Speaker
Description
Network reconstruction from high-dimensional omics data remains an open challenge. While one can calculate gene-gene co-expression statistics for every pair of genes, these complete weighted graphs must then be sparsified to obtain an interpretable network. Common sparsification approaches (such as thresholding) can lead to an excessively fragmented network, masking the relationship between genes and neglecting the "strength of weak ties," wherein a critical but weak relationship may serve as a vital link between two subnetworks. Here we present Network Skeleton Extraction (NSE), a co-expression network generation method using spectral sparsification to sparsify co-expression statistics into minimal co-expression graphs. Spectral sparsification has the advantage of maintaining connections among genes in a manner that preserves the coarse-grained structure and dynamical properties of the input graph. This yields networks that are highly sparse while still being predictive of gene expression. We also present a probabilistic model to generate a null distribution of networks with similar spectral properties, against which inferred networks can be compared. We illustrate the method by applying it to Xenopus transcriptome data across six developmental stages (from pluripotency to lineage commitment), identifying networks whose structure changes over the course of development to give rise to cell-type specific networks.
