Abstract
Purpose:
In the past 20 years, analyses of genetic perturbation in mice have significantly extended our understanding of mammalian lens development. We hypothesized that this wealth of molecular functional information can be “converted” into systems-level information and used as a basis for constructing a detailed spatiotemporal gene regulatory network (GRN). Therefore we comprehensively analyzed functional data on individual regulatory molecules and their targets and developed an algorithmic approach to infer the global gene regulatory relationships over different stages in the lens. These GRNs are publically available at http://bioinformatics.udel.edu/Research/iSyTE/LensGRN.
Methods:
Our approach to construct lens-GRNs is based on a state-of-the-art method successfully applied to derive GRNs for tooth morphogenesis (http://compbio.med.harvard.edu/ToothCODE/). We manually curated perturbation evidences in mouse lens from 54 research articles published in the past 20 years and applied a Bayesian data integration method to infer GRNs in lens initiation, primary fiber cell differentiation, secondary fiber cell differentiation and post-natal stages.
Results:
The current evidence-based lens-GRN is derived from >50 unique regulators and >400 targets, representing ~1300 regulatory relationships in 40 stages from distinct cell types. Based on these integrative approaches, we identified several active networks in different stages. Not surprisingly, Pax6 was identified as a major regulator in all stages, while Fgfr1-3 were critical for regulating gene expression in differentiation. Furthermore, overlay of cataract-associated genes from Cat-map onto these GRNs identified regulatory relationships for 44 genes, with some clustering in distinct pathways. Interestingly, the GRNs revealed that expression of many cataract-associated genes was initiated early in lens development.
Conclusions:
We have constructed lens-GRNs based on hundreds of molecular interactions in the lens. The principle advantage of the curated lens GRN resource is that every edge (relationship) between individual nodes (regulators and targets) is supported by at least one piece of published experimental evidence. In addition to effective visualization of regulatory relationships in lens development, these GRNs will assist in predicting and interpreting how specific perturbations in genetic pathways cause cataract.
Keywords: 445 cataract •
533 gene/expression •
497 development