* Many people are not yet subscribed to the statslausanne mailing-list; please forward appropriately to your colleagues * Our next statistics forum for people interested in quantitative methods in biology and medicine will take place this Thursday 19 April at 11h00 in the reading room (2020) of the Genopode building (2nd floor, same level as the entrance from the Amphimax building). The speaker will be Jérôme Goudet from UNIL, who will talk about "Partial least squares regression (PLS) versus patterned statistics in a Spatial Approximate Bayesian Computation (ABC) context" Abstract: Approximate Bayesian Computation (ABC, http://en.wikipedia.org/wiki/Approximate_Bayesian_computation ) is an increasingly used method to estimate parameters when obtaining the expression for the full likelihood is difficult or impossible. The idea is simple: it consists in simulating data under a process similar to that that generated the observation, and to compare the simulated and observed data via summary statistics. The simulations giving summary statistics closest to the observed are retained, and provide estimation of the though parameters. A key question is how many and what type of summary statistics should be used? In this presentation, I will illustrate the use of ABC for the estimation of demographic and genetic parameters during the colonization of Europe after the last glaciation by a bird of prey, the barn owl. Interestingly we find that a blind approach using several hundred summary statistics post processed via Partial Least Squares regression (PLS http://en.wikipedia.org/wiki/Partial_least_squares_regression ), gave essentially the same results as an analysis where we carefully picked 4 summary statistics that we thought would best describe the information contained in our data. You can find more information about the forum on our web page at http://moodle.unil.ch/course/view.php?id=2729 Hope to see you there ! Frédéric