Transfer Learning and Permutation‐Invariance Improving Predicting Genome‐Wide, Cell‐Specific and Directional Interventions Effects of Complex Systems

With the rise of precision medicine, single-drug treatments alone may not meet its demands. However, extensive data and deep learning models exist for single compounds. Leveraging transfer learning to use this data for predicting complex system intervention effects is promising. In this study, a deep model based on permutation-invariance is used as the core module, pre-trained on a large amount of single-compound intervention data in cell lines, and fine-tuned on a small amount of complex system (like natural products) intervention data in cell lines, resulting in a predictive model, Set Embedding and Transfer learning model for Complex systems (SETComp). The two versions of SETComp achieved an accuracy of 93.86% and 92.70%, respectively, on the complex system-cell-gene association test set, improving by 5.82% to 27.59% compared to the baseline. When predicting the intervention effects of those complex systems the model has never encountered before, the accuracy increased by up to 24.83% compared to the baseline. In the in vitro validation, up to 88.65% of the predictions are confirmed to be correct, and the model's output showed a significant positive correlation with the real-world foldchange. SETComp's potential in various biomedical scenarios is further observed, achieving good performance in applications such as mechanism uncovering and drug repositioning.