Deciphering Myopia: Insights From Single-Cell Data and Machine Learning Approaches

Myopia, the most prevalent refractive error, is a leading cause of vision impairment worldwide, particularly in East Asia. High myopia can lead to irreversible vision loss, emphasizing the need for early identification and intervention. This study provides a comprehensive analysis of single-cell and bulk transcriptome data, combined with machine learning models, to elucidate the molecular mechanisms and key genes in myopia. Results showed significant changes in fibroblasts and endothelial cells in myopia groups. Machine learning models, especially LightGBM and SVM, effectively classified myopia and control samples. Feature importance analysis identified key genes crucial for model prediction. The findings not only reveal pathological changes at the cellular level in myopia but also provides a new way for predicting myopia, highlighting potential therapeutic targets.