Genotyping-by-sequencing derived SNP markers reveal genetic diversity and population structure of Dactylis glomerata germplasm

Abstract

Orchardgrass (Dactylis glomerata L.), a widely cultivated cool-season perennial, is an important forage crop due to its adaptability, high nutritional value, and substantial biomass. Understanding its genetic diversity and population structure is crucial for developing resilient cultivars that can withstand climate change, diseases, and resource limitations. Despite its global significance in fodder production, the genetic potential of many regional accessions remains unexplored, limiting breeding efforts. This study investigates the genetic diversity (GD) and population structure of 91 accessions of D. glomerata from Turkey and Iran using genotyping-by-sequencing based single nucleotide polymorphism (SNP) markers. A total of 2913 high-quality SNP markers revealed substantial genetic variability across provinces. Notably, accessions from Erzurum exhibited the highest GD (mean GD: 0.26; He: 0.5328), while provinces such as Bursa and Muğla demonstrated lower GD (mean GD: 0.15; He < 0.22), suggesting potential genetic bottlenecks. Population structure analysis using Bayesian clustering, PCoA and UPGMA dendrograms divided the accessions into three distinct clusters, with cluster membership largely reflecting geographical origins, and dry biomass content. Cluster II revealed higher GD, associated with enhanced biomass production (128 g/plant), the most important agronomic trait in forage species, supporting the notion of heterosis in breeding programs. The majority of the genetic variation (85.8%) was observed within clusters, with minimal differentiation among clusters (FST = 0.007). Genome-wide association studies (GWAS) identified significant marker-trait associations for dry biomass weight, a critical agronomic trait, with markers DArT-100715788, DArT-101043591, and DArT-101171265 and DArT-101090822 located on Chromosomes 1, 6, and 7 respectively. These findings highlight the importance of regional diversity for maintaining adaptive potential in future breeding programs.

Keywords: Dactylis glomerata; GWAS; SNP markers; forage breeding; genetic diversity; population structure.

Figure 1

Collection points of the studied Dactylis glomerata L germplasm.

Figure 2

Analysis of the optimum number of populations (K) for estimating genetic structure with SNP data. The post hoc method proposed by Evanno et al. (2005) was used with 100 replicates of K ranging from 1 to 12, considering their mean likelihood (A) and delta K (B) as criteria for selection of the most parsimonious model.

Figure 3

Genetic relationships and population structure for 91 D. glomerata accessions based on 2937 SNP markers. (A) UPGMA dendrogram based Hamming genetic distance (GD). Major clusters (I, II, and III) and sub-clusters within clusters I (Ia, Ib) and II (IIa, IIb) are delimited by bold black and gray dashed lines. (B) Genetic structure of the accessions considering different optimal number of populations (i.e., K=2, K=3, and K=4). Each accession is represented by a horizontal bar partitioned into two (K=2), three (K=3), or four-colored segments (K=4), indicating their relative membership to the considered clusters. ISO country codes are as follows: TR: Turkey; IR: Iran. Abbreviations of province names are as follows: Afy, Afyon; Ağr, Ağrı; Ank, Ankara; Ant, Antalya; Ard, Ardahan; Ayd, Aydin; Bay, Bayburt; Bin, Bingöl; Bur, Bursa; Çan, Çanakkale; Ela, Elazığ; Erd, Erdebil (Iran); Erzi, Erzincan; Erz, Erzurum; Güm, Gümüşhane; Ist, Istambul; Kar, Kars; Kay, Kayseri; Mal, Malatya; Muğ, Muğla; Sam, Samsun; Siv, Sivas; Tun, Tunceli; Yoz, Yozgat. Dry biomass values are expressed in grams of dry weight per plant (g dw/plant) and quartiles for this trait are color-coded.

Figure 4

Principal coordinate analysis (PCoA) of the 91 D. glomerata accessions using SNP markers.

Figure 5

Histogram of dry biomass weight based on UPGMA clustering. Bars represent mean values for each cluster ± standard error. Mean values with different letters are statistically significant using least significant difference (LSD) test at p<0.05.

Figure 6

Pseudo Manhattan plot for dry biomass weight in studied Dactylis glomerata germplasm.