1. As a first approximation, whole-body metabolic rate can be considered as the sum of metabolic rates of constituent cells. Yet, among several current explanations of metabolic rate scaling, only two explicitly invoke cell architecture of organisms: (1) the Metabolic Theory of Ecology, which predicts size invariance of metabolically active cells, such as erythrocytes and (2) the cell metabolism hypothesis postulating partial dependence of metabolic scaling on the cell size (CS), which is mediated by nucleus/genome size variation. 2. Here, we tested (1) and (2) by comparing standard metabolic rate (SMR), erythrocyte size (used as a proxy of CS) and nucleus size (NS) between diploid and triploid individuals of a small fish (body mass of c. 3 g) belonging to the Cobitis taenia hybrid complex. 3. We demonstrated a positive correlation of CS with genome/nucleus size and an inverse relationship between those traits and SMR. SMR scaled to body mass with a 0·92 ± 0·05 exponent, which significantly differed from the ¾ value, while CS scaled with body mass with an allometric exponent of 0·05 ± 0·007, which significantly differed from 0. Ploidy level explained c. 85% of CS variation. 4. Our results provide empirical support for CS and genome/nucleus size being important determinants of metabolic rate variation and consequently, its allometric scaling. They call attention to the significance of a long-neglected integration of cellular and organismal perspectives in studies of body size–metabolic rate relationships and their consequences for energy utilization in the wild.

Hybridization and polyploidization are important evolutionary processes whose impacts range from the alteration of gene expression and phenotypic variation to the triggering of asexual reproduction. We investigated fishes of the Cobitis taenia-elongatoides hybrid complex, which allowed us to disentangle the direct effects of both processes, due to the co-occurrence of parental species with their diploid and triploid hybrids. Employing morphological, ecological, and RNAseq approaches, we investigated the molecular determinants of hybrid and polyploid forms. In contrast with other studies, hybridization and polyploidy induced relatively very little transgressivity. Instead, Cobitis hybrids appeared intermediate with a clear effect of genomic dosing when triploids expressed higher similarity to the parent contributing two genome sets. This dosage effect was symmetric in the germline (oocyte gene expression), interestingly though, we observed an overall bias toward C. taenia in somatic tissues and traits. At the level of individual genes, expression-level dominance vastly prevailed over additivity or transgressivity. Also, trans-regulation of gene expression was less efficient in diploid hybrids than in triploids, where the expression modulation of homoeologs derived from the “haploid” parent was stronger than those derived from the “diploid” parent. Our findings suggest that the apparent intermediacy of hybrid phenotypes results from the combination of individual genes with dominant expression rather than from simple additivity. The efficiency of cross-talk between trans-regulatory elements further appears dosage dependent. Important effects of polyploidization may thus stem from changes in relative concentrations of trans-regulatory elements and their binding sites between hybridizing genomes. Links between gene regulation and asexuality are discussed.