The document discusses various methods for high-throughput genotyping used in molecular breeding, including SNP arrays, genotyping by sequencing, and TaqMan assays. It also covers marker-assisted selection strategies and factors critical for successful molecular breeding, such as choice of germplasm, accurate phenotyping, high-quality genotyping data, and involvement of breeders. Genome-wide association studies are presented as a method for population mapping of traits compared to bi-parental mapping of families.

1. Application of high-throughput
genotyping to molecular breeding
From “too many” SNPs to the field
Diego Micheletti

2. High Throughput Marker Analysis
• SNP Arrays
• Illumina Infinum assay (Bead Chip assay)
• Affymetrix Axiom genotyping
• TaqMan SNP Genotyping Assays - Life Technologies
• Genotyping By Sequencing (GBS)

3. Cost Effective Molecular Breeding Strategies
The cost of molecular markers application is now very low.
Whole genome scan (mainly for parents) is becoming a routine
application instead of single marker screening.

4. Marker Assisted Selection
Assumption: DNA markers can reliably predict phenotype
Very useful when selecting for:
• Low heritability traits (Traits dependent from the
environment)
• Traits hard to measure (aroma)
• Traits measured late in the plant life cycle (Fertility genes)
• Disease resistance genes (Multiple assay needed for each
selection)

5. Critical factors for the success of molecular
breeding
Germplasm
• Elite vs. non-elite
• Wild relatives
• Structure

6. Critical factors for the success of molecular
breeding
Germplasm
• Elite vs. non-elite
• Wild relatives
• Structure
Phenotyping
• Identify key traits
• Stress management
• Accurate

7. Critical factors for the success of molecular
breeding
Germplasm
• Elite vs. non-elite
• Wild relatives
• Structure
Genotyping
• High-throughput level
• Cost-effectiveness
• Informatics
• Data management
Phenotyping
• Identify key traits
• Stress management
• Accurate

8. Critical factors for the success of molecular
breeding
Germplasm
• Elite vs. non-elite
• Wild relatives
• Structure
Genotyping
• High-throughput level
• Cost-effectiveness
• Informatics
• Data management
Phenotyping
• Identify key traits
• Stress management
• Accurate
Breeding
• Choice of MB strategy
• Breeders fluent with MB
• Stakeholders involvement

9. QTL mapping
Family mapping
• Developing of the experimental
population
• Collection of genotypic and
phenotypic data
• Construction of linkage map
• Linkage analysis between the
trait and the markers
• Identification of QTL
• QTL cloning

10. Construction of Genetic Map
A high-density genetic map of SNP markers in apple produced
using the 20K Infinum Array

11. Family (linkage) QTL mapping
PROs
• Highly controlled experiments
• High statistical power to detect ‘major QTL’
• A limited number of markers is requested and limited samples to be
phenotyped
CONs
• Experimental crosses, time consuming, laborious, long generation time
• QTL intervals are quite long, 5-10 cM (1 cM might be equal to 300 kb up
to 3000 kb of DNA) and contain MANY genes
• Fine mapping to identify individual genes in QTL (necessity of making
large number of crosses to elicit sufficient number of meiotic events)
• QTL identification is based on bi-parental crosses and quite often QTL
are specific to the bi-parental population and not useful in a wide
genetic backgrounds

12. Population mapping vs Family mapping
• Sampling natural variation: existing populations/germplasm
collections
• Breeding accessions (high agronomic value, low diversity)
• Landraces and local varieties
• Wild varieties (low agronomic value, high diversity, source of new alleles)
• Rely on the detection of Linkage Disequilibrium
• Take advantage of recombination events accumulated over
many generations
• Searching for genotype-phenotype correlations in unrelated
individuals
closed system open system

13. Linkage Disequilibrium
Definition:
“non-random association of
alleles at different loci on the
same chromosome”
Depends on:
– recombination fraction
– number of generation
Flint-Garcia et al. 2003

14. How does it look at molecular level?
Marker/marker correlation
Disequilibrium matrix for
polymorphic sites
Selfing species
selfing reduces opportunities
for recombination . . . DNA
variation tends to be structured
into distinct haplotypes coupled
with extensive LD
Outcrossing species
recombination among
heterozygotes breaks up
haplotypes and reduces LD

15. Why linkage disequilibrium matters?
– LD refers to the correlation between polymorphisms in a population
– The power of association study depends on the strength of this
correlation
– The resolution with which a QTL can be mapped is a function of
how quickly LD decays over distance
High LD
Whole genome scan, low number
of marker and resolution.
Typical of selfing plants
Low LD
Candidate gene approach, high
number of marker and resolution.
Typical of outcrossing plants

16. Population structure
Presence of a systematic difference in allele frequencies
between subpopulations in a population.
Causes:
Geographic origin: natural selection may favor genes for
adaptation (fixation of alleles flanking a favored
variant)
Nonrandom mating: geographic isolation, assortative mating,
breeding history
Coancestry
Domestication: Selection of desired agronomic and end-use
quality characteristics, co-selection of loci during
breeding for multiple traits

17. Population Structure estimation

18. Population mapping
PROs
• No need for the development of biparental populations
• High resolution mapping, take advantage of LD as well as historical
recombinations present within the gene pools of an organism and
natural genetic diversity
• Availability of broader genetic variations with wider background for
marker-trait correlation (i.e., many alleles evaluated simultaneously)
and also more QTLs underlying the traits
CONs
• GWA blind approach, facilitates the identification of new regions
containing no a priori candidate genes
• Less statistical power (presence of alleles wih low MAF, rare alleles)
• Spurious signals of association derived from population genetic
structure or non-random mating (relatedness)
• Population admixture, selection and genetic drift can bias the detected
association
• Large sample size – significant phenotyping

19. Genome Wide Association Studies (GWAS)
Different statistical methods are available for AM studies

20. GWAS Example
• 4271 SNPs, (561 Peach, 205 Nectarine)
• Trait: fruit pubescence
• Association made with GAPIT (Lipka et al. 2012) using the EMMAx
(Efficient Mixed-Model Association eXpedited) algorithm
●
●
●
●
●
●
●●●
●●●●
●
●
●
●● ●●● ●●● ● ●●●● ●● ●● ●●●●● ●● ● ● ●● ●● ●●●●● ● ●●● ● ●●●● ●● ● ● ●●● ●● ● ● ●● ●● ●● ● ●●● ● ●●● ●● ●● ● ●●● ● ●● ●● ● ●●● ● ●●● ● ●●● ●● ●●●● ●● ● ●● ●●● ● ●● ●● ● ●●● ● ●●● ●●●● ●● ● ● ●● ●● ●●● ●● ●●● ● ● ●●● ● ●●●●●● ● ● ●● ●● ●● ●● ●●●●● ●● ●●●● ●●●● ● ●●● ●●● ●● ●● ●●● ●● ●● ●●●● ●●● ● ●●● ●●● ●●● ●●● ●●●● ● ● ●●●● ● ●● ●●● ● ● ●● ●● ●● ● ●● ● ●●● ● ●● ●● ●● ●●●● ● ●●●● ● ●● ●● ● ● ●●● ● ●● ●●● ●● ●● ● ●●● ●●● ●● ●● ● ●●● ●● ●● ● ●●● ●●● ● ● ●● ●● ●●●● ● ● ●●● ●●●● ●● ●● ●● ●●● ●●●● ●● ●● ●●● ● ● ● ●● ● ● ●●● ●●● ● ● ●●● ●● ●●●● ●● ●●● ●● ● ●●● ●● ● ●●● ●●●●● ● ●● ●● ● ●● ●● ●● ●● ●● ●● ● ●●● ●●● ●● ●●● ●● ● ●● ●● ● ●● ● ●● ●●●● ●●● ●●●●● ●● ●●● ●● ● ●● ●●● ●● ●●● ●● ●● ●●●● ●●● ●● ● ●●● ● ● ●● ●●● ● ●●●● ●● ●●● ● ●● ●●● ●● ● ●● ●●● ●● ●●●● ● ●● ●●●● ●● ●●● ● ●● ● ●● ●● ●● ●● ●● ●● ●●● ●● ● ●● ●● ●● ●● ● ●●●● ●● ●●●●● ●●● ●● ●●● ● ●● ● ●● ●● ●● ●● ●● ● ●● ●● ● ●●● ● ●●●● ●●● ● ●●● ●● ●● ●● ●●● ●●● ●● ●●●●●● ●● ●●● ● ●●●●● ● ●● ●●● ●● ● ●● ● ●●● ●● ● ●● ●● ●● ● ● ●● ●● ●●● ● ●●● ●●● ● ●●●● ● ●●●● ●● ● ●● ●●● ●●● ●●● ● ●● ● ●● ●● ●●● ●●● ●●●● ● ●●●● ●● ●● ●● ●●● ●● ● ●● ●● ● ●● ●● ●● ●●● ● ●● ●●●●● ● ●● ●● ●●● ●● ●●● ●● ●●●●●●●● ●● ●●● ● ●● ●●● ● ● ●●● ●●● ●● ●●●● ●● ●● ●● ●● ●● ●●● ●● ●●●● ● ●●● ● ● ●● ● ●● ●● ●● ● ●●● ●● ●● ●● ●●● ●●● ● ● ●● ● ● ●● ●●● ●●●● ●●● ● ●●● ●● ● ●● ● ●●●●● ●● ● ●●● ● ●●●●● ●● ●●●● ● ● ● ● ●● ● ●● ●● ●● ● ●● ●● ● ●● ●● ●●●● ● ● ●●●● ● ● ●●●● ●● ●● ●● ●● ● ●●● ●● ● ●●● ● ●● ● ●● ●●●● ● ●●● ●● ● ● ●● ● ● ●●● ●● ●●● ● ● ●●● ●● ● ●● ●●●●●●● ●● ●●● ●●●● ●● ● ●●● ●●● ●●● ●●● ●● ● ●● ●●● ●●● ●● ● ●●● ●● ●● ●● ●● ● ●● ●●● ●●●● ●● ●● ●● ●●●● ●● ●●●● ● ●●● ●● ●● ●● ●● ●●● ●● ●●●●●●●● ●● ●● ● ●● ●● ●●●● ● ● ●●● ●●●● ● ●●● ●●● ● ●● ●● ●● ●●● ● ●● ● ●●● ● ●●●● ● ●●●●● ●●●● ●● ●● ●● ●●● ●●● ●●● ●● ●● ●● ●●● ●●●●● ● ●●●● ● ●●● ●● ●● ●●● ● ●●●●●● ●● ●●● ● ● ●● ●● ●● ●●●● ●●● ●●● ●● ● ●●● ●●●● ●● ● ●● ●●●● ● ● ●● ● ●●● ●● ●● ●● ●● ●● ●●● ●● ● ●●● ● ● ●●● ●● ●● ●● ●●● ●● ●●● ●● ● ●● ●●● ● ●●●●●● ●●●● ●● ●● ●●● ● ● ● ●●●● ●●● ●● ● ●●● ● ●● ● ●●●● ● ●●●● ●● ●●●●● ●● ●● ● ●● ●● ●●● ●●●● ● ●●●●● ● ●● ●●● ● ●●●● ● ●●●● ●●● ● ●●● ● ●● ● ● ●● ●● ●● ●●● ● ● ● ●●● ●● ● ●●●● ●● ●●● ●● ● ●●● ● ●●● ● ●●● ● ● ●● ●● ●● ●● ● ●●● ●● ●●● ●●● ● ● ●●● ●● ● ●●● ●● ● ●●● ●●●●●●● ●●● ●● ● ● ●● ● ●●● ●● ●● ● ●● ● ●● ●● ●●● ● ●● ● ●●●●●●●●●● ● ●●● ●●● ● ● ●● ●● ● ● ●● ● ●● ●● ●● ●●●● ● ●● ● ●● ●●●● ● ● ●●● ●●● ●●●●● ●● ●● ● ● ●● ●● ●●● ● ● ●●●● ●● ●●●● ●● ● ● ●● ●● ●● ● ●●● ● ●● ●● ● ●●●● ●● ● ●●●●● ●●● ● ●● ●● ●●● ● ●●● ●● ●●● ●●● ● ●● ● ●●●●●●●● ●●● ● ● ●● ●●● ●● ● ●● ●●● ●● ●● ●● ●● ●●● ● ● ●●●● ●● ●●●● ●● ●●● ●● ● ●●● ●● ● ● ●●●●● ●● ● ●● ●● ● ● ●●●● ●● ●● ● ●● ● ●● ●● ●● ● ● ●●●● ●● ●●● ●●● ●●● ●●● ●● ● ●●● ●● ●●●● ●● ●● ●● ● ● ● ●●● ●●● ●● ●●●●● ●● ●● ● ● ●●● ● ● ●● ● ●●● ● ●● ● ●●● ● ●●● ●● ● ●●● ● ● ●●● ●●●● ●●●●● ● ●● ●● ●● ● ●●● ●● ●● ●●● ●● ●● ●●● ● ●● ● ● ● ●● ●● ● ●●●● ●● ●● ●● ●●● ●● ●●● ● ●●● ●● ●● ●●● ●● ●●● ● ●● ● ●●●●● ●● ● ●● ● ●●● ● ●● ● ● ●● ●● ●● ● ●●●● ● ●●●●● ●● ● ● ●● ●●●● ●●●●● ● ●● ●● ●● ●● ●● ●●● ● ● ●● ● ● ●● ●●● ●● ●● ●●●●● ● ●●● ●● ● ● ●● ●●●● ● ●●● ● ●● ●● ● ●●●● ●● ●●● ●● ●●● ●●● ●● ●● ●● ●● ● ●●● ●● ● ●●●● ● ●●● ●●● ● ●● ●●● ●●● ● ●● ●●● ● ●● ● ●●● ●●● ● ●●● ● ● ●●●● ● ●●●● ●● ● ●● ●●● ●● ● ●●●●●● ● ● ● ● ●●● ●●● ●●● ●●● ● ●● ●●●● ●●●● ●●● ●●● ● ●● ● ●● ●● ●●● ●●● ●● ● ● ●●● ●●●● ● ●● ●● ● ● ●●● ●● ●● ●●● ● ● ●●● ● ●●● ● ●● ●● ● ●● ● ● ●●● ● ●● ●● ● ●● ●●● ●●● ●●● ●● ● ●● ●●●●● ●● ● ●● ●●● ●● ●●● ●● ● ●● ● ●●● ● ● ●●● ●● ●●● ● ●●● ●● ● ●●● ●● ●●● ●● ●● ●● ● ●●● ●● ●● ● ●●●●● ●● ●●● ●● ●● ●●●● ●●● ●●●●●● ●● ●● ●● ● ●● ●●●● ●● ●● ●● ●● ●● ●● ●●● ●● ●●● ●● ●● ● ●●● ●●●●● ●●●●● ●● ● ●● ● ●● ● ●● ●●● ●● ● ●●●● ● ●●●●● ● ● ●● ●● ●● ●●● ● ● ● ●●● ●● ●●●●● ●●● ●● ● ●● ● ● ●●● ● ●●●● ●● ● ●●● ●● ●● ● ●● ● ●● ● ●●● ●●● ●● ●●●● ●● ● ●●● ● ● ●●●●●● ● ●● ● ●● ● ●●● ●● ●● ● ●● ● ●●●● ● ●●● ●●● ● ● ●●● ●●●● ●● ● ●● ● ●● ● ●● ●●● ●● ●● ●● ●● ●●● ●●● ● ● ● ● ●● ●●● ● ●●● ●●●●● ●●● ●●● ● ● ●●● ●●● ●●●●● ●● ● ● ●●●● ● ● ●● ● ●●●● ●●●●● ● ●●● ●●● ● ●●● ●●●● ●● ●● ● ●●●●● ●●●● ● ● ●● ● ●●●● ●● ●● ●●● ●● ●● ● ●●● ●●●● ● ● ●●● ● ●●●● ●●● ●● ●● ●●● ●● ●●● ●●● ●●● ●●● ●● ● ● ●● ●●● ●●● ● ●● ●●● ●● ● ● ●● ● ● ●● ●● ●● ●● ● ●● ●●●● ●●● ● ● ●●● ●●● ●● ● ●●●● ●● ●●● ●● ●●● ●●●● ●● ●● ● ● ●● ●●● ● ●●● ● ● ●●● ● ●● ●● ● ● ●● ●● ●● ●● ●● ●● ●● ●● ● ●●● ● ●●● ●● ●● ● ●●● ●● ●● ● ●●● ● ● ●●● ●●● ●● ●● ●● ●●● ●● ● ●● ● ● ●● ●●● ●●● ● ●●● ● ● ●● ●● ●● ●● ●●● ●● ● ●● ●● ●● ●●●● ●●●● ●●● ●● ●● ● ●● ●●● ●● ● ●● ● ●●● ●● ●●● ●● ●● ● ●● ●● ●● ● ● ●● ● ●● ● ●● ●● ●● ●●● ● ● ● ●●● ● ●●● ● ●●●● ● ● ●● ● ●● ●● ●● ● ●●● ●● ● ●● ●●●● ● ●●●● ●● ●● ●● ● ●●●● ● ●● ● ●●●●● ●●●● ● ● ●● ●●● ●● ● ●● ●●● ● ●● ● ●● ●● ●● ● ●● ●● ● ● ●●●● ●●●●● ●● ●●● ●● ● ●●● ●● ● ●● ●●●●● ● ●●● ● ●●●●●●● ●●●● ●● ●●● ● ● ●● ●●● ● ● ●● ●● ●●●● ● ●●● ● ●●●● ● ●● ●● ●● ●● ● ●● ● ● ●●●● ●● ●● ●● ●●● ● ● ●●●● ●●● ●● ●●● ●●●● ●●●● ●● ●●● ●● ●● ● ●● ● ●●● ● ●● ●● ●●●● ●● ●● ●● ●●● ● ●● ●●●● ●●● ●● ●●● ● ●●●● ●●●● ● ●●● ● ●● ●● ●●● ●● ●● ●● ●●● ●● ●●● ● ● ● ●●●●●●● ● ●● ● ●●● ● ●● ● ●● ●●●● ● ●●●● ●●● ●●● ●● ●●● ●● ● ●● ●● ●● ● ●●● ●● ●●● ●●●●● ●● ●●● ●● ● ● ●●● ●●● ●● ● ●● ●● ●●●● ●● ●● ● ● ●● ●● ●● ● ●●●●●● ● ●●● ● ●●●● ●●●● ●● ● ●● ●● ●● ● ●● ● ● ● ● ● ●●● ●● ●● ● ●●●●● ●● ● ●●●●● ●●● ●● ●● ● ●●●●● ● ●●●● ●● ● ●●● ●● ●● ●●● ●● ●● ●● ● ●●●● ●● ●●●● ●● ●● ●●●●● ●● ●● ●● ● ●●● ● ●●● ●● ● ●● ●●● ●● ●● ●● ● ●● ●● ● ●●●● ● ●●● ●●● ●● ●● ● ● ●● ●●● ●● ●● ● ● ●● ●● ●● ● ● ● ●●● ●●● ●● ● ● ●● ● ●●● ● ●● ●●● ●● ●● ●● ●● ●●● ●●●● ●●● ●● ●●● ● ●●● ●● ●●●●● ● ●●●●● ●● ●●●●● ●●● ●●● ●●●●● ●● ●● ●● ●● ●●● ●● ●● ● ●● ●● ●● ●●●●●● ●●●● ● ●● ● ●●●● ● ●● ●●●● ●●● ● ●●● ● ●●● ●●● ●●● ● ● ● ●●●●●●● ●●● ●● ● ● ●● ●●●● ● ●●● ●● ●●● ●●● ● ●● ●● ●●● ●● ●●● ● ● ●● ● ●● ● ● ●●● ●●●● ● ●● ●● ●●● ● ●● ●● ●●●● ●●● ●●●● ● ●● ●●● ● ●● ● ●● ● ●● ● ●● ● ●●● ●●● ●● ●●●● ●● ●●● ●● ●●●● ●● ●●● ● ● ●●● ● ●● ●●●●● ●●●● ●●● ●● ●● ●● ●● ● ● ●●● ●●● ●●● ●● ● ● ●● ● ● ●● ● ●●● ●● ●●● ● ●●●● ● ●●● ●● ●●●● ●●●●● ●● ●● ●●● ●● ●●●● ● ●● ● ● ●● ●● ●●●● ●●● ● ● ●● ● ● ● ●● ● ●● ●● ● ●● ● ●● ● ●● ●● ●● ● ●● ● ●● ● ●●● ●●● ● ●● ●● ●● ●● ● ●● ●●● ● ● ●●● ●●● ●●● ● ●●● ●●● ●● ●●● ●● ●● ●●● ●● ●● ●● ●●●●● ● ●●● ●●● ● ●● ● ●●● ●● ●● ● ● ●● ●●●●● ●● ●●● ●● ●● ● ● ●● ●●●● ● ●●● ●●● ●●● ●● ●● ●●● ● ●● ●●● ● ● ●●● ●● ●●● ● ● ●● ●●● ●● ●●● ●● ● ● ●● ●● ●●●● ● ●● ●●● ●●●● ● ●●● ●●●●● ●●
.fruit_pubescence
−log10(p)
1 2 3 4 5 6 7 8
135791113
●
●
●
●
●
●
●●●
●●●●
●
●
●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●
0 1 2 3
02468101214
.fruit_pubescence
Expected −log10(p)
Observed−log10(p)

21. Test of associated SNP as candidate for MAS
Case 1: Perfect association

22. Test of associated SNP as candidate for MAS
Case 1: Perfect association
Case 2: Obfuscated association

23. Genomic Selection
Genomic selection (GS) was first suggested in animal breeding
(Meuwissen, 2001)
GS is a form of MAS that simultaneously consider the effect of
all markers (as many loci as possible) in the whole genome to
calculate the Genomic Estimate of Breeding Value (GEBV).
GS avoids QTL mapping altogether
• In GS, the joint effects of all markers are fitted as random
effects in a linear model
• Trait values are predicted from a weighed index calculated
for each marker

24. Genomic Selection

25. Genomic Selection
• Simulation studies have shown that across different
numbers of QTLs (from 20 up to 100) and levels of h2,
responses to GS were 18 to 43% larger than MAS
(Bernardo and Yu, 2007)
• GS was found more useful with complex traits and low h2
• GS focuses on the genetic improvement of quantitative
traits rather than on understanding their genetic basis
• GS and QTL discovery are not mutually exclusive.

26. Genomic Selection

27. Genomic Selection Models
Prediction problem due to large p small n
• Best Linear unbiased prediction (BLUP)
• Ridge regression
• Bayesian regression
• Kernel Regression
• Machine learning model
• Partial least square regression (PLS)
• Principal component regression (PCR)

28. Thank You for your
attention