Enhanced Gene Expression Prediction Models

This document is a research article that presents an investigation into improving gene expression predictive models through the incorporation of protein-mediated spatial chromatin interactions. The study introduces the Spatial Gene Expression (SpEx) algorithm, which enhances the existing ExPecto model by integrating spatial genomic information. The authors detail their methodology, which includes the use of ChIA-PET interactions mediated by proteins such as cohesin, CTCF, and RNAPOL2 across various cell lines. The results indicate statistically significant improvements in predictive accuracy, with the SpEx algorithm achieving higher Spearman’s rank correlation coefficients compared to the baseline ExPecto model. The article discusses the limitations of previous models that relied solely on epigenetic factors and highlights the necessity of incorporating 3D chromatin architecture to better understand gene expression dynamics. The findings contribute to the ongoing advancements in machine learning applications within genomics, emphasizing the role of spatial interactions in gene regulation.