Résumé
Genetic improvement of Hevea brasiliensis is a long process given both the immature period before latex production, the choice of parents for recombination and a three stage clone selection scheme covering 20 years or more. There are frequent contentions that genomics and biotechnologies will reduce drastically this duration. On the contrary, breeders can argue that phenotypic studies at field level, based on the observation of the global expression of the genotypes on rather mature trees (5 to 10 years of age), will always be necessary for clone recommendation. In rubber, improving the accuracy of evaluation is probably more efficient than reducing its duration for achieving a better genetic gain and ensuring regular release of new efficient, stable, and diversified clones. Integration of genomics and biotechnology tools can strongly contribute to very early screenings for a quick enrichment of the breeding populations with the most favourable alleles, or to improve cultivated clones by genetic modification. However, genetic recombination or genetic insertion into the genome respectively upon meiosis and transgenesis strongly modify genetic and epigenetic interactions. These changes impact both on plant metabolism and development. Gene: expression analyses cannot be successfully applied directly to breeding. They must be validated on young plants, under different stress treatments (notably wounding and ethylene: applications), and in controlled conditions (greenhouse). With this view, a range of gene expression analyses based on RT-PCR, macro-arrays, etc. was initiated on mature buded plants, on juvenile plants issued from in-vitro, as well as on transgenic plant material for characterizing early response to stress of contrasted clones (PB 260, PB 217, RRIM 600). Avaibility of young grafted plants and capability to apply on them a set of experimental strest treatments in controlled condition are necessary conditions to study in details molecular mechanisms. This better understanding of plant metabolism can lead to conceptual functioning model for identifying the key genes involved in the response of plants to the exploitation stress (tapping and stimulation) or other biotic and abiotic stresses. Describing s gene ontology might allow to select the key markers/genes responding in the same manner aft various stages of development. From the results of that research will be drawn a reduced set of assumptions that can then be submitted to validation on mature rubber trees. This approach is discussed in relation to rubber tree genetic improvement. (Texte intégral)