基于脉冲神经膜系统的细胞核输出信号识别方法研究

利用高性能计算模型从大规模生物数据中识别细胞核输出信号是计算生物学研究的重要科学问题之一。脉冲神经膜系统是受人类大脑神经元以脉冲方式传递信息的生物功能启发得到的分布式并行神经计算模型，具有计算性能可观、计算效率高、系统扩展性强等优点。本项目研究基于脉冲神经膜系统的细胞核输出信号识别方法，旨在充分发挥脉冲神经膜系统的计算高效性和并行性，为快速准确地识别细胞核输出信号提供新方法，主要研究内容包括：（1）建立处理生物数据的脉冲神经膜系统；（2）发展基于生物数据的脉冲神经膜系统训练方法；（3）利用编程语言和并行元器件实现细胞核输出信号识别方法，并将其应用于细胞核输出信号识别。这方面研究成果可为发展高性能、高并行性、高容错性的神经计算装置提供新模型，也可为从大规模数据中快速准确地识别细胞核输出信号提供新方法，还可为靶标蛋白设计、疾病基因信号识别和分子诊疗提供理论支持。

Using high-performance computing models to identify nuclear export signals is one of the hot and important scientific researching topics in computational biology. Spiking neural P systems (shortly called SN P systems) are a class of distributed and parallel neural-like computing models, which are inspired from the way of neurons in human brain communicating with each other by means of spikes. SN P systems have powerful computability, high computational efficiency and well scalability. In this project, we focus on the research of developing methods of using SN P systems to identify nuclear export signals. The main contents are as follows: (1) constructing new variants of SN P systems, which take biological data as data structure; (2) developing training algorithms for SN P systems to solve biological data; (3) using programmable languages and parallel computing devices to simulate the obtained nuclear export signals identifying algorithms and using the obtained method to predict nuclear export signals from huge biological sequences data. The expected researching results can provide new candidates in designing high-performance, high parallel, high fault-tolerant neural-like computing models and new efficient algorithms to identify and predict nuclear export signals. The obtained results will also give theoretical support to design target protein for regulating disease signals, identify genetic disease signals and disease diagnosis and treatment in molecular level.