Lignocellulosic biomass from short-rotation coppice poplar is a renewable resource of interest for producing second generation biofuels. However, current poplar varieties have neither been selected for this specific cultivation system nor for the conversion process of lignocellulose into simple and fermentable sugars (saccharification). The factors affecting biomass yield and chemical properties need thus to be studied. Over the last decades, some efforts in the main riparian poplar species involved in poplar cultivation in Europe have been devoted to the evaluation of the genetic variability of target traits for a bioenergy end-use. Furthermore, several genomic regions underlying this variation (quantitative trait loci – QTL) have been successfully mapped on the reference genome sequence. However, as they have been carried out in bi-parental crosses, the confidence intervals of the QTLs detected in such studies were quite large and their transferability in several backgrounds was rare, limiting severely their use in marker-assisted breeding. Efficient genetic improvement needs population based screenings, for which QTL mapping is rapidly limiting. To overcome such limitation, association mapping – the detection of QTLs in more complex populations – represents an obvious alternative in forest tree species because they are generally characterized by a rapid decay of linkage disequilibrium (LD). As a result, two recent association studies have been carried in two poplar species for biomass quality traits. Despite their novelty, as they were the first association mapping studies in riparian poplar species, both studies were characterized by a relative small number of significant associations and, more importantly, most of the reported associations explained a low proportion of phenotypic variability. Several hypotheses might explain these disappointing results, including the lack of exhaustiveness both in terms of genetic diversity and polymorphism screened, and the complexity of underlying determinisms. This latter point could involve gene by environment and/or gene by gene interactions, as well as rare variants, all of which could not be detected in previous studies as they were not explicitly addressed. By tackling such drawbacks, the proposed project aims at deciphering the genetic architecture of biomass yield and quality as target traits for poplar lignocellulose valorization in biorefinery. More specifically a systems biology approach, integrating polymorphism, expression and phenotypic data, is proposed in natural populations of black poplar (Populus nigra) covering the native range of the species in Western Europe and of specific interest in the French national poplar breeding program. Besides fundamental knowledge on the genetic factors controlling plant cell wall construction and biomass yield, the proposed project will evaluate the feasibility of phenotype prediction using multiple sources of information which can be of direct interest for breeding.