The fourth largest seed company in the world, Limagrain is an international cooperative group created and directed by French farmers. As a creator and producer of plant varieties, Limagrain markets field seeds, vegetable seeds and cereal products. HM.CLAUSE, Business Unit of Group Limagrain, is an innovative global leader in the development, production, and commercialization of vegetable seeds. HM.CLAUSE belongs to the top 5 vegetable seed entities in the world. With over 2,000 varieties in more than 20 vegetable crops, HM.CLAUSE provides innovative solutions to growers worldwide.  

Référence : 2019-4885

Context and objectives

Breeding programs generate large amount of phenotypic data (field trials) to evaluate plant material and select best performing genotypes for subsequent generations of selection. This systematic field screening along breeding cycles leads to the accumulation of phenotypic information over years and locations which might not be used anymore since selection decisions have been taken.

Combining these phenotypic data with genotypic information would allow the implementation of recent molecular based approaches such as genomic selection (GS) to leverage the use of these historical breeding data for predicting breeding value of new selection candidates.

The efficiency of GS relatively to phenotypic selection might differ according to breeding stages. We propose to assess the efficiency of GS at two different breeding stage: i) in early stage of line selection within biparental segregating families (breeding starts) as well as ii) in early stage of hybrid selection during first crossing block. The objective will be to propose a sampling strategy within the large training dataset to optimize prediction accuracies at these two different breeding stages. Cross validations with different sampling strategies will be carried out to identify the best training set within historical data to perform GS at different breeding stages.

The present internship proposition will focus on one vegetable hybrid breeding program of HM.Clause. Historical breeding data from this program consists in more than 10000 hybrid combinations observed in the trialing network across several years and locations.

Required skills

• Quantitative genetics
• R coding
• Good English written skills 
• Statistics
• Team working 
• Plant breeding
• Enthusiasm

Click HERE to learn more and to apply online 
https://talent.limagrain.com/Pages/Offre/detailoffre.aspx?idOffre=4885&idOrigine=3448&LCID=1036&offerReference=2019-4885