A study on the concentration mechanism of autistic children based on graph attention network and gene editing technology

This study focuses on the attention mechanism in children with Autism Spectrum Disorder (ASD) and proposes an analytical framework combining Graph Attention Networks (GAT) with gene editing technology to systematically explore key gene modules affecting attention performance in ASD children. By constructing a simulated gene network containing 150 ASD-related genes and generating gene expression and attention score data for 200 virtual individuals, the research demonstrates that the GAT model significantly outperforms traditional MLP and linear models in predicting attention levels. Model interpretability analysis reveals distinct patterns of key genes, including synaptic function genes such as SYNGAP1 and SHANK3, and chromatin regulation genes such as CHD8 and ADNP, between high and low attention groups. Based on these findings, three specific gene editing strategy recommendations are proposed: upregulation of synaptic function genes, downregulation of chromatin regulation genes, and combined multi-target interventions during critical developmental periods. The results not only confirm the effectiveness of the GAT model in ASD biological networks but also provide a theoretical basis for precise gene intervention strategies. Future work will focus on validating these findings using real clinical data and implementing CRISPR-based experiments.