Random intercept hierarchical linear model for multi-regional clinical trials.

In multi-regional clinical trials, hierarchical linear models have been actively studied because they can reflect that patients in the same region share common intrinsic and extrinsic factors. In this paper, we investigate the statistical properties of the hierarchical linear model including a random effect in the intercept. The big advantage of the random intercept hierarchical linear model is that it can control the type I error rates of testing the overall treatment effect when there are no or clinically negligible regional differences in the treatment effect. Moreover, we compare the pros and cons with the hierarchical linear model in which the random effect is included in the slope. For the two hierarchical linear models, the model selection criteria are determined according to the magnitude of the difference in treatment effect across the regions, and we provide the criteria through simulation studies.

Machine learning model for predicting age in healthy individuals using age-related gut microbes and urine metabolites.

Age-related gut microbes and urine metabolites were investigated in 568 healthy individuals using metataxonomics and metabolomics. The richness and evenness of the fecal microbiota significantly increased with age, and the abundance of 16 genera differed between the young and old groups. Additionally, 17 urine metabolites contributed to the differences between the young and old groups. Among the microbes that differed by age, Bacteroides and Prevotella 9 were confirmed to be correlated with some urine metabolites. The machine learning algorithm eXtreme gradient boosting (XGBoost) was shown to produce the best performing age predictors, with a mean absolute error of 5.48 years. The accuracy of the model improved to 4.93 years with the inclusion of urine metabolite data. This study shows that the gut microbiota and urine metabolic profiles can be used to predict the age of healthy individuals with relatively good accuracy.

ACGCN: Graph Convolutional Networks for Activity Cliff Prediction between Matched Molecular Pairs.

One of the interesting issues in drug-target interaction studies is the activity cliff (AC), which is usually defined as structurally similar compounds with large differences in activity toward a common target. The AC is of great interest in medicinal chemistry as it may provide clues to understanding the complex properties of the target proteins, paving the way for practical applications aimed at the discovery of more potent drugs. In this paper, we propose graph convolutional networks for the prediction of AC and designate the proposed models as Activity Cliff prediction using Graph Convolutional Networks (ACGCNs). The results show that ACGCNs outperform several off-the-shelf methods when predicting ACs of three popular target data sets for thrombin, Mu opioid receptor, and melanocortin receptor. Finally, we utilize gradient-weighted class activation mapping to visualize activation weights at nodes in the molecular graphs, demonstrating its potential to contribute to the ability to identify important substructures for molecular docking.