Evolution of regulatory networks associated with traits under selection in cichlids

Tarang K Mehta,Federica Di-Palma,Federica Di-Palma,Christopher Koch,Wilfried Haerty,Sara A Knaack,Luca Penso-Dolfin,Tamas Korcsmaros,Marton Olbei,Sarah Bastkowski,Sushmita Roy,Will Nash,Padhmanand Sudhakar

doi:10.1186/s13059-020-02208-8

Abstract

BackgroundSeminal studies of vertebrate protein evolution speculated that gene regulatory changes can drive anatomical innovations. However, very little is known about gene regulatory network (GRN) evolution associated with phenotypic effect across ecologically diverse species. Here we use a novel approach for comparative GRN analysis in vertebrate species to study GRN evolution in representative species of the most striking examples of adaptive radiations, the East African cichlids. We previously demonstrated how the explosive phenotypic diversification of East African cichlids can be attributed to diverse molecular mechanisms, including accelerated regulatory sequence evolution and gene expression divergence.ResultsTo investigate these mechanisms across species at a genome-wide scale, we develop a novel computational pipeline that predicts regulators for co-extant and ancestral co-expression modules along a phylogeny, and candidate regulatory regions associated with traits under selection in cichlids. As a case study, we apply our approach to a well-studied adaptive trait—the visual system—for which we report striking cases of network rewiring for visual opsin genes, identify discrete regulatory variants, and investigate their association with cichlid visual system evolution. In regulatory regions of visual opsin genes, in vitro assays confirm that transcription factor binding site mutations disrupt regulatory edges across species and segregate according to lake species phylogeny and ecology, suggesting GRN rewiring in radiating cichlids.ConclusionsOur approach reveals numerous novel potential candidate regulators and regulatory regions across cichlid genomes, including some novel and some previously reported associations to known adaptive evolutionary traits.

Highlights

Seminal studies of vertebrate protein evolution speculated that gene regulatory changes can drive anatomical innovations
Orthologous genes of each species can be assigned to non-orthologous modules (Fig. S-R1a), indicative of co-expression divergence and potential transcriptional rewiring from the last common ancestor (LCA); this is referred to as “state changes” in module assignment
To ensure orthologous genes of all branches are included in subsequent analysis, we focused on state changes of 6844 1-to-1 orthologous genes to assess convergent and unique state changes along the phylogeny (Fig. 1b)

Summary

Introduction

Seminal studies of vertebrate protein evolution speculated that gene regulatory changes can drive anatomical innovations. Seminal studies by King and Wilson [1] analyzing protein evolution in vertebrates speculated the importance of evolutionary changes in “regulatory processes” for morphological diversity [2, 3] These ideas were soon expanded on by François Jacob [4], Mehta et al Genome Biology (2021) 22:25 who suggested that the molecular “tinkering” of pre-existing systems is a hallmark of evolution where, for example, regulatory processes can either be transformed or combined for functional gain [4]. While there are efforts to collate and integrate several genomic datasets for vertebrates, including human and mouse [14], comparative analysis of regulatory networks from these data alone remains a major computational challenge and very little is known about the phenotypic effect of genome-wide regulatory network rewiring events in non-model vertebrates [15]

Methods

Results

Discussion

Conclusion