Tracking the genetic legacy of past human populations

Tracking the genetic legacy of past human populations through the grid UNIVERSITY OFENEVA G NICOLAS RAY & UNEP/GRID-EUROPE Swiss Grid Day, Bern, November 26 th 2009

Human migrations [12, 000] [55, 000] Homo sapiens Adapted from Cavalli-Sforza & Feldman, 2003

Why aiming at a good demographic model 1. Better understand human evolution • Origin of modern human (when, where, how many? ) • Relationship with other members of the Homo genus 2. Distinguish between the effect of demography and those of selection (biomedical applications)

A complex past demography fluctuation in effective pop. size substructure migrations Gene-specific factors mutations recombination selection Observed patterns of genetic diversity in contemporary populations

A complex demography [10, 000] [55, 000] demographic and spatial expansions population bottlenecks secondary contacts population isolation fast migration events Adapted from Cavalli-Sforza & Feldman, 2003

SPLATCHE SPatia. L And Temporal Coalescences in Heterogeneous Environment (http: //cmpg. unibe. ch/software/splatche)

From environment to demography Carrying capacity low high Spatial resolution: 100 km

From environment to demography Friction low high

Demographic simulations stepping-stone model (cellular automata) Pop. size Cell or deme time

Demography and spatial expansion Population density

Genetic simulations

Simulated genealogy ACCTAGTACAATCGGTAATGCCATTGGT TCCTTGTA…ATTGGT Summary statistics – Within population: • S, p – Between populations • Pairwise FST • Global FST – Globally • S, p Modèle de mutation Mutation ACCGAGTA…GTTGGT

APPROXIMATE BAYESIAN COMPUTATIONS (ABC) COMPUTATIONAL ISSUES Computer clusters Draw parameter values from priors Compute summary statistics 1 -10 mio. Generate genetic data Simulate one genealogy UBELIX (>500 nodes) Zooblythii (~40 nodes)

Computational issues A fully spatially-explicit model using 500 loci in 800 individuals: 10 CPU-years Adding long-distance dispersal: 20 CPU-years

SPLATCHE on the grid early 2005: joined the Biomed VO of the EGEE project mid 2005: tested on GILDA test bed, and deployed on the Grid since mid 2006: production mode and optimization

Use of SPLATCHE on the grid GRID N simulations Statistical tools Posterior distribution of demographic/genetic parameters of interest Independent simulations: - the more CPUs, the better - job failures are not that bad

Optimizations Submission time multi-threaded application using up to 30 RBs (used for the WISDOM project) Fetching time of job outputs in-house multi-threaded solution for checking status and getting outputs GRID 5 mio. simulations Reduction of the number of simulations (Daniel Wegmann) By MCMC. Promising results (~50 times less sims)

Geographic origin of human dispersal Ray et al. (2005) Genome Research

Interactions among populations Interaction between modern humans and Neanderthals in Europe Currat & Excoffier (2004), PLo. S Biol.

Cane toad invasion in Australia Estoup, A. , Baird, S. J. E. , Ray, N. , Currat, M. , Cornuet, J. -M. , Santos, F. , Beaumont, M. A. and L. Excoffier. Combining genetic, historical and geographic data to reconstruct the dynamics of the bioinvasion of cane toad Bufo marinus. Submitted

Take-home message A good human demographic model is important Realistic spatially-explicit approaches are essential The grid is key for sufficient exploration of parameter space User support and connections outside one’s discipline is crucial

THANK YOU !