Courses

The courses will take place on June 15th.

Sign up for the courses via ConfTool. Deadline is May 11th.

Overview

Temporal and spatio-temporal modelling and monitoring of infectious diseases (Full day course)

Michael Höhle, Department of Mathematics, Stockholm University, Sweden

Sebastian Meyer, Epidemiology, Biostatistics and Prevention Institute, University of Zurich, Switzerland

Analyzing latent times and competing risks in survival data (Full day course)

Laura Antolini, Paola Rebora, Maria Grazia Valsecchi, Department of Health Science, University of Milano-Bicocca, Italy

RNA-Seq: From analysis workflows to high-dimensional modeling (Half day course, afternoon)

Harald Binder, IMBEI, Mainz University, Germany

Details

1. Temporal and spatio-temporal modelling and monitoring of infectious diseases

Full day, 15 June 2015

Room U7-LAB732

Lecturers:

Michael Höhle, Department of Mathematics, Stockholm University, Sweden

Sebastian Meyer, Epidemiology, Biostatistics and Prevention Institute, University of Zurich, Switzerland

Course contents:

Infectious diseases remain a continuous threat to human and animal health. Understanding and controlling infectious disease spread is thus a key element in public health. The role of statistics is to combine stochastic models with observational data for applications in epidemiology and public health. Although our case studies originate from public health, the methods equally well find application in other contexts such as ecology and environmental sciences.

Introduction to infectious disease epidemiology
Transmission models and their parameter estimation
- Continuous time Susceptible-Infectious-Recovered (SIR) model
- Effective reproduction numbers
Latency periods and reporting delays
- Back-calculation method
- Nowcasting
Spatio-temporal modelling of registry data
- Multivariate time series models for count data
- Counting process models for individually-referenced data
Monitoring of univariate count data time series
- Aberration detection by the Farrington algorithm and beyond
- Methods from statistical process control
Outlook

The course content will be illuminated both from a theoretical and an applied perspective. In order to enhance the practical understanding of the methods, R code is given where possible – especially, the R package ‘surveillance’ will be used.

Requirements:

Knowledge of likelihood inference, generalized linear models and survival analysis as well as a basic understanding of stochastic processes.

2. Analysing latent times and competing risks in survival data

Full day, 15 June 2015

Room U7-LAB714

Lecturers:

Laura Antolini, Paola Rebora, Maria Grazia Valsecchi, Department of Health Sciences, University of Milano-Bicocca, Italy

Course contents:

Competing risks play an important role when analysing the clinical course of a disease when several events, fatal and non fatal, may originate the failure and are seen as competing causes of failure.

In this setting the cumulative incidence of failure, related to any event (or the corresponding event free survival curve) is not the only quantity of interest. The cumulative incidence function of each specific type of event, its relation to treatment and covariates and its contribution to the overall incidence are also on interest.

The analysis of the first signal of treatment failure is focused on the event occurring as first and is formalized by observable quantities such as the crude cumulative incidence and sub-distribution hazard.

Each event can be analyzed also by its occurrence even after previous signals treatment failure. The corresponding quantities of interest, such as the net incidence and net hazard, are generally unobservable since a fatal could indeed event prevent the observation of further events.

The course will deal with these two types of analysis starting from the clinical questions and corresponding quantities. This will deliver to the suitable data analysis and results interpretation.

Introductory aspects and notation
- Censoring and truncation
- Competing risks and latent times
Clinical questions and theoretical quantities
- crude cumulative incidence
- net incidence
- related hazard functions
Crude cumulative incidence Net incidence and related hazard functions
- nonparametric estimation and testing
- regression models
Results intepretations
Common errors

Examples of application with the software R will be illustrated and code fragments provided.

Requirements:

Knowledge of basic theoretical and applied survival analysis.

3. RNA-seq: from analysis workflows to high-dimensional modeling

Half-day (afternoon), June 15

Room U7-LAB712

Lecturer:

Harald Binder, Division Biostatistics and Bioinformatics, University Medical Center of the Johannes Gutenberg University Mainz, Germany

Course contents:

Next generation sequencing (NGS) comprises many different measurement platforms and approaches, e.g. for determining gene expression (RNA-seq). This includes corresponding bioinformatics and biostatistical approaches. While the basic measurement platforms are now fairly established, there is a multitude of tools for data processing and analysis, and it is difficult to establish a standard workflow for performing all necessary steps. This course will provide a brief overview of the biological and measurement platform foundations and primarily focus on data processing and statistical modeling. Specifically, different approaches for normalization and statistical testing with RNA-Seq data will be introduced. Furthermore, tools for fitting multivariable regression models and potential pitfalls will be discussed.

Basic biology and measurement platforms
Early processing steps of RNA-seq data
- Quality control
- Mapping
- Counting
Differential expression
- DESeq, edgeR, and baySeq
- Alternative splicing
Multivariable models
- Adjusting for confounders
- Regularized regression

The course includes demonstration of analysis steps in the R environment.

Requirements:

Basics of statistical testing and generalized linear models.