Browse list of available series in The Biomedical & Life Sciences Collection
Recommend The Biomedical & Life Sciences Collection to your librarian
Request to be kept updated on The Biomedical & Life Sciences Collection
Apply for a free trial of The Biomedical & Life Sciences Collection
View all talks & series with full access

Guest (sample access)

Optional Login with Username


        
> Home  /  Biomedical & Life Sciences  /  Series  /  Gene-Environment Interactions  /  Talk Details
SHARE:   Email via Outlook   Email via Web   Facebook   Twitter   |   more
Close SHARE
Please select one of the methods below to share the following page with your collegues:
Statistical Issues in Epidemiological Studies of Gene-Environment Interaction
Blogger Google Reader
Delicious Linked In
Digg MySpace
Favorites PDF Online
Gmail

Hello. As a guest, you do not have full access to this talk. Click the PLAY button below to watch the sample extract. Alternatively, you may login or review methods of obtaining more access.

    SPEAKER(S)

Dr. Peter Kraft - Harvard School of Public Health

Peter Kraft is an Assistant Professor of Epidemiology and Biostatistics at the Harvard School of Public Health. His research focuses on statistical methods for genetic association studies, with particular emphasis on gene-gene and gene-environment interaction. He has collaborated on both family- and population-based studies of cancer, asthma and schizophrenia.

Prof. Donna Spiegelman - Harvard School of Public Health

Dr. Spiegelman completed her ScD jointly in biostatistics and epidemiology at the Harvard School of Public Health in 1989, after receiving a MS in biostatistics from the same institution in 1985. She then went on to join the Department of Community Health, Division of Biostatistics, at Tufts University School of Medicine in 1989. In 1992, she accepted a position as Assistant Professor in the Department of Epidemiology at the Harvard School of Public Health, and in 2002, she was promoted to Professor of Epidemiology Methods in the Departments of Epidemiology and Biostatistics. Her primary area of statistical research is the design and analysis of epidemiologic studies with exposure measurement error, with the goal of developing new study designs coupled with new methods of analysis that permit valid effect estimates and statistical inference when there is measurement error and/or misclassification in the exposure of primary interest or potential confounders.

Talk Online Publication: Oct 2007

TOPICS COVERED IN STATISTICAL ISSUES IN EPIDEMIOLOGICAL STUDIES OF GENE-ENVIRONMENT INTERACTION

Gene-environment interaction in the development of human traits, including disease - Conceptual and analytical challenges - Joint analysis of data on both genetic and environmental factors from epidemiologic studies as a way to increase power to detect a polymorphism (or an exposure) that is associated with disease risk - Study designs - Statistical modeling issues in the analysis of data - Sample size and power calculations for a range of designs - Specification of parameters - Ordinal coding - Hierarchical models - The analysis of highdimensional "pathway" data (multiple related genes and environmental exposures)

How to cite this talk:
Kraft, P. and Spiegelman, D. (2007), "Statistical Issues in Epidemiological Studies of Gene-Environment Interaction", in Christiani, D. and Fraser, P. (eds), Gene-Environment Interactions: Role in the Modulation of Pulmonary and Autoimmune Disease Risks, The Biomedical & Life Sciences Collection, Henry Stewart Talks Ltd, London (online at http://hstalks.com/bio)

Direct talk access link:
http://hstalks.com/lib.php?t=HST11.1285_1_2&c=252

    DETAILED SLIDE INDEX

1. Introduction
2. Sir Archibald Garrod
3. JBS Haldane
4. Examples of diseases
5. Metabolism of Dietary Folate
6. Conceptual and analytic challenges
7. Outline - Designs
8. Study designs
9. Outline - simple pairwise analyses
10. Definition - NOTA BENE
11. Example: risk of disease and lof odds of disease
12. Different scale gives different in interaction (1)
13. Different scale gives different in interaction (2)
14. The standard "test for interaction"
15. Example: all relationships should be known
16. Screening for stratum-specific effects
17. Gene associated with risk in any subgroup
18. Exposure associated with risk in any genotype
19. Example: DNA repair gene XRCC1
20. Example: test for genetic effect
21. Beyond binary G and E
22. Outline - power considerations
23. Power
24. Programs for power calculations
25. Parameters to specify
26. Ge_trend example
27. Ge_trend example: results
28. Ge_trend example: results under minimum power
29. Ge_trend example: results under maximum power
30. Sample size is necessary to detect genetic effect
31. Test for "interaction"
32. Screening for genetic effect
33. Outline - more sophisticated analyses
34. Beyond one gene-one environmental factor
35. Ordinal coding
36. Ordinal coding - multiple loci
37. Ordinal coding - three parameter model
38. Hierarchical Models
39. Example [Aragaki et al. 1997]
40. Machine learning methods
41. Relation between training- and test-set error
42. Caveat Emptor
43. Final Thoughts
44. Software
45. References (1)
46. References (2)
47. Acknowledgements
48. END