligent searching and filtering of the numerous, complex data types to answer specific biotechnology questions, ranging across specialist research fields outside of the knowledge of any individual. These research fields include gene and genome sequencing, comparative and functional genomics, molecular genetics, gene and protein expression analysis, protein structure prediction, protein design and kinetics, metabolite flux analysis, phenotypic modeling, the development of data standards, ontologies and data representations, applied statistics, analytical tools and systems modeling. Bioinformatics now sits as an umbrella over biotechnology, providing the design, development and application of data analysis algorithms, integrated database systems and web-tools. As the data continues to increase in quantity and complexity, there will be an increasing reliance on the fusion of biotechnology and information technology to provide bioinformatics answers to demanding biotechnology questions. Through this fusion, we can gain a greater knowledge and understanding of the relationship between the heritable material, DNA, and the observed phenotype, from the molecular to the organism and environmental scale. This special issue comes from a collection of papers presented at 2005 International Conference of Intelligent Computing. After extensive reviews and revisions, 16 papers were accepted. These papers cover a wide variety of areas of bioinformatics and are of exceptionally high quality and significance. Below, we provide a summary of the papers published in this journal issue. Huiyu Xia, Fei Li, Tao He, and Yanda Li analyze the hairpin structures of 557 miRNA genes from six eukaryotic organisms using a random model. Their results provide a new character of stem-loop structure for miRNA prediction and indicate that this character might facilitate miRNA biogenesis. Min Hu, Wei Chen, Tao Zhang, Qunsheng Peng and Liguang Xie present a novel technique for estimating the potential similarity of protein 3D shape. Such quantitative measurement is helpful to reduce the search space when we retrieve protein structures from large data sets. Yanhong Zhou, Huili Zhang, Lei Yang, and Honghui Wan develop a gene prediction program GeneKey. Statistical analysis demonstrates that some structure features, such as splicing signals and codon usage, of CG-poor genes are quite different from that of CG-rich ones. The results imply that careful construction of training dataset is very important for improving the performance of various prediction tasks. Mahmood Akhtar presents a detailed comparison of time-domain and frequency-domain techniques for the detection of both short and long coding regions that are both closely and widely spaced. Liangliang Wang, Jinwen Ma proposes a post-filtering gene selection algorithm to discover informative genes of a tumor. Compared with the conventional methods, the constructed diagnosis system with the post-filtering gene selection algorithm can reach higher diagnosis accuracy on both the colon and leukemia data, with a smaller number of informative genes. Yan Li and Zheru Chi present a new unsupervised MR image segmentation method based on self-organizing feature map (SOFM) network. Paul C. Conilione and Dianhui Wang explore the application of feature selection by the Correlation based Feature Selection (CFS) algorithm on the problem of classification of E.coli promoters using neural networks, Support Vector Machines (SVM) and Extreme Learning Machines (ELM). It was found that even though in general the classification accuracy can be reduced by a statistically significant amount, in real terms this was only a few percent. Yue Ma and Cao An Wang propose a new anchor-based model for global multiple alignment of whole genome sequences. The proposed chaining procedure is based on evolutionary theory and can align whole genome sequences not only for close homologs, but also distant species. Wengang Zhou, Hong Zhu, Guixia Liu, Yanxin Huang, Yan Wang, Dongbing Han and Chunguang Zhou present a computational based approach to select the most relevant information for searching binding motifs from the long upstream regions. Wenming Cao and Shoujue Wang consider a learning similarity for classifying gene expression patterns, and therefore create a neural network-based proximity measure. It was observed that the clusters obtained using Euclidean distance, correlation coefficients, and mutual information were not significantly different. Lin Wang, Minghu Jiang and Stefan Wolfl investigate the different activation and silencing of Akt/ ERK/p70S6 signaling in treated human cancer cell lines with proteomic analysis of co-localization using western blot, FACS and nano-gold phospho-antibody microarray by reaction with nuclear and cytoplasmic proteins. This study implies that different activations and silencings of Akt/p70S6/ERK signaling reflect different functions. Wenkai Lu and Yong Yang present a superresolution algorithm using the minimum entropy criterion to improve the depth resolution of Optical Coherent Tomography (OCT) signals. Pan-Gyu Kim, Kiejung Park and Hwan-Gue Cho devise five functions as quality measures for signal noise, background noise, scale invariant, spot regularity, and spot alignment, which can check the quality of microarray and validate the correctness of experiments. This paper also suggests a linear-weighted integration model combining the P-values of the five quality measures. Xiao-Run Wu uses random graph theory to analyze seven different organism protein interaction networks. Three topological properties (degree distribution, clustering coefficient and average shortest path) were used to characterize these networks. Junfeng Gu, Xicheng Wang and Jincheng Zhao develop a novel method for multiple sequence alignment based on wavelet package transform (MAWPT). By means of wavelet package analysis, homologous regions can be found rapidly. Dr. Dianhui Wang, Guest Editor Department of Computer Science and Computer Engineering La Trobe University, Melbourne, VIC 3086, Australia Email: csdhwang@ieee.org Dr. David Edwards, Guest Editor Primary Industries Research Victoria La Trobe University, Melbourne, VIC 3086, Australia Email: Dave.Edwards@dpi.vic.gov.au Dianhui Wang |