One of the major hurdles in predicting how biological diversity will respond to environmental factors in the 21st century is the current limits to assemble large and standard diversity data that are repeatable over time. Sequencing technologies are currently evolving at a fast pace and should offer unique opportunities to make a significant progress for biodiversity surveys in the near future. The METABAR project proposes to combine concepts from metagenomics (analysis of cellular microbial DNA from the soil) and from the recently emerged field of DNA barcoding (use of small DNA fragments that serve to discriminate among species). Based on recent results, we are now confident that virtually any soil contains enough extracellular DNA from decomposed tissues (even degraded and in small quantities) to be extracted, amplified, and sequenced using next generation sequencers. Second, an analysis of sequence repositories shows that, for most taxonomic groups, it is possible to find small DNA regions (ideally around 30 bp) that will efficiently discriminate across the taxa. Using this approach, we have been able to survey the recent and past diversity of plants in the arctic permafrost. We also demonstrated that such surveys are possible for temperate and tropical regions, not only for plants, but also for animal with high biomass. The METABAR project proposes to develop innovative protocols to standardize and automate the acquisition of high-throughput biodiversity data based on environmental (extracellular) DNA samples. These novel protocols will be tested in two large-scale experiments, including broadly different environmental conditions (alpine and tropical), and taxonomic groups (plants, ants, frogs, mammals, fungi). We will then conduct an analysis of these biodiversity patterns based on ecological models bridging across species-level and community-wide approaches. In particular we will test whether patterns of spatial species turnover previously evidenced for tropical trees may also be detected for other groups. We will also contrast levels of biodiversity between pristine (or unmanaged) and disturbed (or managed) ecosystems.

Tropical plant diversity is extraordinarily high both at local and regional scales, including a significant component of beta-diversity, or the turnover in species composition across habitats and regions. Yet we still know little about the factors underlying species distributions, with more than half of all tropical plant species having been collected only once. In particular, the relative roles of biogeography, abiotic factors, and biotic factors in limiting plant species distributions remain a subject of debate. Natural enemies (both fungi and insect herbivores) have recently been shown to exert strong forces on plant community composition, and it has been hypothesized that such biotic interactions are far more important in driving plant species turnover than environmental drivers. Alternatively, natural enemies may be important at small scales, (local diversity) but may not influence turnover at the beta scale. The NEBEDIV project represents a comprehensive evaluation of tropical forest beta-diversity across broad geographic and environmental gradients. We will integrate not only plot level analyses of more than 100 tree communities across Amazonia but also the first characterizations of soil fungi and insect herbivore communities at this scale. Critically, we will not only examine community-level correlations between these communities across spatio-environmental scales but also characterize host specialization of fungi and herbivorous insects, building on the extensive molecular phylogenetics work and collaborations we have established in previous international projects. We believe our teams are in a unique position to make this vital contribution given the extensive infrastructure we have established in recent years. The databases and modeling approaches we develop in NEBEDIV will contribute to an increased understanding of the factors that influence species turnover in the most diverse forests on earth. All three groups on which we focus are critically understudied in the tropics, and no attempt to date has been made to study these groups simultaneously for the same sites, especially with a well-replicated experimental design. The datasets we generate will therefore be essential for regional estimates of biodiversity, to assist policy makers to choose protected areas across the region, and to improve models of biodiversity dynamics in response to climate and land use change scenarios. The databases and modeling approaches we develop in NEBEDIV will contribute to an increased understanding of the factors that influence species turnover in the most diverse forests on earth. All three groups on which we focus are critically understudied in the tropics, and no attempt to date has been made to study these groups simultaneously for the same sites, especially with a well-replicated experimental design. The datasets we generate will therefore be essential for regional estimates of biodiversity, to assist policy makers to choose protected areas across the region, and to improve models of biodiversity dynamics in response to climate and land use change scenarios.

Improving feed efficiency in pig production has two main consequences: (i) increasing competitiveness of pig industry by reducing the cost of pork production – in a context of dramatic rise of the cost of feedstuffs – and (ii) decreasing environmental impact of pig farms, in particular that concerning phosphorus and nitrogen excretion. The measure of feed intake for individual pigs is, however, difficult and costly. Until now, the genetic improvement of feed efficiency has mainly been achieved through selection for higher average daily weight gain and lower lipid content of the weight gain. However, variations of the two latter traits explain only 50 % of the variation of feed intake in pig. Thus, it is tempting to improve “feed efficiency per se”, i.e. at constant growth rate and body composition. Residual feed intake (RFI) is a measure for assessing feed efficiency: it estimates the difference between actual feed intake and feed intake predicted from production and maintenance requirements. This trait is moderately heritable in livestock species, including pigs. It shows phenotypic and genetic correlations with major production traits. Relationships with reproduction traits were essentially explored in mice and poultry, showing low to moderate unfavorable correlations. Furthermore, lower RFI is suspected to reduce fitness and to increase susceptibility to stress factors. The main biological sources of genetic variation of RFI have been investigated in poultry and beef cattle, but mostly remain unknown in pigs. The PIG_FEED project is based on two experimental pig lines divergently selected for RFI and aims to provide 1) a better understanding of the genetic basis of feed efficiency and its genetic relationships with traits of interest in pig production, 2) a better understanding of the physiological consequences of selection for RFI in the growing pig. Through the PIG_FEED project, we will be able to evaluate the impact of selection on sustainable goals for pig breeding, such as reduced environmental impacts, reduced food inputs, reduced drug distribution, increased tolerance to hot environments. The lines result from a five-generation divergent selection experiment, from a common Large White population. Previous genetic studies on pig RFI have mainly targeted breeds, whereas responses to selection in “high” and “low” lines selected from the same base population should essentially reflect consequences of selection. Previous studies of RFI+ and RFI- lines have validated a significant divergence for the selection criterion. Correlative responses in meat quality, carcass composition and energy expenditure measurements were obtained. The PIG_FEED project is composed of four tasks. In the first task, we propose to analyze the divergent lines to measure correlative responses on production traits, reproduction traits, feeding patterns and activity related traits. These phenotypes will be analyzed under different feeding levels and consequences in terms of excretion levels and biological regulations will be inferred. In the second task, we propose to find associations between RFI and particular marker haplotypes, taking advantage of a pig SNP chip of 60K and the biological samples collected on founder animals and across all generations of selection. Moreover, a QTL detection design based on a backcross with females from the RFI experiment should permit to refine the haplotypes first detected. Strategies to take benefit of genomic associations with RFI in commercial populations should be proposed. In the third task, consequences of selection for RFI will be examined, using three series of transcriptomic and proteomic analysis, for the following aspects: muscle metabolism and adipose tissue metabolism during growth, muscle metabolism at slaughter and consequences on meat quality, and immune response. These data will be explored to identify physiological pathways affected by selection on RFI. In the fourth task , responses to stress challenges will be explored in the two RFI lines, with an immune challenge, a digestive and a nutrient challenges and a challenge to heat exposition. When required, the direction of nutrient use from growth to other physiological functions might be less efficient in RFI– than in RFI+ pigs. The results obtained in the PIG_FEED project should give us novel ideas for proposing breeding strategies for various selection objectives.