As any breeder will tell you, some hybrids exhibit vigor as compared to their parents, but others come up short. To better understand this seeming paradox of hybrid behavior, Enrico Coen at the John Innes Centre and his colleagues turned to studying gene expression in snapdragons. In particular, they focused on genes involved in flower asymmetry, CYC and RAD, and mapped how the expression levels of those genes, measured by RT-PCR and quantitative sequencing, affect plant phenotype. Their study was published in PLoS Biology.
Previously, Coen says, hybrid performance did not make much sense: if the fitness of the offspring was lower than that of the parents, then genes that evolved in one, or both, of the species reduced viability and if the fitness was better than the parents', those species had not evolved to their optimal level. "This really bothered Charles Darwin," Coen says.
To examine the question of hybrid performance, Coen and his colleagues devised a way to analyze genotype data, expression levels, and an index of morphological variations in three dimensions. With data from a variety of snapdragon mutants, they developed a 3D map of the gene expression-morphological space and found that the hybrids fall on a plateau, meaning that the variation seen in gene expression levels had no effect on the phenotype — it was just noise.
The researchers then turned to see whether there was an undetected effect of gene expression on the fitness of the snapdragons — a minor morphological change could have a major effect on fitness, Coen says. From this, they determined that hybrids were likely to do better with basic traits, such as growth, though worse with others.
To resolve the hybrid performance question, Coen says there may be mutations so slight that natural selection "can't scrutinize it." For example, Coen says that a mutation may reduce the fitness of a population by 0.001 percent and that just won't affect a large population — natural selection never detects it, he says. And so those populations can drift within that neutral space.
At the individual locus level, there won't be consequences, but across multiple loci that slight variation may have an effect. A hybrid, then, combines all those accumulated mutations from both populations. Since different regions were under different selection constraints, the hybrids can then seem to be either better or worse off than their source populations. If the variations cancel each other out or otherwise muffle the noise, the offspring shows vigor. But if they do not, or are additive, the hybrid seems worse off.
"It's an argument from statistics," Coen adds.