On the Consistency between Gene Expression and the Gene Regulatory Network of Corynebacterium glutamicum

Abstract

Background: Transcriptional regulation of gene expression is crucial for the adaptation and survival of bacteria. Regulatory interactions are commonly modeled as Gene Regulatory Networks (GRNs) derived from experiments such as RNA-seq, microarray and ChIP-seq. While the reconstruction of GRNs is fundamental to decipher cellular function, even GRNs of economically important bacteria such as Corynebacterium glutamicum are incomplete. Materials and Methods: Here, we analyzed the predictive power of GRNs if used as in silico models for gene expression and investigated the consistency of the C. glutamicum GRN with gene expression data from the GEO database. Results: We assessed the consistency of the C. glutamicum GRN using real, as well as simulated, expression data and showed that GRNs alone cannot explain the expression profiles well. Conclusion: Our results suggest that more sophisticated mechanisms such as a combination of transcriptional, post-transcriptional regulation and signaling should be taken into consideration when analyzing and constructing GRNs.

System Principles Governing the Organization, Architecture, Dynamics, and Evolution of Gene Regulatory Networks

“…The consistency between GRNs and expression data has been previously studied by assuming a causal effect between the expression of the TFs and their target genes (TGs). Recent studies using expression data in E. coli and C. glutamicum have assessed this causal effect by using correlations to show a weak correlation of the known regulatory TF-TG pairs compared to all the possible random pairs as background ( Larsen et al, 2019 ; Parise et al, 2021 ). Moreover, repressor interactions were associated with a positive correlation, rather than the expected negative correlation.…”

Improving gene regulatory network inference and assessment: The importance of using network structure

“…Therefore, inference tools based solely on gene expression data tend to also infer non-direct interactions, especially COEX tools (Figure 4C). Perhaps, this consideration may shed light on the search for consistency between GRNs and gene expression data (Larsen et al, 2019; Parise et al, 2021). On the other hand, every tool performs better with the “strong” GS on AUROC (Supplementary figure 12), but this is because of the highly unbalanced positives/negatives ratio (Saito and Rehmsmeier, 2015).…”

“…Recently, Simon Larsen et al performed an in-deep analysis on this matter, their results show that the correlation of pairs of random genes is indistinguishable from those involved in known regulatory interactions in E. coli (Larsen et al, 2019). Doglas Parise et al confirmed the results also on Corynebacterium glutamicum (Parise et al, 2021).…”

An Integrated Database of Small RNAs and Their Interplay With Transcriptional Gene Regulatory Networks in Corynebacteria

“…Several studies have shown the interplay between TFs and sRNAs when regulating gene expression by forming regulatory circuits, as reviewed by Beisel and Storz (2010); Nitzan et al (2017), Brosse and Guillier (2018). Furthermore, consistency assessments in E. coli (Larsen et al, 2019) and C. glutamicum (Parise et al, 2021) showed that regulation driven by transcription factors is not able to satisfactorily explain gene expression and suggested other layers of regulation to be integrated into the networks in order to model the complexity of gene expression. Our work contributes to expanding the regulatory landscape of two biotechnological and four pathogenic Corynebacterium species by predicting their sRNA regulatory networks and by integrating them into the corresponding GRNs.…”