Learning massive interpretable gene regulatory networks of the human brain by merging Bayesian Networks

Abstract

We present FGES-Merge, a new method for learning the structure of gene regulatory networks via merging locally learned Bayesian networks, based on the fast greedy equivalent search algorithm. The method is competitive with the state of the art in terms of the recall of the true structure while also improving upon it in terms of speed, scaling up to the tens of thousands of variables and being able to use empirical knowledge about the topological structure of gene regulatory networks. We apply this method to learning the gene regulatory network for the full human genome using data from samples of different brain structures (from the Allen Human Brain Atlas). Our goal is to develop a Bayesian network model that predicts interactions between genes in a way that is clear to experts, following the current trends in interpretable artificial intelligence. To achieve this, we also present a new open-access visualization tool that facilitates the exploration of massive networks and can aid in finding nodes of interest for experimental tests.

Explainable artificial intelligence for omics data: a systematic mapping study

“…Although all included studies follow a design approach (i.e. design, implementation and evaluation of specific AI models), four studies [ 49–52 ] also conduct an extensive literature review and/or case study, qualifying them as mixed methods approaches. 247 studies only provide an implementation or prototype, while 158 develop a general concept in addition to implementation.…”

A survey of Bayesian Network structure learning

“…The authors find that GIES orientates more edges and has increasingly better accuracy than GES as the number of different intervention datasets or number of intervention targets grows. Bernaola et al (2020) introduce FGES-Merge which is focussed on learning very large networks, with tens of thousands of variables, some with very large degree, typical of those encountered when modelling gene regulation networks. FGES-Merge uses FGES to learn sub-graphs around each node and then merges these to create the whole graph.…”

Discover the Dehydration Response Genes in <i>Boea hygrometrica</i> Transcriptome Using Bayesian Network Approach

“…Bayesian network approach [3] is a promising tool for transcriptome data analysis [4][5][6][7] and biological network reconstruction [8][9][10][11][12]. Bayesian network approach is a kind of data-driven analysis method.…”