 |
 |
|
|
| Date/Time | Activity | Topic/Content | Faculty | Locations |
| Tues 3/31/2009 8:15 AM-10:15 AM | Lecture | Permutation tests After outlining the course, we will discuss permutation tests - a simple but widely-applicable class of statistical tests. The beauty of permutation tests lie in their minimalist simplicity. This comes at the cost of computational complexity. They constitute a class of non-parametric statistical methods that make as few assumptions as possible. | S Sen | CB 5721 |
| Tues 4/7/2009 8:15 AM-10:15 AM | Lecture | Likelihood and likelihood ratios Likelihood methods provide powerful statistical methods to extract information from data if we have credible knowledge on the nature of random variation in data. Competing hypotheses can be compared by calculating likelihood ratios based on observed data. We will discuss a few examples of likelihood methods in genetics including the ubiquitous LOD score. We will introduce the class to PLINK, a software program for genomewide genetic analysis. | S Sen | CB 5721 |
| Tues 4/14/2009 8:15 AM-10:15 AM | Lecture | Family based association tests The TDT tests for the absence of both linkage and association, and is robust to population structure. After reviewing its connections to a randomized trial, we will discuss its extensions to quantitative traits, and other family-based tests of association. | S Sen | CB 5721 |
| Tues 4/21/2009 8:15 AM-10:15 AM | Lecture | Population based candidate gene association studies We will discuss population-based association studies including case-control studies. We will discuss methods for probabilistic reconstruction of haplotypes when parental data on subjects in unavailable, and see examples of how to implement such statistical analysis. | S Sen | CB 5721 |
| Tues 4/28/2009 8:15 AM-10:15 AM | Lecture | Advanced methods: population structure and multiple comparisons Population structure has the potential to confound association studies. We will discuss three methods to detection and adjustment: genomic control, Bayesian population membership estimation (STRUCTURE) and principal components. Genomewide studies necessitate multiple comparisons, which has the potential to increase false positives. We will discuss how to approach the problem, and techniques such as the Bonferroni correction and the False Discovery Rate. | S Sen | CB 5721 |
| Tues 5/5/2009 8:15 AM-10:15 AM | Lecture | Genomewide association studies: Methods We will discuss the design and analysis of genomewide association studies. These will include genotyping strategies, selection of cases and controls, single and multi-point analyses, multiple comparisons, choice of validation and followup studies. | S Sen | CB 5721 |
| Tues 5/12/2009 8:15 AM-10:15 AM | Lecture | Genetics case study with discussion Case study: "Telomere length, aging and genetics: what have we learned from studies of populations and families" | W Hsueh | CB 5721 |
| Tues 5/19/2009 8:15 AM-10:15 AM | Lecture | Genetics case study We will take a brief tour of modern machine learning techniques for prediction derived from the computer science literature. Case study:"Searching for the genetic basis for a complex cognitive trait - absolute pitch" | J Gitschier | CB 5721 |
| Tues 5/26/2009 8:15 AM-10:15 AM | Lecture | Student presentations Session 1Session 2 | E Jorgenson | CB 5721 |
| Tues 6/2/2009 8:15 AM-10:15 AM | Lecture | Student presentations Session 3Session 4 | E Jorgenson | CB 5721 |
