Identification of Ideal Allele Combinations for the Adaptation of Spring Barley to Northern Latitudes

Abstract

The northwards expansion of barley production requires adaptation to longer days, lower temperatures and stronger winds during the growing season. We have screened 169 lines of the current barley breeding gene pool in the Nordic region with regards to heading, maturity, height, and lodging under different environmental conditions in nineteen field trials over 3 years at eight locations in northern and central Europe. Through a genome-wide association scan we have linked phenotypic differences observed in multi-environment field trials (MET) to single nucleotide polymorphisms (SNP). We have identified an allele combination, only occurring among a few Icelandic lines, that affects heat sum to maturity and requires 214 growing degree days (GDD) less heat sum to maturity than the most common allele combination in the Nordic spring barley gene pool. This allele combination is beneficial in a cold environment, where autumn frost can destroy a late maturing harvest. Despite decades of intense breeding efforts relying heavily on the same germplasm, our results show that there still exists considerable variation within the current breeding gene pool and we identify ideal allele combinations for regional adaptation, which can facilitate the expansion of cereal cultivation even further northwards.

Keywords: GWAS; Hordeum vulgare; QTL; earliness; maturity; plant breeding; plant height.

FIGURE 1

An overview of the multi-environment field trials (MET). (A) Field tests were performed at eight different locations in six different countries, ranging from Laberweinting, Germany in the south to Korpa, Iceland in the north, for up to 3 years, resulting in 19 distinct environments. The number of years is shown inside circles. (B) The different day lengths between test locations, with the respective sowing days of the trials. (C) The accumulated heat sum (°C) from sowing day until maturity of the latest maturing line in the eight trials where maturity was screened. The highest and lowest value are shown in the graph (at Korpa, IS 2013 not all lines reached maturity).

FIGURE 2

GWAS results for the traits early spring growth stages 31, 32, and 34 (Zadoks et al., 1974), for panels (A) PPP169 and (B) PPP124, respectively. The Manhattan plots show significant associations (Bonferroni threshold in red, and the suggestive threshold in purple) between trait and marker, the x-axis shows the physical distance over all seven barley chromosomes. The bar under the x-axis shows the SNP distribution on each chromosome, where 0 ->18 depicts SNP density (the number of SNPs per 1 Mbp bin). The non-significant and significant associations are displayed as open and solid filled circles, respectively. The significance of the associations at values between these two extremes are displayed as circles with varying amounts of fill. Relevant QTL names and putative loci are presented in the figure. Note that the ^∗ indicates a SNP marker located at 3H position 105 cM using the POPSEQ 2017 reference map (http://floresta.eead.csic.es/barleymap/).

FIGURE 3

GWAS results for the traits straw length, straw breaking and lodging for panels (A) PPP169 and (B) PPP124, respectively. The Manhattan plots show significant associations (Bonferroni threshold in red, and the suggestive threshold in purple) between trait and marker, the x-axis shows the physical distance over all seven barley chromosomes. The bar under the x-axis shows the SNP distribution on each chromosome, where 0 ->18 depicts SNP density (the number of SNPs per 1 Mbp bin). The non-significant and significant associations are displayed as open and solid filled circles, respectively. The significance of the associations at values between these two extremes are displayed as circles with varying amounts of fill. The most significant values, located at the top, are displayed as solid filled dots. The significance of the associations at values between these two extremes is indicated by all other dots with varying amounts of fill. Relevant QTL names and putative loci are presented in the figure.

FIGURE 4

GWAS results for the traits heat sum heading day, heat sum maturity, and heat sum grain filling period for panels (A) PPP169 and (B) PPP124, respectively. The Manhattan plots show significant associations (Bonferroni threshold in red, and the suggestive threshold in purple) between trait and marker, the x-axis shows the physical distance over all seven barley chromosomes. The bar under the x-axis shows the SNP distribution on each chromosome, where 0 ->18 depicts SNP density (the number of SNPs per 1 Mbp bin). The non-significant and significant associations are displayed as open and solid filled circles, respectively. The significance of the associations at values between these two extremes are displayed as circles with varying amounts of fill. The most significant values, located at the top, are displayed as solid filled dots. The significance of the associations at values between these two extremes is indicated by all other dots with varying amounts of fill. Relevant QTL names and putative loci are presented in the figure.

FIGURE 5

GWAS results for spike morphology for panel PPP169. The Manhattan plots show significant associations (Bonferroni threshold in red, and the suggestive threshold in purple) between trait and marker, the x-axis shows the physical distance over all seven barley chromosomes. The bar under the x-axis shows the SNP distribution on each chromosome, where 0 ->18 depicts SNP density (the number of SNPs per 1 Mbp bin). The non-significant and significant associations are displayed as open and solid filled circles, respectively. The significance of the associations at values between these two extremes are displayed as circles with varying amounts of fill. The most significant values, located at the top, are displayed as solid filled dots. The significance of the associations at values between these two extremes is indicated by all other dots with varying amounts of fill. Relevant QTL names and putative loci are presented in the figure.

FIGURE 6

Boxplot of allele combinations showing BLUEs for heat sum to maturity (HSMD) (°C) in the PPP169 panel. Allele combinations were based on three significant markers that passed the suggestive threshold value in the GWAS analysis. The effect of each allele combination for combinations with at least five observations (lines) was calculated based on BLUE values and the significance of the effects was tested using the lm () function in R. ^∗ marks significant combinations (p ≤ 0.05). Allele combinations followed by the same lower case letter do not differ statistically.