Development And Evaluation On Novel Prediction Tools For RNA-Ligand Docking

Medical drug research and development is a long-term,high-input and high-risk process,so it is great significance to speed up drug research and development,reduce capital investment costs and improve the success rate of drug research and development.As a kind of ribonucleic acid,RNA is not only an important genetic material of life system,but also an important biological target of many major diseases.At the beginning of drug research and development,molecular docking tools are used for high-throughput screening to determine drug lead compounds,which can provide instructive computational data for drug design.Therefore,the research on RNA-ligand docking tool has great scientific and practical value.In this article,taking the RNA-ligand docking conformation search as the scientific problem and improving the molecular docking accuracy as the starting point,an efficient docking conformation space search algorithm with high docking accuracy is proposed.The RNA-ligand docking problem is defined as an optimization problem whose dimension is related to the number of rotatable bonds of the ligand,the objective function of the optimization problem is defined as minimizing the binding energy of the docking conformation,and the constraint condition of the optimization problem is defined as the bonding between the atoms of the ligand molecule.The artificial bee colony algorithms are used to solve the multidimensional optimization problem.Artificial bee colony algorithms are widely used,but it has the disadvantages of slow convergence and easy to fall into local optimization.In this paper,based on the traditional artificial bee colony algorithm,the random honey source search strategy of the leading bee and the following bee is improved.The leading bee adopts the global optimal subspace honey source update mode of adaptive nonlinear search strategy,while the following bee adopts the local optimal subspace search strategy of random nonlinear update dimension.Therefore,an artificial bee colony algorithm based on optimal subspace adaptive search is proposed.In the docking experiment,the semi-empirical potential energy evaluation is used as the combined energy evaluation function to guide the conformational optimization search direction of the optimization algorithm.This paper mainly takes FIPSDock as the research basis,and extends the framework of FIPSDock,and introduces the proposed optimal subspace search artificial bee colony algorithm EPSABC into FIPSDock.Hence,we have achieved the RNA-ligand oriented docking tool of EPSDock.In the same experimental environment,multiple open source RNA-ligand complex data sets were compared with widely used docking tools such as Auto Dock4,Auto Dock Vina,r Dock,Glide and FIPSDock,and the docking success rate of EPSDock was 84%.The experimental results show that the molecular docking problems of high dimension and high flexibility of RNA-ligand,EPSDock shows higher docking success rate and docking accuracy,which further verifies the effectiveness,stability and accuracy of the improved artificial bee colony algorithm proposed in this research.