Los Angeles doesn’t really get full-on summer heat until September, after months of building warmth and time elapsed since that last gasp of winter rains and spring fog. This year we (and most of the rest of the western U.S.) were hit with a record-breaking heatwave, which locally meant daily high temperatures above 100°F (38°C) from August 31 to September 8 — and then on the 9th, we were rescued by an incoming tropical storm. Through it all I hiked across campus from the faculty parking lot every morning, various short-sleeved dress shirts stuck to my back with sweat, and then tweeted excitedly when relief stirred the surface of my apartment building’s swimming pool.
And here’s some of what I read, in my thoroughly air-conditioned campus office:
Abdellaoui A, CV Dolan, KJH Verweij, and MG Nivard. 2022. Gene-environment correlations across geographic regions affect genome-wide association studies. Nature Genetics doi: 10.1038/s41588-022-01158-0
GWAS in humans (but also in other organisms where we can’t get phenotypes measured in common-garden-type experiments) is complicated by the possibility of environmental confounding, in which genetic variation aligns with environmental factors that affect a phenotype of interest. The authors here isolate gene-environment correlations in data used to estimate polygenic scores for income and educational attainment in a large (N = ~43,000) dataset of siblings from the UK, and re-run complex trait GWAS in an even larger (N = ~250,000) set of UK participants, controlling for regional gene-environment correlations.
- Controlling for gene-environment correlations significantly reduced estimated heritability of multiple traits, led by a bunch of socio-economic measures.
- But also, maybe more surprisingly, controlling for regional gene-environment correlations decreased heritability of (nominal) health indicators like BMI and waist circumference.
- The authors usefully distinguish between “active” mechanisms creating gene-environment correlations, and “passive” ones — the former are the result of individuals moving to environments that may change a complex phenotype (i.e., to a better school district), the latter are things like regional government policy.
Moutinho AF, A Eyre-Walker, and JY Dutheil. 2022. Strong evidence for the adaptive walk model of gene evolution in Drosophila and Arabidopsis. PLOS Biology doi: 10.1371/journal.pbio.3001775
Classic theory predicts that early steps in an “adaptive walk” should be larger in magnitude than later ones — because new mutations with large effects relative to the distance between a population’s current average trait value and the nearest adaptive optimum are more likely to take individuals away from that optimum than closer to it. This should apply to genes as they age: when a gene is newly created by a duplication event and hasn’t fully settled into a novel function, it can “afford” big mutational changes; but when it is established in the biology of an organism, big changes are more likely to break the gene’s function than improve it. The authors assess this using dN/dS ratios for annotated adaptive and nonadaptive variation in Drosophila and Arabidopsis data, and find that
- Gene age is indeed strongly correlated with overall dN/dS, and both adaptive and nonadaptive dN/dS, consistent with more rapid adaptive evolution of young genes;
- dN/dS ratios are also correlated with known confounding factors such as gene expression (young genes are less expressed) protein length, and structural features of the proteins that may reduce purifying selection;
- But the overall signal is generally robust to systematically controlling for those confounding factors, with some exceptions in the Arabidopsis data;
- They further apply a new variant of the dN/dS (McDonald-Kreitman) test that allows accounting for multiple confounding variables simultaneously, which looks interesting.
Anghel I et al. 2022. Reference genome of the color polymorphic desert annual plant Linanthus parryae. Journal of Heredity doi: 10.1093/jhered/esac052
Linanthus parryae is an annual wildflower found in the Mojave Desert and adjacent arid lands, growing in low carpets of white and blue flowers after winter rains. Its stable color polymorphism has been a model for development of some key ideas in evolution and ecology, including isolation by distance and good old FST. And now it has a reference genome assembly, covering 1.5Gb of an estimated ~2Gb genome (about 2/3rds of a human genome) with long-read HiFi data and chromatin conformation capture. Almost 72% of the assembly is annotated as repetitive elements, which is high for flowering plants but apparently aligns with the only other assembly for a member of the Polemoniaceae, Gilia yorkii. The authors emphasize potential conservation value for the genome, citing development, especially renewable energy build-out, and disruptive recreational usage in the Mojave.