Condition-Specific Modeling of Biophysical Parameters Advances Inference of Regulatory Networks

Abstract

SUMMARY Large-scale inference of eukaryotic transcription-regulatory networks remains challenging. One underlying reason is that existing algorithms typically ignore crucial regulatory mechanisms, such as RNA degradation and post-transcriptional processing. Here, we describe InfereCLaDR, which incorporates such elements and advances prediction in Saccharomyces cerevisiae. First, InfereCLaDR employs a high-quality Gold Standard dataset that we use separately as prior information and for model validation. Second, InfereCLaDR explicitly models transcription factor activity and RNA half-lives. Third, it introduces expression subspaces to derive condition-responsive regulatory networks for every gene. InfereCLaDR’s final network is validated by known data and trends and results in multiple insights. For example, it predicts long half-lives for transcripts of the nucleic acid metabolism genes and members of the cytosolic chaperonin complex as targets of the proteasome regulator Rpn4p. InfereCLaDR demonstrates that more biophysically realistic modeling of regulatory networks advances prediction accuracy both in eukaryotes and prokaryotes.

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

“…We tested two methods for TF activity estimation: (1) based on TF mRNA levels and (2) based on prior knowledge of TF-gene interactions. Although prior-based TFA improved TRN inference in Bacillus subtilis and yeast (Arrieta-Ortiz et al 2015;Tchourine et al 2018), neither method consistently outperformed the other in this study. As a result, final TRNs were built using both TFA methods.…”

“…TFA estimation based on prior knowledge of TF target genes provides an alluring alternative as it appears to be technically feasible, requiring only partial a priori knowledge of TF-gene interactions and gene expression data (Methods). "Prior-based" TFAs improved TRN inference in unicellular organisms (Arrieta-Ortiz et al 2015;Tchourine et al 2018). Here, we evaluate this approach in a mammalian setting.…”

“…where X ∈ R |genes|×|samples| is the expression matrix for genes in the prior, P ∈ R |genes|×|TFs| is the prior matrix of known TF-gene interactions, and A ∈ R |TFs|×|samples| contains the unknown TFAs. Equation 2 has no unique solution, but the least-squares solution has worked well in simpler organisms (Arrieta-Ortiz et al 2015;Tchourine et al 2018). Given first-time application to a mammalian setting, we tested both methods of TFA estimation: (1) TF mRNA levels and (2) prior-based (Equation 2).…”

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

“…We validated our approach using two model organisms, a gram-positive bacteria, B. subtilis , and an eukaryote, S. cerevisiae . Availability of validated TF-target regulatory interactions, hereafter referred to as the gold-standard, make both organisms a good choice for exploring inference methods (3040 interactions, connecting 153 TFs to 1822 target genes for B. subtilis [17, 46], 1198 interactions connecting 91 TFs to 842 targets for S. cerevisiae [51]). For B. subtilis , we use two expression datasets.…”

“…In the absence of such information, we hypothesized that orthogonal high-throughput datasets would provide insight. Because the yeast gold-standard [51] was built as a combination of TF-binding (ChIP-seq, ChIP-ChIP) and TF knockout datasets available in the YEASTRACT [47] and the SGD [57] databases, we propose to derive prior knowledge from chromatin accessibility data [22, 23] and TF binding sites [58] (as this is a realistic and efficient genomic experimental design for non-model organisms). Open regions in the genome can be scanned for transcription factor binding sites, which can provide indirect evidence of regulatory function [59].…”