Présentation 1 (Alexandre Vaillant) - Recherche d'un déterminisme génétique dans l'expression du transcriptome d'hybrides d'eucalyptus au champ
Studies of phenotypic variations in plants are classically addressed to assess their adaptive potentials in changing enviroments and their ability to generate genetic gain through breeding program. Enabling easy snapshots of transcriptomes,RNA-Seq may bring expression profiling of genotypesas new and intermediate information in thisresearch field. However the existence of a genotype-based specificity of thegene expression levels, especially in actual cultivation conditions, is still to beclarified. We investigated leaf and xylem transcriptomes of three /E. urophylla x E. grandis/genotypes, contrasted in terms of growth in field, to look for differential expression linked to genotype and for heritability.Transcriptomes were very similar between clonal repeats. Out of the 25,665±634 genes expressed in leaf and the 23,637±1,241 expressed in xylem, 4.5 to 14% were differentially expressed (FDR<0.05) in leaf and 7.1 to 16% in xylem between clones. Several of these genes belonged to the secondary cell wallbiosynthesis pathways that play an important role in wood formation. Some significant heritability was found in leaf and xylemin 3431 (248) genes (at FDR<0.2) and 160 (3) (at FDR<0.05) respectively, showing that gene expression is influenced by genetic control. No relationship was detected between heritability estimates and gene expression levels or genomic regions, but Gene Ontology revealed that genes presenting significant heritability were significantly enriched in leaf in the 3 GO domains. Moreover the heritability of gene expression was significantly different between functions suggesting that heritability varied with gene function. These preliminary results are promising inunderstanding of the relationship between genes and phenotypes and, in the context of plant breeding, to evaluate whether gene expression can serve as possible ‘intermediate phenotypes’ that could improve the accuracy of selection of grossly observable traits.

présentation 2 - (Jean-Marc Bouvet) - covariation entre transcriptome et croissance de l'arbre chez eucalyptus - exploration des mécanismes fonctionnels reliant le gène au phénotype
Analysis of the transcriptome variation among Eucalyptus clones, for different tissues such as leaf and xylem, via RNA-seq technology, has shown that the genetic control could be predominant in gene expression  (Vaillant et al 2016 submitted). Numerous genes have been detected presenting significant heritability using the number of reads as quantitative variable .
The questions now are to see whether we can find some relationship between those peculiar genes and the complex traits related to tree growth and whether we can design some general schemes of gene functioning associated with trait response. For that we coupled multivariate analyses (sparse partial least square regression) to detect co-variation and  systems biology approach (protein protein interaction network) to define the patterns of gene functioning. The analysis used growth data from three replicated Eucalyptus urophylla*grandis clones and their transcriptome collected on two tissues leaf and xylem.
Multivariate analyses allow to detect numerous genes co-varying with height and circumference. Interactome analyses conducted with the selected genes, lead to the detection of the most relevant gene networks and generic functions and some specific genes (hub/bottleneck) that could play an critical role in complex trait expression.

Contact :

JM. Bouvet