Multi-study inference of regulatory networks for more accurate models of gene regulation

Abstract

Gene regulatory networks are composed of sub-networks that are often shared across biological processes, cell-types, and organisms. Leveraging multiple sources of information, such as publicly available gene expression datasets, could therefore be helpful when learning a network of interest. Integrating data across different studies, however, raises numerous technical concerns. Hence, a common approach in network inference, and broadly in genomics research, is to separately learn models from each dataset and combine the results. Individual models, however, often suffer from under-sampling, poor generalization and limited network recovery. In this study, we explore previous integration strategies, such as batch-correction and model ensembles, and introduce a new multitask learning approach for joint network inference across several datasets. Our method initially estimates the activities of transcription factors, and subsequently, infers the relevant network topology. As regulatory interactions are context-dependent, we estimate model coefficients as a combination of both dataset-specific and conserved components. In addition, adaptive penalties may be used to favor models that include interactions derived from multiple sources of prior knowledge including orthogonal genomics experiments. We evaluate generalization and network recovery using examples from Bacillus subtilis and Saccharomyces cerevisiae, and show that sharing information across models improves network reconstruction. Finally, we demonstrate robustness to both false positives in the prior information and heterogeneity among datasets.

Leveraging chromatin accessibility for transcriptional regulatory network inference in T Helper 17 Cells

“…We included the highest confidence edges until we reached a model size of average 15 TFs/gene (3578 genes × 15 TFs/gene = 53,670 TF-gene interactions). Given the complementary performance of TF-mRNA and prior-based TFA, we combined resulting TRNs by taking the maximum edge confidence to preserve the individual strengths of each (Kittler et al 1996;Castro et al 2019). See Supplemental Note 8 and Supplemental Figures S33 and S34 for performance comparison of max-to rank-combine (Marbach et al 2012) relative to individual TRNs as well as performance combining TRNs from different priors.…”

Dual threshold optimization and network inference reveal convergent evidence from TF binding locations and TF perturbation responses

“…As long as the network inference algorithm does not use any binding data, DTO can provide independent, convergent evidence. There are also network inference algorithms that weigh and integrate data sources, including gene expression and TF binding location data or curated sources influenced by binding data (e.g., Siahpirani and Roy 2017;Wang et al 2018;Castro et al 2019). These algorithms are not suitable for our current purpose, which is to assess the convergence of independent evidence from gene expression and binding location data.…”

A 1p/19q Codeletion-Associated Immune Signature for Predicting Lower Grade Glioma Prognosis

“…Next, univariate Cox proportional hazards regression and LASSO (least absolute shrinkage and selection operator) Cox regression analyses were done on the 1p19q codeletion associated immune-related DEGs to prognosis-associated genes. The LASSO regression algorithm is used to reduce over tting highdimensional prognostic genes [12,13]. Multivariate Cox proportional hazards regression analysis was then used to establish a prognostic signature with a coe cient (β) based on all the genes included in the signature [14].…”