AUTHOR=Wang Bo , Ma Fangjian , Du Xiaoxin , Zhang Guangda , Li Jingyou 

TITLE=Prediction of microbe–drug associations based on a modified graph attention variational autoencoder and random forest

JOURNAL=Frontiers in Microbiology

VOLUME=Volume 15 - 2024

YEAR=2024

URL=https://www.frontiersin.org/journals/microbiology/articles/10.3389/fmicb.2024.1394302

DOI=10.3389/fmicb.2024.1394302

ISSN=1664-302X

ABSTRACT=The identification of microbe-drug associations can greatly facilitate drug research and development. Traditional methods for screening of microbe-drug associations is time-consuming beside manpower and money costly, so computational methods is a good alternative. However, most of them ignore the combination of abundant sequence, structural information, and microbe-drug network topology. In this study, we developed a computational framework based on a modified graph attention variational autoencoder (MGAVAEMDA) to infer potential microbial-drug associations by combining biological information with the variational autoencoder. In MGAVAEMDA, we first used multiple databases, which include microbial sequences, drug structures and microbe-drug association databases to establish two comprehensive feature matrices of microbes and drugs after multiple similarity computation,fusion, smoothing and thresholding. Then we further adopted a combination of variational autoencoder and graph attention to extract low-dimensional feature representations of microbes and drugs. Finally, the low-dimensional feature representation and graphical adjacency matrix were input into the random forest classifier to obtain the microbial-drug association score to identify the potential microbial-drug association. Moreover, in order to correct the model complexity and redundant calculation to improve efficiency, we introduced a modified graph convolutional neural network embedded into the variational autoencoder for computing low-dimensional features.The experiments results demonstrate that the prediction performance of MGAVAEMDA better than 5 state-of-the-art methods. For the major measurements (AUC = 0.9357, AUPR = 0.9378), the relative improvements of MGAVAEMDA compared to the suboptimal methods are 1.76%, and 1.47%, respectively. In addition, we conducted case studies on two drugs and found that more than 85% of the predicted associations have been reported in PubMed. The comprehensive experimental results validated the reliability of our models in accurately inferring potential microbe-drug associations.