3D drug design method for targeting receptor based on artificial intelligence

Abstract

The invention relates to a 3D drug design method for targeting a receptor based on artificial intelligence. The method mainly aims at solving the problems that an existing artificial intelligence training model can only grow into 1D or 2D drug molecular fragments and cannot directly target a receptor 3D active pocket. 1D or 2D drug fragment molecules are converted by utilizing an MMFF94 force field, a starting fragment in a 3D active pocket of the receptor is randomly specified or specially specified as a growth starting point of the drug molecule, the rotatable bond of the drug molecule is used as a segmentation point, and 3D drug design is carried out on the 3D active pocket of the receptor target through adoption of a genetic algorithm, so that the problem that a drug model trained by artificial intelligence cannot be directly applied to the 3D active pocket of the receptor target to carry out 3D drug molecule design is solved.

Description

technical field [0001] The invention belongs to the technical field of drug molecules, and in particular relates to a 3D drug design method for targeting receptors based on artificial intelligence. Background technique [0002] At present, the molecules grown from drug models trained based on artificial intelligence are all one-dimensional (1D) or two-dimensional (2D), such as the SMILES sequence structure, which cannot be based on receptor targets (such as proteins, nucleic acids, etc.) The three-dimensional (3D) active center grows a 3D drug molecule structure, and the present invention mainly solves the technical problem that the drug fragment growth model trained by artificial intelligence performs 3D drug molecule growth in the receptor target active pocket. [0003] Rational drug design in computer-aided drug design is to find the optimal 3D drug molecule in the 3D active pocket of the receptor target (eg, protein, nucleic acid, etc.). Artificial intelligence has been...

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to provide a 3D drug design method for targeting receptors based on artificial intelligence, mainly for the current artificial intelligence training model can only grow into 1D or 2D drug molecule fragments and cannot directly target receptor 3D activity The problem of the pocket, by using the MMFF94 force field to convert 1D or 2D drug fragment molecules, randomly designate or specifically designate the starting fragment in the 3D active pocket of the receptor as the starting point of the growth of the drug molecule, and use the rotatable bond of the drug molecule as the starting point for the growth of the drug molecule Segmentation points, 3D drug design for the 3D active pocket of the receptor target through the genetic algorithm, and then solve the problem that the drug model trained by artificial intelligence cannot be directly applied to the active pocket of the receptor target for 3D drug molecular design

Images