Protein structure prediction method and device based on Cycle-GAN

Abstract

The invention relates to a protein structure prediction method and device based on Cycle-GAN. The method comprises the steps: acquiring an X-ray crystal diffraction image, a nuclear magnetic resonanceimage and a frozen electron microscope image of a plurality of proteins, and recording the X-ray crystal diffraction image as a first image and the nuclear magnetic resonance image as a second image;constructing a model data set according to the first image, the second image and the nuclear magnetic resonance image; training the CycleGAN by utilizing the model data set, and stopping training when a loss function value of the CycleGAN is lower than a threshold value; and inputting the first image or the second image of the structure of a to-be-predicted protein into the trained Cycle-GAN to obtain a three-dimensional image of the to-be-predicted protein. According to the method, the crystal diffraction image, the nuclear magnetic resonance image and the frozen electron microscope image are reconstructed and fused by using the Cycle-GAN to obtain the frozen electron microscope image, so that the resolution is improved and the cost for predicting the high-resolution structure of the protein is reduced.

Description

technical field [0001] The present invention relates to the fields of biological information and deep learning, in particular to a method and device for predicting protein structure based on Cycle-GAN. Background technique [0002] At present, the experimental methods for determining the three-dimensional structure of proteins include X-ray crystallography, nuclear magnetic resonance (NMR) and cryo-electron microscopy that have emerged in recent years. However, X-ray crystallography requires protein separation, purification and crystallization, and can only measure single crystals, reflecting static structural information, and cannot measure information in solution. For some flexible and complex biomacromolecular proteins, it is difficult to obtain the desired crystal structure. The NMR method does not require crystallization, and can measure the three-dimensional structure in the liquid state, but the resolution is not high. At present, NMR can only be used to determine th...

AI Technical Summary

Problems solved by technology

Cryo-electron microscopy has no crystallization and length limitations, and is the most promising technology for protein structure analysis, but the disadvantage is that the cost of experimental equipment and experiments is too high

Images