Synergistic anti-cancer drug combination recognition method based on molecular fingerprints and multi-target proteins

Abstract

The invention provides a collaborative anti-cancer drug combination recognition method based on molecular fingerprints and multi-target proteins, which comprises the following steps: firstly, using aChemoPy toolkit to carry out molecular fingerprint feature calculation on drug compounds, and then using PSI-BLAST and PSI-PRED software to carry out feature extraction on multi-treatment target protein of cancer cell lines; and on the basis, inputting the molecular fingerprint characteristics of the drug compound and the multi-treatment target protein characteristics of the cancer cell line intoa two-stage deep convolutional neural network, and carrying out network training and prediction by taking a multi-classification cross entropy loss function as a target. According to the method, the prediction precision of the collaborative anti-cancer drug combination is improved, and the recognition speed of the collaborative anti-cancer drug combination is increased.

Description

technical field [0001] The invention belongs to bioinformatics drug disease effect analysis technology, specifically a method for identifying synergistic anticancer drug combinations based on molecular fingerprints and multi-target proteins. Background technique [0002] Drug synergy means that the curative effect of the combination of drugs exceeds the sum of the curative effects of the individual components of the drug. With the further understanding of the pathogenic factors and treatment methods of major diseases such as cancer, the combination drug method is becoming the mainstream drug method in cancer treatment because of its low side effects and long-lasting efficacy. The interaction of drug combination can be divided into additive effect, antagonistic effect and synergistic effect according to the curative effect. Identifying drug-drug interactions, especially synergy, is important for maximizing drug therapeutic effects. [0003] Traditional clinical analysis and...

AI Technical Summary

Problems solved by technology

Researchers in Finland used a machine learning method based on matrix decomposition to achieve a high-precision prediction model for synergistic drug combinations for cancer, malaria, Ebola and other major diseases; the prediction performance of the method based on matrix decomposition is also heavily dependent on The prior data set is rich and susceptible to noise interference; if there are a small number of data outliers in the data set, the predictive performance of the method based on matrix decomposition will be significantly reduced

Chinese researchers used a method based on a deep neural network, combined with a gated recurrent unit and a convolutional unit to predict the synergistic effect. This method achieved an accurate analysis of the relationship between drugs and diseases. The complexity of the model has increased significantly, and the prediction is very time-consuming

Images