Abstract: 【Purpose/significance】Academic virtual community has become a new important position for contemporary researchers to open access to knowledge, academic exchange and innovation. Based on the mining of hot topics in academic virtual communities, pro⁃ moting the early detection of discipline development trends and frontier hot topics in the field will help scholars grasp the discipline development process and predict the future direction of discipline hot topics as soon as possible, and then support and serve scientific research management decisions.【Method/process】This study focuses on the knowledge exchange in academic virtual communities. Starting from the needs of scientific researchers, it aims to form a method framework for identifying hot topics and predicting their trends in the perspective of academic virtual communities with versatility, mobility and attention to semantic information. Based on the academic virtual community in the field of bioinformatics, empirical research is carried out.【Result/conclusion】This study identifies and predicts the trend from the front end of the evolution of hot topics (demand side), which can effectively reduce the negative impact of the time lag of scientific literature on hot topic prediction The empirical results show that the hot topics in the papers in the field of bioinformatics lag behind the hot topics in the community by 2 years.【Innovation/limitation】It is a useful exploration in the forwardlooking prediction of hot topics, but it needs to be improved in the application scope of empirical research and the fusion analysis of multi-source heterogeneous data.