Constructing Local Cell Sepcific Networks from Single Cell Data

Gene co-expression networks yield critical insights into biological processes, and single-cell RNA sequencing provides an opportunity to target inquiries at the cellular level. However, due to the sparsity and heterogeneity of transcript counts, it is challenging to construct accurate gene networks. We develop an approach that estimates cell-specific networks (CSN) for each cell using a method inspired by Dai et al. (7). Although individual CSNs are estimated with considerable noise, average CSNs provide stable estimates of network structure, which provide better estimates of gene block structures than traditional measures. The method, called locCSN, is based on a non-parametric investigation of the joint distribution of gene expression, hence it can readily detect nonlinear correlations, and it is more robust to distributional challenges. The individual networks preserve information about the heterogeneity of the cells and having repeated estimates of network structure facilitates testing for difference in network structure between groups of cells. The original CSN algorithm showed promise; however, it had shortcomings which locCSN overcomes. Additionally, we propose new downstream analysis methods using CSNs, to utilize more fully the information contained within them. Finally, to further our understanding of autism spectrum disorder we examined the evolution of gene networks in fetal brain cells and compared the CSNs of cells sampled from case and control subjects to reveal intriguing patterns in gene co-expression changes.