Association Mapping for Important Agronomic Traits in Safflower (Carthamus tinctorius L.) Core Collection Using Microsatellite Markers

Abstract

Carthamus tinctorius L. (safflower) is an important oilseed crop producing seed oil rich in unsaturated fatty acids. Scarcity of identified marker-trait associations is a major limitation toward development of successful marker-assisted breeding programs in safflower. In the present study, a safflower panel (CartAP) comprising 124 accessions derived from two core collections was assayed for its suitability for association mapping. Genotyping of CartAP using microsatellite markers revealed significant genetic diversity indicated by Shannon information index (H = 0.7537) and Nei's expected heterozygosity (I = 0.4432). In Principal Coordinate Analysis, the CartAP accessions were distributed homogeneously in all quadrants indicating their diverse nature. Distance-based Neighbor Joining analysis did not delineate the CartAP accessions in consonance with their geographical origin. Bayesian analysis of population structure of CartAP demonstrated the unstructured nature of the association panel. Kinship analysis at population (G_ij ) and individual level (F_ij ) revealed absence of or weak relatedness between the CartAP accessions. The above parameters established the suitability of CartAP for association mapping. We performed association mapping using phenotypic data for eight traits of agronomic value (viz., seed oil content, oleic acid, linoleic acid, plant height, number of primary branches, number of capitula per plant, 100-seed weight and days to 50% flowering) available for two growing seasons (2011-2012 and 2012-2013) through General Linear Model and Mixed Linear Model. Our study identified ninety-six significant marker-trait associations (MTAs; P < 0.05) of which, several MTAs with correlation coefficient (R²) > 10% were consistently represented in both models and in both seasons for traits viz., oil content, oleic acid content, linoleic acid content and number of primary branches. Several MTAs with high R²-values were detected either in a majority or in some environments (models and/or seasons). Many MTAs were also common between traits (viz., oleic/linoleic acid content; plant height/days to 50% flowering; number of primary branches/number of capitula per plant) that showed positive or negative correlation in their phenotypic values. The marker-trait associations identified in this study will facilitate marker-assisted breeding and identification of genetic determinants of trait variability.

Keywords: SSR markers; association mapping; core collections; kinship analysis; population structure; safflower.

Figure 1

(A) Neighbor Joining dendrogram elucidating the genetic relationships between 124 CartAP accessions using 93 SSR loci based on simple matching coefficient. NJ clusters (NJ I–NJ III) are demarcated through dashed lines. (B) Scatter plot for Principal Coordinate Analysis (PCoA) exhibiting the distribution of CartAP accessions on two main axes. Primary axes 1 and 2 captured 29 and 25% of the total variance, respectively. Color codes used for different regional pools are provided.

Figure 2

Population structure of CartAP collection inferred using Bayesian-based program STRUCTURE. (A) Estimation of hypothetical sub-populations using ΔK-values. The number of identified sub-populations were two (K = 2). (B) Population structure analysis of CartAP accessions at K = 2 based on inferred ancestry (Q matrix). The sub-populations were named STR I and STR II.

Figure 3

Hierarchical population structure analysis (A) Estimation of sub-population for STR I. Maximum number of sub-populations were inferred at K = 2. (B) Population structure for STR I at K = 2 based on Q matrix. The sub-populations were designated as STR I(a) and STR I(b) (C) Estimation of sub-populations for STR II. Maximum number of subpopulations were inferred at K = 5. (D) Population structure for STR II at K = 5 based on Q matrix. The sub-populations were named as STR II(a)–STR II(e). Color codes for each sub-population are provided.

Figure 4

Matrix showing pairwise F_ST-values between sub-populations inferred through hierarchical STRUCTURE analysis. Color codes are provided for each sub-population. ^*Denotes F_ST-values between same sub-population.

Figure 5

Kinship estimates at individual and population level. (A) Frequency distribution for global pairwise kinship estimates (F_ij) for CartAP accessions. (B) Matrix showing pairwise G_ij-values between sub-populations inferred through hierarchical STRUCTURE analysis. Color codes are provided for each sub-population. ^*Denotes G_ij-values values between same sub-population.

Figure 6

Demarcation of two major STRUCTURE populations (STR I and STR II) and seven sub-populations obtained by further hierarchical structure analysis on (A) PCoA scatter plot. (B) Neighbor Joining dendrogram. Color codes are provided for each sub-population.