Deep learning reveals a systems perspective on the origin and evolution of insect wings

The origin and evolutionary mechanisms of insect wings are important and long-standing scientific questions in evolutionary biology. Although some key genes related to wing formation have been identified, there is still a lack of systematic explanation of the regulatory mechanisms of these genes in different insect taxa. In this study, we analyzed the molecular regulatory mechanisms of insect wing origin and evolution by deep learning techniques. We constructed models based on Bidirectional Gated Recurrent Unit and Attention Mechanisms, integrated genomic protein sequence information from 119 species, and successfully identified a series of highly weighted genes that may be involved in the origin and evolution of insect wings. The results suggest that the origin and evolution of insect wings may be a complex systematic engineering. In summary, this study not only reveals the molecular mechanism of insect wing development, but also reveals the potential value of deep learning in complex biology research.