基于深度学习和PNP模型的离子通道体系特性研究

离子通道是细胞与周围环境进行物质交换的重要途径，通道内的离子电流对其生物物理特性研究具有重要意义。基于连续模型假设的PNP方程是常用来研究离子电流的分子模拟技术，而其受限于模型的近似性和数值计算的困难，当应用到真实复杂的离子通道体系时，仍不能准确模拟离子电流强度。本项目旨在改进传统的PNP模型，研究变介电系数和Born溶剂化能耦合的混合修正模型，并且发展基于深度学习和PNP模型的离子通道模拟方法，可用于预测离子通道体系在复杂环境下的离子电流强度。项目的研究成果可用于深入研究离子通道体系的生物物理特性，为实验研究提供更多的理论指导，降低实验成本。

Ion channels are an important way for cells to exchange with the surrounding environment. Ion currents in channels are important for their biophysical properties. The PNP equation based on the continuous model is a molecular simulation technique commonly used to study ion currents, which is limited by the approximation of the model and the difficulty of numerical calculation. The ion current cannot be accurately simulated by the PNP model for a real complex ion channel system. This project aims to improve the traditional PNP model, study the mixed modified PNP model of variable dielectric coefficient and Born solvation coupling, and develop an ion channel simulation method based on deep learning and PNP model, which can be used to predict the ion currents of the ion channel system in complex environments. The research results of the project can be used to study the biophysical properties of the ion channel system, which provides more theoretical guidance for experimental research and reduces the experiment cost.