Skip to Main content Skip to Navigation
Theses

L'évaluation expérimentale des innovations variétales. Proposition d'outils d'analyse de l'interaction génotype - milieu adaptés à la diversité des besoins et des contraintes des acteurs de la filière semences.

Abstract : Among the methods developed by technicians and researchers and aimed at helping experimental evaluation of new varieties, many have not been used in the agricultural sphere, despite their theoretical interest. Assuming that this lack of use results, at least partially, from lack of appropriateness between the tools and the real needs of the professionals, this work combines the analysis of the diversity of use of varietal experimentation and the conception of a tool to interpret results in variety trials. This study was carried out on winter soft wheat.
Firstly, interviews were carried out with about twenty different professionals involved in breeding, development, multiplication, variety distribution and cereal products transformation. The interviews highlighted great diversity in the use of varietal experimentation. Four objectives underlie these uses : (1) to sort the genotypes, (2) to position them, with a geographical or marketing point of view, (3) to obtain knowledge on the genotypes, (4) to transfer knowledge about them. Each objective is distinguished by a small number of criteria to evaluate genotypes. Based on these objectives and criteria, and on the experimental network characteristics (for example, trial and genotype number, partnership part), we propose a classification of the different uses in 10 categories. Differences within each category may be related to individual company strategies. The interviews also show a gap between both pursued and effectively obtained information from the experimentation. This gap can be explained by organisational constraints in the performing of trial networks and by difficulties encountered during data processing (for example, lack of available time to process the results, required competence to use the genotype x environment interaction analysis tools). The most expressed need concerned better extraction of the information contained in the experimental networks. For each of the 10 uses, we discuss the interest of 5 tool types, able to improve data collecting and processing.
Secondly, according to certain needs of professionals, a toll has been developed to improve extraction of information from multilocal trials and to optimise trial networks, by helping interpretation of genotype performance variations. This tool is based on the association of a crop diagnosis on probe genotypes, using a multiple linear regression, and an interaction analysis on all genotypes, using a factorial regression. This association aimed at validation of explicative variables of the interaction in an agronomic point of view. Depending on the probe genotype, the diagnosis explains from 70 to 99% of the yield variations for networks amounting 10 to 35 trials. Thus, each trial is characterised by the nature and the contribution of the limiting factors to the yield losses. In the interaction analysis, the genotypic parameters of the factorial regression make it possible to interpret the variations of genotype behaviour in terms of tolerance to the limiting factors. The tolerance scores appear well correlated to the scores deduced from the observations for visible limiting factors as diseases (we obtained a correlation score of 0.9 for brown rust). Thus, this method allows also evaluation of the genotype tolerance to limiting factors which effect is difficult to observe (for example, water stress, lack of nitrogen). Taking into account the needs expressed by the professionals, and the diversity of the varietal experimentation uses, we discuss the ways to simplify and improve the method in order to analyse the trial results, and we evoke adjustments of the actual evaluation practices which could result from the adoption of this tool.
Document type :
Theses
Complete list of metadatas

https://pastel.archives-ouvertes.fr/pastel-00003471
Contributor : Ecole Agroparistech <>
Submitted on : Thursday, February 28, 2008 - 8:00:00 AM
Last modification on : Monday, October 19, 2020 - 11:03:51 AM
Long-term archiving on: : Thursday, September 30, 2010 - 8:29:58 PM

Identifiers

  • HAL Id : pastel-00003471, version 1
  • PRODINRA : 244938

Collections

Citation

Christophe Lecomte. L'évaluation expérimentale des innovations variétales. Proposition d'outils d'analyse de l'interaction génotype - milieu adaptés à la diversité des besoins et des contraintes des acteurs de la filière semences.. Agronomie. INAPG (AgroParisTech), 2005. Français. ⟨NNT : 2005INAP0016⟩. ⟨pastel-00003471⟩

Share

Metrics

Record views

1373

Files downloads

11622