Prediction of Drug-Target Interaction Based on Multi-Head Self-Attention
Xin Ma, Yujing Cheng · 2022 IEEE Conference on Telecommunications, Optics and Computer Science (TOCS) · 2022
Excellent drug research and development technology not only plays a positive role in promoting the country's medical status in the world, but also has a supporting function in the steady development of social health. At present, the traditional drug research and development process has the disadvantages of long cycle, high cost, and high failure rate. Therefore, it is very important to find the interaction between drugs and targets as soon as possible for the research and development of new drugs and the new use of old drugs. In this study, a prediction method of drug-target interaction based on multi-head self-attention mechanism is proposed and implemented. Based on the data sets of DrugBank, UniProt, and PubCham, 24106 pieces of data screened were used for 100 epochs of training. The accuracy of 4821 samples in the validation set is as high as 89%. The detection accuracy of untrained new data is also close to 100%, and the recall rate is slightly lower than that of validation set data.