• STATUS and ADDRESS

Researcher (DR), Dept. Forest, Grassland and Freshwater Ecology (EFPA).
Also associated to the unit Ecology of Mediterranean Forests.

INRA - Unité de Biostatistiques et Processus Spatiaux (UR546)
Domaine St-Paul - Site Agroparc
84914 Avignon Cedex

Tél.: + 33 4 32 72 21 54
Fax: +33 4 32 72 21 82

 

• POSITIONS AND EDUCATION

INRA Researcher since 2003
Assistant professor in statistics, department OMIP, Institut National Agronomique Paris-Grignon (1999-2003)
PhD Student, department Ecology, Systematics & Evolution, University Paris-Sud Orsay. 1996-1999

 

• RESEARCH INTERESTS

Within population biology, my research interests concern dispersal of plant propagules (pollen and seeds). Estimating dispersal kernels in various experimental designs, understanding the impacts of long distance dispersal and forecasting gene flow in mosaic landscapes are my main activities. During my PhD, I worked on transgene escape in agro-ecosystems. I am still engaged on this topic through collaborations, but I have mostly worked on gene flow in forest species since my arrival in dept EFPA. I am currently focusing on the estimation of dispersal kernels and components of male/female fertilities from multi-locus genetic data (typically microsatellites markers).

• Mixed Effects Mating Model
  

Programs to analyse pollen dispersal from microsatelites data following the approach developed in Klein et al. 2008 are now available here.

• Colon SGS PROJECT

A project named "ColonSGS: Set-up of a spatial genetic structure during a colonisation process: mathematical models and two case studies concerning wood species (cedar and beech)", granted by the ANR started in 2008. It will focus on the importance of long-distance dispersal event for spatial genetic structure.

• DETAILS

 

• Long distance dispersal

Fat-tailed dispersal kernels were known to behave qualitatively differently than thin-tailed dispersal kernels concerning the speed of advance of a colonization wave (Mollison 1977, Clark et al 2001). We proved that the exponential tails are also a critical point concerning the mixture of propagules from different sources : fat-tailed kernels lead to even propagule clouds when far from the sources, whereas propagule clouds become dominated by the closest source for thin-tailed kernels. The consequences of this property for the spatial pattern of genetic diversity during colonization will be fully investigated during the PhD of Julien.

Julien Fayard started his PhD in october 2006 supervised by François Lefèvre and myself, with a grant from the French Research Ministry. He works on Variability of the dispersal kernel along an expanding population. Impacts of the heterogeneity in the environment.

Characterizing the tail of dispersal kernels, particularly with respect to the property mentioned above, was also a main objective of numerous statistical studies in the sections below.

FAYARD J, KLEIN EK, LEFEVRE F, submitted. Long-range dispersal and the fate of a mutation arising in the colonization front. JEB.
KLEIN EK, LAVIGNE C, GOUYON PH, 2006 Mixing of propagules from discrete sources at long distance: comparing a dispersal tail to an exponential. BMC Ecology. 6:3

 

• Mating systems in forest trees species

We extended the likelihood approaches for neutral markers previously used for estimation of selection gradients (Burczyk et al 2002) to estimate explicit dispersal kernels and investigate the shape of their tail. The approach was originally designed on Sorbus torminalis, and for this species the approach provided results quite consistent with those from the now classical TwoGener approach (Smouse et al 2004). With Pierre Gerard, we extended the method to account for genotyping errors typical of microsatellites (± 2bp) for a study on hybridization between two Fraxinus species. Finer characterizations of the composition of pollen clouds and how they depend on the spatial pattern of individuals around a mother tree are also of interest to me. This could lead to alternative spatial mating models no more based on a dispersal kernel. These topics will be central in the PhD of Florence and Gwendal.

On theis topic, I also collaborated with Catherine Bastien and her team in Orléans in the framework of INTERPOPGER. This was a project on Populus and Melampsora, granted by the national program Ecoger, that will finish at the end of 2008.

Gwendal Restoux defended his PhD in June 2009 on Spatio-temporal variability of mating system in Abies Alba. Bruno Fady (URFM) supervised with me this PhD granted by INRA EFPA and PACA region.

Florence Carpentier defended her PhD in june 2010 on Measuring pollen or seed dispersal from genetical markers partially observed. This PhD was supervised by Joël Chadœuf and me and granted by French Research Minister.

 

• Pollen dispersal in crop species and GMO release — Oilseed rape

During my PhD, we estimated the dispersal kernel of oilseed rape from an experimental field with a GMHT patch in the middle. We proposed a method to analyse data from such experiments accounting for the spatial design used. Actually, it consisted in estimating a dispersal kernel using an inverse method such as that used for seed dispersal (Ribbens 1997, Tufto 1997, Clark 1998). This function was used in the model GeneSys (Colbach et al 2001) to forecast transgene escape in spatially explicit agronomical landscapes. During her PhD Celine estimated a dispersal kernel directly at landscape scale using microsatellites data and a statistical approach derived from those used on forest species. She found much fatter tailed kernels than those found at field scale.

My contirbution to this field is now limited to collaborations in the project GMBIOIMPACT. This project on oilseed rape is granted by the national program ANR OGM and managed by Jane Lecomte in Orsay. The team of Bernard Vaissière working at INRA Avignon on pollination by bees is also engaged in this project.

Céline Devaux defended her PhD on 19th of may 2006 entitled Modelling and estimating the long-distance and efficient dispersal of oilseed rape pollen, at the landscape level and within a field, supervised by Claire Lavigne and me, granted by INRA.

• Pollen dispersal in crop species and GMO release — Maize

Using an experimental field of yellow corn with a patch of blue corn in the middle we estimated dispersal kernels for corn. The statistical approach was similar to that used for oilseed rape but we used quasi-mechanistic models accounting for wind speed, main wind direction and difference of height between ears and panicles. The integration of this dispersal model with a model for flowering phenology was achieved within the MAPOD model that allows forecasting pollen flow within a spatially explicit corn landscape and an explicit meteorological scenario. This model is used to investigate coexistence rules for GM, non-GM and organic crops (Angevin)

• Other publications

 

• OTHER PAST AND PRESENT MAIN COLLABORATORS

Catherine Larédo from MIA Jouy-en-Josas supervised my PhD.
Claire Lavigne thought to extricate herself from the field of research on GM crops by transmitting me her PhD topic. She finally had to put up with the topic and me …
Frédéric Austerlitz supervised my M.Sc. thesis on coalescence and we collaborated on numerous projects during last years.
Sylvie Oddou-Muratorio converted me to microsatellites data and parentage analyses during her PhD on Sorbus torminalis.
Frédérique Angevin had the bravery to convert the maize dispersal model set up during my PhD into the MAPOD model…