Workshop 1

5, 6 July, 10:00 – 12:00, Seminar Room, 2nd Floor, Building VII

Alexandra Posekany, Research Unit of Computational Statistics, Vienna University of Technology

Bayesian inference in medical statistics and epidemiology: a brief introduction

The Workshop will introduce the principles of Bayesian modelling through examples from building simple posteriors and HPDIs in conjugate prior scenarios showing an example from epidemiology where classical statistics is inapplicable. Then, we turn to the concepts behind creating complex models through standard software for simulation of Markov Chain Monte Carlo algorithms or Hamiltonian Monte Carlo algorithms, JAGS and stan for R. Installing R, RStudio and the R packages "rstan" and "rjags" prior to attending the workshop is necessary.

Workshop 2

11, 12 July, 10:00 – 12:00, Room 1.16, 1st Floor, Building VII

Antonio Gómez Corral, Complutense University of Madrid

Markov Chains in Epidemiology

The aim of this four-hour course is to increase the student's understanding of the spread of infectious diseases using stochastic processes, and specifically Markov chain models. First, we focus on single-type branching processes and their application to the SIRS model, which is an epidemic model with susceptible, infectious, and recovered individuals in which there is only temporary immunity to reinfection. Second, we present basic elements on quasi-birth-death (QBD) processes, which can be thought of as multivariate versions of the well-known univariate birth-death process, and we show how to derive a general-purpose algorithmic solution for a variety of stochastic descriptors. In the context of SIR and SIS models, the algorithmic approach is exemplified with the study of the first extinction of the pathogen, the peak of infection, and a random version of the basic reproduction number.

In the setting of these continuous-time and discrete-state processes, the emphasis is on describing the mathematical results being used and showing how to apply them to relatively simple models rather than on detailed proofs and complex distributional assumptions. Details on recent advances in the application of single- and multi-type branching processes, and of QBD processes, to epidemic models are briefly provided, along with a selection of related references.

Workshop 3

11, 12 July, 14:00 – 16:00, 1st Floor, Building VII

Sonia Torazona, Applied Statistics and Operations Research and Quality, Universitat Politècnica de València

Bioinformatics tools for multi-omics analysis

This short course will give an overview of the different analysis approaches and goals in multi-omics analysis. We will also discuss how a multi-omics experiment should be designed, and the MultiPower R package will be used to estimate statistical power in multi-omics studies. Finally, we will learn how to use the MORE R package to generate multi-omics regulatory networks for gene expression. The course includes practical sessions using R language and Rstudio.

REGISTRATION FOR THE WORKSHOPS IS FREE BUT MANDATORY