David CROS

Teaching (program, download)




Practical – requirements: computer, no prior knowledge of R

30+ slides powerpoint:     [download]

o   What is R?

o   Installing R and RStudio

o   The RStudio environment

o   The R language and commands: elementary operations, objects, functions, graphs, etc

o   Packages

~ 2 hours – 2 ½ hours depending on number of participants,

+ homework: make an ANOVA and multivariate plots




This is not an exhaustive course about plant breeding, but a focus on a common method (reciprocal recurrent selection) and on the state-of-the-art method of marker assisted selection for quantitative traits (genomic selection), with a preliminary presentation of the basics necessary to understand them.

lecture and practical work – requirements (for the practical work): computer, R software, knowledge of R

150+ slides powerpoint [download part 1+2, download part 3]:

o   PART 1 – Basic concepts of quantitative genetics (QG): the model of QG, gene effects, heterosis, genetic variances, coancestry (genealogical and molecular), mixed model for genetic analysis, heritability and selection accuracy, response to selection, linkage disequilibrium, correlations between traits, population structure


Practical: computation of coancestry in oil palm (excel / R)

Practical: simple example to estimate genetic values with a mixed model using R (MCMCglmm package applied by participants)

~ 2 ½ hours

NB: see “PART2” and “R for genomic studies” for other related exercises


o   PART 2 – Breeding (1/2)

§  Mass selection

§  Reciprocal recurrent selection (RRS): principle, genetic model for GCA and SCA, accuracy of GCA and SCA, the oil palm example (why RRS in oil palm?, crossing and experimental designs, results)

Practical: real life oil palm example to estimate genetic values with a mixed model using R (MCMCglmm package applied by participants and demo of asreml package on my laptop)

~ 2 hours

+ homework: answer questions about breeding paper


o   PART 3 – Breeding (2/2)

§   Genomic selection (GS)

(GS is about predicting the genetic value of individuals from their genotype on a large number of molecular makers, usually using the BLUP or Bayesian approaches, for quantitative traits)

o   Principle

o   Accuracy

o   Models and statistical methods

o   Genotyping

o   Imputing missing data

o   Etc


3 ½ hours


+ homework: answer questions about GS papers, among 2:

- on a tropical species

- on methodological aspects


NB: see “R for genomic studies” for GS practicals



R FOR GENOMIC STUDIES (starting from a VCF file)

practical – requirements: computer, R software, knowledge of R, knowledge from lecture “Breeding for quantitative traits”

50+ slides powerpoint:             [download]

1.    The Variant Call Format

2.    Using R for genomic studies (with 20K SNP maize data as an example):

·         Importing & exploring a VCF file

·         Missing SNP data

·         Physical map

·         Linkage disequilibrium

·         Population structure

·         Genomic relationships

·         Genomic predictions: GBLUP, Bayesian Ridge Regression, Bayesian LASSO & RKHS

  ~ 3 - 4 hours (depending on R level and number of participants, including 15 min break)

+homework:   1/use SNPRelate package to compute Fst between maize populations,

                     2/ loop on number of SNP and see how it affects accuracy




Requirement: knowledge from lecture “Breeding for quantitative traits”

130+ slides powerpoint             [download]

1. Introduction

a.     Palm oil market

b.     The oil palm (Elaeis guineensis and E. oleifera)

2. Traits of interests

3. Reciprocal recurrent selection (RRS) for yield

a.     Why RRS in oil palm?

b.     Implementation of the RRS scheme

c.      Seed production

d.     Genomic RRS

4. Interspecific breeding

5. Breeding for resistance to diseases

6. Marker assisted selection for simple traits

Practical: Mendelian inheritance of the low lipase trait (15-20 min)

~ 4 hours

+ homework: answer questions about 1 oil palm paper, among several (genome sequence [Singh et al 2013], virescens gene [Singh et al 2014], mixed model [Soh 1994], …)



  See this page for specific teaching sessions.


Il semble que vous naviguiez depuis un appareil de type 'bureau'

Tester plus en détail