Early stirrings of landscape genomics: Awaiting next-next generation sequencing before take-off

Development of low impact sustainable agriculture and a growing use of adapted breeds is of priority to most countries in the world, and is in particular of key importance to developing countries/emerging regions. The level of adaptation of livestock breeds to their environment has to be measured in order to reach better understanding of the relationship between environment and the adaptive fitness of livestock populations, as well as to favour production systems based on adapted breeds. According to ILRI, landscape genomics is a long-term promising approach when regarding study of adaptation. A wide definition considers this field being an 'emerging area at the interface of natural resources management and the genome sciences' (Dr Williams, Nicholas School of the Environment and Earth Sciences, Duke University, USA). Population genomics, geoenvironmental and statistical methods can be combined to assess the level of association between specific genomic regions of living organisms and environmental factors. Processing of these many association models is based on data sets made up of hundreds of environmental parameters and thousands of molecular markers of thousands of animals. Applied to livestock genetics, landscape genomics purpose is to uncover what genetic material fits with which region of the world. A widespread use would support decision making and recommendations in order to favour raising of the right breed in the right place, thus leading to sustainable and adequate income for resource-poor farmers, reducing impacts on the environment, and making better goods available to consumers. Early stirrings of landscape genomics applied to livestock research were heretofore limited to a few studies carried out on sheep and goats. They were based on the use of SAM software (http://www.econogene.eu/software/sam/) as well as of population genomics theoretical approaches for results validation, and references therein). In comparison with the potential of landscape genomics, existing analyses were restricted to approximately one hundred environmental parameters related to a reduced number of genetic markers (<1'000) for a maximum number of 2'000 animals. Let us see how these figures are expected to evolve within a rather short lap of time. On the one hand, the number of available environmental databases (NASA SRTM, NASA MODIS, Norwich CRU, etc.) gradually increases and their quality improves (better resolution). Most of these global data sets are freely available throughout the Internet and can be used for a comparable description of production environments worldwide. No revolution will occur in the short-term with this category of information, but the diversity of the sources (new satellites, new environmental monitoring capacities) and the quality of data will go on improving progressively. On the other hand, the amount of molecular data to be analysed will more than likely very shortly and rapidly expand (P. Taberlet, personal communication). In 2004, after the 13 years and a 3 billion dollars human genome sequencing project, the American National Institutes of Health (NIH) proposed a new challenge: sequencing one human genome for $1'000. The objective has not yet been reached, but next generation of sequencers will allow researchers to get to genomic data faster and at a lower cost. In particular, theoretical potential of Single-molecule/Nanopore sequencing is undeniable. Based on this technology, with a 100 nanopores in parallel, a mammalian genome could be sequenced in 24 hours with the main cost being the chip itself, probably around $1'000 [9]. Several alternative low cost sequencing technologies are under way, and the $100 complete sequencing of individuals is about to become real (P.Taberlet, personal communication). It means that in the immediate future, any research project in livestock genetics will be given the opportunity to analyse the entire 3.4 billion base pairs per animal. We therefore have little time to pave the way. Naturally, researchers in molecular biology already started the conception of new approaches to analyse the large data sets produced by the 'next next' generation sequencing technologies. On the GIS and statistical side, it is necessary i) to stress the importance of geographic coordinates recording in any new project with animal sampling campaigns, and ii) to reinforce software and to make them more sophisticated. As the association analysis process is rather straightforward, the challenge will mainly consist in improving algorithm's efficiency and robustness. This is important because now the perspective clearly is the opportunity to analyse the complete genome of hundreds of thousands of sampled animals worldwide, and to compare it to hundreds of variables constituting progressively enhanced environmental data sets. In other words: models potentially constituted of billions of base pairs time hundreds of eco-climatic parameters corresponding to hundreds of thousands of individuals. In such a context, let us imagine the richness of the information collected by the GLOBALDIV project (www.globaldiv.eu) with the gathering of data from different European projects on livestock genetics (RESGEN, ECONOGENE, AVIANDIV, EFABIS). Forthcoming global analyses will really push us towards a new dimension of analysis, and therefore lead to new ways to assess genetic diversity. From then on, all conditions will be right for landscape genomics to take off, and to make it possible to fully profit from its contribution in order i) to better understand animal's adaptation to their local environment, and ii) to support adequate decision making in livestock biodiversity conservation, i.e. raise the right breed in the right environment. (author)