This application relates to the field of
bioinformatics computing technology, and discloses a method, device,
computer equipment, and storage medium for predicting
pathway activity status. The method includes: acquiring multi-
omics data of a target object, wherein the multi-
omics data includes multiple different
omics entities; performing
interaction strength analysis on each omics entity to obtain edge weight data between each omics entity; evaluating
pathway activity on the edge weight data based on a pre-trained multilayer heterogeneous graph convolutional network to obtain
pathway activity evaluation results of the edge weight data; and predicting the pathway
activity status of the target
object based on the activity evaluation results. This method can effectively capture the activity changes of genes in different omics entities, thereby improving the accuracy of
biological pathway activity prediction.