1. Association between SSR markers and
fiber related traits in an exotic germplasm
derived from multiple crosses among
tetraploid species in Gossypium
Linghe Zeng and William R. Meredith
Crop Genetics & Production Unit,
USDA-ARS, Stoneville, MS

2. Association Mapping
Association - “Polymorphic fragments are identified with significant higher frequencies in one type of phenotype than the contrasting phenotype” (Schafer and Hawkings)
Definition – an approach to identify marker-phenotype associations based on linkage disequilibrium in natural populations, cultivars, breeding lines, or breeder’s populations

3. Association mapping vs. linkage-based QTL mapping
Linkage-based Association mapping
1. Identify polymorphic Identify polymorphic markers
markers in a given in cultivars, breeding lines,
segregating population natural populations, or
breeder’s populations
2. Require linkage info not necessary
among markers before
association analysis
3. Genetic distance LD, population structure
4. Two alleles multiple alleles
5. Contrasting phenotype limited phenotype

4. Procedures
1. Identify polymorphic markers in cultivars or natural populations
2. Estimate association: mean comparisons of marker class (1, 0) using experiment wise t tests (permutation>1000)
3. Analyze population structure
4. Analyze genetic differentiation (Ht, Fst) in population
5. Analyze association with population structure as a factor

5. Objectives
* Screen polymorphic SSR markers in a germplasm derived from multiple crosses among tetraploid species in Gossypium
* analyze population structure and their effects on associations.
* Identify and confirm associations between SSR markers and fiber related traits with consideration of population structure

6. * Initiated by P.A. Miller in 1967
SP Germplasm
* Initiated by P.A. Miller in 1967
* Crossing 12 cultivars and strains in G. hirsutum with G. barbadense, G. tomentosum, G. mustelinum, and G. darwinii.
* Partial random mating in F2 and subsequent 10 generations at Raleigh, NC
* Predominant selfing for 12 generations at Stoneville, MS (2,000 plants)
* 260 plants were samples and 20 bolls were collected from each plant in 2004.

7. Field Evaluation
* Experiment design: 260 lines; Randomized complete block with three locations and 1-2 reps at each location
* Three environments: Env1=Field Location1, 2005; Env2=Field Location2, 2005; Env3=Field Location1, 2006.
* Single row plots (4.57 m long, 1 m row space); 30 bolls collected each plot; saw gin; lint percent and boll wt were measured separately from each plot
* Fiber quality was analyzed at Starlab, Knoxville, TN

8. Effects of genotype and environment (fiber properties)
Source Micronaire Elongation Strength Length Length
(50%) (2.5%)
G *** *** *** *** 0.012***
E *** ** *** *** ***
G x E * *** ***
Error

9. An example of SSR by capillary electrophoresis (ABI3730XL DNA Analyzer)
BNL285

10. Population Structure Analysis
Software: STRUCTURE v. 2.1, AFLP-SURV 1.0
Admixture model: assume mixed ancestral backgrounds in SP lines
Allele frequency model: Allele Frequencies Correlated
Burning length: 100,000
Replication: 50,000
K (number of groups): 1 to 10; p(X︱Y)
Genetic differentiation:
Ht, Fst using AFLP-SURV 1.0 (Vekemans, 2002)

11. * Phenotype = structure + marker + structure*marker + residue
Association Model
* Phenotype = structure + marker + structure*marker + residue
* ‘Within Group’ – Reduce genetic differentiation (Fst values) among groups by removing sub-structures
* Significant ‘Within Group’ – confirm association (independent of population structure)
* Significant S x M, non-significant ‘Within Group’ – false association due to substructure
* Non-significant S x M, non-significant ‘Within Group’ – false association, attributed to ?

12. Screening for polymorphic SSR
* 84 SSR primer pairs
* 296 polymorphic fragments
* 3.52 polymorphic fragments per primer pair
* Assumption was made for convenience: Each polymorphic fragment is a locus

13. Results of structure analysis(1)
K = 6 based on p(X︱Y)
Structure No. lines probability
Group >0.75
Group >0.75
Group >0.75
Group >0.75
Group >0.75
Group >0.75
‘Mixed group’ <0.75

14. Results of structure analysis(2)
Structure lines Ht Fst p values
Overall <0.0001
Group <0.0001
+ ‘Mixed’

15. Averages of fiber related traits among population structures (1)
Structure lines Lint percent Boll wt Micronaire
(number) (%) (g/boll)
Group
Group
Group
Group
Group
Group
Mixed group
Significance p< p< p<0.01

16. Association between SSR and traits (1)
Marker Class Line Lint % Boll wt Micronaire
BNL *** *
Interaction p= p<0.01
‘Within group p< p<0.01
BNL *** *
Interaction p< p=0.61
‘Within group’ p= p<0.001
CIR ***
Interaction p=0.163
‘Within group’ p=0.318

17. Number of markers with ‘confirmed’ association or ‘false’ association
Significance Significance False assoc.
(t tests) (‘Within grp’) (structure)
Lint%
Boll wt
MIC E T
Length(50%)
Length(2.5%)

18. A list of markers with confirmed association
BNL – lint%(r2=0.29) BNL elongation(r2=0.23)
BNL lint%(r2=0.15) CIR elongation(r2=0.21)
JR – lint%(r2=0.25) CIR elongation(r2=0.21)
BNL boll wt(r2=0.15) BNL strength(r2=0.19)*
BNL boll wt(r2=0.22) BNL – length(50%)(r2=0.27)
BNL boll wt(r2=0.32) BNL – length(50%)(r2=0.23)*
BNL boll wt(r2=0.19) BNL40971 – length(2.5%)(r2=0.24)
CIR boll wt(r2=0.26) BNL – length(2.5%)(r2=0.18)
BNL micronaire(r2=0.20) BNL – length(2.5%)(r2=0.26)
BNL MIC(r2=0.18) BNL – length(2.5%)(r2=0.25)*
CIR micronaire(r2=0.38)* BNL – length(2.5%)(r2=0.36)
CIR9985-MIC(r2=0.21) BNL length(2.5%)(r2=0.13)
CIR10590-MIC(r2=0.17) CIR – length(2.5%)(r2=0.22)

19. Conclusion
* Twenty-seven marker-trait associations confirmed with seven fiber related traits
* No effect of population structure for marker-micronaire association
* Moderate effect of population structure for boll wt, elongation, and length(2.5%)
* Dramatic effect population structure for lint percent and length(50%)

20. ACKNOWLEDGEMENT
Dr. Brian Scheffler and Ms. Sheron Simpson (MSA Genomics Laboratory).
Dr. Jack C. McCarty (USDA-ARS, Mississippi State University
Osman A. Gutiérrez (Mississippi State Universtiy)
Dr. Jodi Scheffler, Dr. Jeffery Ray (USDA-ARS, Stoneville, MS
Ms. Deborah Boykin (USDA-ARS, Mid South area Statistician)

21. Thanks