Common gardens in teosintes reveal the establishment of a syndrome of adaptation to altitude

Abstract

In plants, local adaptation across species range is frequent. Yet, much has to be discovered on its environmental drivers, the underlying functional traits and their molecular determinants. Genome scans are popular to uncover outlier loci potentially involved in the genetic architecture of local adaptation, however links between outliers and phenotypic variation are rarely addressed. Here we focused on adaptation of teosinte populations along two elevation gradients in Mexico that display continuous environmental changes at a short geographical scale. We used two common gardens, and phenotyped 18 traits in 1664 plants from 11 populations of annual teosintes. In parallel, we genotyped these plants for 38 microsatellite markers as well as for 171 outlier single nucleotide polymorphisms (SNPs) that displayed excess of allele differentiation between pairs of lowland and highland populations and/or correlation with environmental variables. Our results revealed that phenotypic differentiation at 10 out of the 18 traits was driven by local selection. Trait covariation along the elevation gradient indicated that adaptation to altitude results from the assembly of multiple co-adapted traits into a complex syndrome: as elevation increases, plants flower earlier, produce less tillers, display lower stomata density and carry larger, longer and heavier grains. The proportion of outlier SNPs associating with phenotypic variation, however, largely depended on whether we considered a neutral structure with 5 genetic groups (73.7%) or 11 populations (13.5%), indicating that population stratification greatly affected our results. Finally, chromosomal inversions were enriched for both SNPs whose allele

Transcriptome Analysis Reveals Potential Roles of Abscisic Acid and Polyphenols in Adaptation of Onobrychis viciifolia to Extreme Environmental Conditions in the Qinghai-Tibetan Plateau

“…It is well documented that phenotypic traits of plants are influenced by environmental gradients [ 65 , 66 , 67 ]. Intriguingly, some phenotypic traits of O. viciifolia plants grown at two different altitudes showed a great difference according to our filed survey in the last three years.…”

“…Parviglumis and Z. mays ssp. mexicana ) were decreased with the increase of elevation [ 66 ]. Since decreased tiller numbers could also be triggered by drought stress [ 23 ], drier and cooler climate in the high-altitude TD region might cause decrease in tiller number of O. viciifolia plants grown in that area in comparison with those grown in the low-altitude MH.…”

Evaluation of the minimum sampling design for population genomic and microsatellite studies. An analysis based on wild maize

“…Given the proximity of teosintes to maize, different genomic resources are available (Hufford et al, 2012; Aguirre-Liguori et al, 2016) and several studies have analyzed their population genomics (van Heerwaarden et al, 2010, 2011; Hufford et al, 2013; Pyhäjärvi et al, 2013; Aguirre-Liguori et al, 2017, 2019a,b; Fustier et al, 2017, 2019; Moreno-Letelier et al, 2018). Briefly, genomic studies suggest that teosintes have high genetic diversity, show patterns of isolation by distance and environment, and show strong patterns of local adaptation (Pyhäjärvi et al, 2013; Aguirre-Liguori et al, 2017, 2019a, b; Fustier et al, 2017, 2019). The vast genomic resources and biological knowledge makes teosintes an ideal system to study the importance of sampling design in analyses of genetic diversity, isolation patterns and identification of candidate SNPs, since sampling biases can be compared with the available knowledge.…”

Polygenic adaptation of rosette growth in Arabidopsis thaliana

“…In humans, where population structure and environmental variation are correlated, insufficient correction of the genetic associations caused by shared ancestry has been shown to create spurious associations [55][80]. Even though environmental variance is much better controlled in common garden experiments including kinship as a covariate, association tests can still be confounded by genetic relatedness [56]. This is of particular concern when many trait/SNP associations are below the Bonferroni significance threshold.…”