This intensive, residential, computational course is aimed at scientists actively involved in genetic analysis of either rare (Mendelian) or complex human traits who anticipate using state-of-the-art statistical analysis techniques on genetic data collected on related and unrelated individuals.The programme provides a comprehensive overview of the statistical methods currently used to map disease susceptibility genes in humans and non-model organisms with an emphasis on data collected on families or populations (which should often be considered a collection of large families).
This is a small residential course, with a low student to instructor ratio, personalized attention, and the instructors actively involved throughout the week. Students present on their own research to the group and receive constructive criticism particularly pertaining to study design and analysis. This course is unique among statistical genetics courses in that it concentrates on approaches that capitalize on families or a combination of families and unrelated individuals in the post-GWAS era.
Why does this course emphasize family data?
In the GWAS and post-GWAS era, gene mapping has concentrated on analysis of unrelated individuals due to simplicity and convenience. However, these approaches tend to treat any relatedness among individuals as a nuisance to be adjusted away rather than a benefit to be exploited. Furthermore, researchers are increasingly aware that the use of unrelated individuals has limitations that family data can overcome. Family studies have many advantages in gene mapping:
- They are extremely powerful in situations where unrelated individuals lack power (e.g., when rare variants underlie the aetiology) since related affecteds are more likely to share the same disease predisposing gene than unrelated affecteds.
- They overcome confounding factors such as population stratification and allow better modelling of environmental factors.
- They allow the examination of a wealth of nuanced genetic models.
- They provide ways to rule out artefacts and false associations that can plague genetic analyses. This course will enable participants to make better use of their data that may include related individuals.
During this course, discussions of the latest statistical methodology are complemented by practical hands-on computer exercises using state-of-the-art software. The statistical principles behind each method will be carefully explained so that participants with a non-statistical background can understand and better interpret their results. Note, however, that the bioinformatics pipelines for calling variants from next generation sequencing data are not covered; the focus of this course is on the downstream analysis of the called variants.
This course is aimed primarily at advanced Ph.D. students and post-docs who are early in their careers, whose projects involve data that could be analysed by the methods covered in this course. Since we emphasize methods for handling family data, there is a preference for candidates who have some family data or who are likely to have access to family data in the near future. Programming experience is not required, but candidates without prior experience with the Unix/Linux/Mac command line will be expected to read through a tutorial on this topic prior to the course.
The programme will discuss fundamental issues needed to increase success in gene mapping studies including:
1. Why families?
Contrasting family and population study designs
Practical aspects of collecting family data
2. Association analysis in samples of unrelated and related individuals
Linear mixed models (LMM, aka variance components)
3. Linkage analysis as an effective tool for gene mapping in the post-GWAS era
4. Quality control strategies
When only using unrelateds
When families are included
5. Using families in order to move beyond simple genetic models
6. Haplotyping using GWAS and sequencing data
7. Analysis of rare traits using sequencing data from families
8. Risk prediction, meta-analysis, and other post-GWAS analyses using families
Dan Weeks (Course Organiser)
University of Pittsburgh, USA
Institute of Genetic Medicine, Newcastle University, UK
Centre Nacional d’Anàlisi Genòmica (CNAG), Spain
University of California, Los Angeles, USA
University of California, Los Angeles, US
Columbia University, New York, USA
University of Dundee, UK
University of Helsinki, Finland
Johns Hopkins University, USA
Erasmus Medical Centre, Rotterdam, The Netherlands
University of California, Los Angeles, USA