APPLICATION OF BIOINFORMATIC PREDICTION FOR SUPPORTING HIGH-THROUGHPUT EXPERIMENTS IN PROTEOME
Recent advances in high-throughput technologies have opened the gate to many frontiers in proteomic researches. However, in spite of the increase of tools, and assay methods, the combinational work of bioinformatics and such high-throughput experiments in proteomic field is still immature. Considering the complexity of proteomic data, a powerful bioinformatic tool to support and enhance the efficiency of exhaustive data acquiring experiments should greatly contribute in the proteomic researches.We here report the efficiency of introducing our fuzzy neural network (FNN) as bioinformatic prediction tool to support high-throughput experiments in proteome. FNN is a type of neural network with implicative model structure. Where usual neural networks offer high prediction accuracy as blackbox, FNN can provide rule matrix by implicating the model structure, and such matrix indicates important feature underlying in the data. With such advantageous feature of FNN, we found that vanous proteomic data could be predicted with high accuracy, and also informative rule matrix could be provided. We applied FNN to lipase engineering data and angiotensin Ⅱ (Ang Ⅱ )binding peptidedata, and examined the efficiency of FNN-supported experiments. As a result, FNN has shown 84. 6％ accuracy in predicting binding-pocket sequence to have inversed enantioselective lipase, and 97. 0％ accuracy in predicting 6-mer peptide that strongly bind to Ang Ⅱ. Furthermore, according to the rule matrix obtained from FNN, we succeeded in designing novel lipase variants and Ang Ⅱ -binding peptides that could not be easily found in simple library screening. Such strategy of introducing bioinformatic prediction tool in biological experiments should ennhnce proteomic researches by decreasing fruitless experiments.
Bioinformatics prediction high-throughput proteome
Ryuji Kato Yasuyuki Tomita Mina Okochi Hiroyuki Honda
Graduate School of Engineering, Nagoya University, Nagoya 464-8603, Japan
国际会议
The 4th International Forum on Post-genome Technologies(4IFPT)(第四届国际后基因组生命科学技术学术论坛)
杭州
英文
26-29
2006-09-25（万方平台首次上网日期，不代表论文的发表时间）