Mapping of genotype-by-environment interactions in phenology identifies two cues for flowering in switchgrass (Panicum virgatum)

Abstract

AbstractPlant phenological timings are major fitness components affected by multiple environmental cues; thus, phenological traits can have important genotype-by-environment interactions (GxE). We use a flexible, data-driven method to characterize GxE in the timing of vegetative growth (‘green-up’) and flowering across eight environments and in two highly divergent switchgrass (Panicum virgatum) populations. We classified polygenic GxE patterns as suggestive of modulation of genetic effects via weather-based cues—or other, unknown modulators. >26% of Gulf population SNPs affecting flowering had effects that covaried with photoperiodicity and >34% of Midwest upland population SNPs affecting flowering had effects that covaried with cumulative growing degree days. 76% of SNP effects on green-up showed antagonistic pleiotropy, a change in effect sign, between environments native to Gulf plants (Texas) and environments native to Midwest plants (North). In contrast, <2% of flowering effects showed antagonistic pleiotropy; the majority (>64%) showed no GxE. Top GxE-associated SNPs were highly enriched in the top associations from an independent pseudo-F2 cross of individuals from the same two populations. Breeding for particular alleles at GxE-associated loci could change flowering responsiveness to photoperiod cues in switchgrass. More broadly, this approach could be used to flexibly characterize patterns of GxE across species, populations and environments.Significance StatementThe timing of plant seasonal development (phenology) has major impacts on fitness because of the negative consequences of plant-environment mismatches. Here we map the genetic basis of two phenological events, the start of above-ground growth and flowering, in two genetically and phenologically distinct populations of switchgrass. We do this at eight field locations spanning the latitudinal range of both populations. Our approach allows us to identify regions of the genome with effects that covary with weather-related environmental features at every location. For flowering, these features differed by population: the Midwest population had genetic effects that primarily covaried with cumulative growing degree days, a temperature-related measure, while the Gulf population had genetic effects that primarily covaried with photoperiod, a day-length-related measure.