Drug Repositioning Method Based on Pre-trained Large Model and Network Embedding Representation

Drug repositioning offers a novel approach to repurpose existing drugs, which plays an important role in the field of drug development. This method can save the time cost and money cost of drug development, and is safe and reliable, so as to improve the efficiency of drug development. Drug repositioning occurs when the old drug is no longer suitable for the current condition, or when a new application area needs to be explored. However, how to find new indications for drugs for precise drug repositioning is a challenge. To solve this problem, we propose a drug repositioning method based on pre-trained large model and network embedding representation. In this method, the features of biological knowledge are extracted by pre-training the large model, the features of network structure are extracted by network embedding representation model, and the two features are aggregated by the attention layer. Finally, a classifier was used to predict drug-disease associations (DDAs). This model has a certain guiding role for drug repositioning strategies.