|
Il giorno 22 marzo 2006, alle h
14.30, si terra' presso il CNR-IMATI, sezione di Milano, Via Bassini, 15, Milano, (piano terra), il seminario BAYESIAN METHODS FOR GENOMICS WITH VARIABLE SELECTION MARINA VANNUCCI Texas A&M University Abstract. Stochastic search techniques for variable selection, originally developed for regression settings, have been applied to a variety of different models, including recent extensions to classification problems. I will briefly review the key ideas of these methods, then propose new developments for the case of unsupervised clustering. Unlike classification, where the group structure is pre-specified in a training set, in clustering the analysis is fully data-based. This calls for different methods for modeling the data and implementing the Bayesian variable selection techniques. I will formulate the problem in terms of a multivariate normal mixture model with an unknown number of components and use the reversible jump MCMC to define a sampler that moves between different dimensional spaces. The work is motivated by applications in genomics, specifically the analysis of the high-dimensional data generated by DNA microarrays. The identification of distinct subtypes of a disease and the detection of discriminating genes are common goals in the analysis of such data. Molecular classes defined on a small number of genes can lead to a better understanding of the underlying biological processes. In addition, the selected genes can serve as biomarkers for improved diagnosis and targets for therapeutic intervention. Tutti gli interessati sono cordialmente invitati a partecipare. Ulteriori informazioni alla pagina web http://www.mi.imati.cnr.it -- Antonella Bodini CNR-IMATI Via Bassini 15, 20133 Milano (Italy) tel +39 02 23699524 fax +39 02 23699538 http://www.mi.imati.cnr.it/~anto/============================================================================= NOTA: Le norme per utilizzare il forum SIS e le istruzioni per iscrizione e cancellazione sono disponibili all'indirizzo . . http://w3.uniroma1.it/sis/forum.asp . L'archivio di tutti i messaggi (aggiornato al giorno precedente) e' disponibile all'indirizzo . . http://www.stat.unipg.it/cgi-bin/wilma/sis/ . ============================================================================= |