Interpretation of Genetic Association Studies: Markers with Replicated Highly Significant Odds Ratios May Be Poor Classifiers
2009

Evaluating Genetic Testing for Disease Risk

publication Evidence: moderate

Author Information

Author(s): Jakobsdottir Johanna, Gorin Michael B., Conley Yvette P., Ferrell Robert E., Weeks Daniel E.

Primary Institution: University of Pittsburgh

Hypothesis

How useful are highly associated SNPs for individual-level risk estimation and prediction?

Conclusion

Strong genetic associations do not guarantee effective discrimination between cases and controls in personalized medicine.

Supporting Evidence

  • Strong associations do not guarantee effective discrimination between cases and controls.
  • The AUC for the three-factor model of AMD was 0.79, but only 30% of those classified as high risk were actual cases.
  • Logistic regression analysis showed odds ratios of around 3 for significant SNPs.
  • Genetic testing results are often poorly understood by individuals and physicians.

Takeaway

Scientists are trying to figure out if genetic tests can really help us know if we will get sick, but just because a gene is linked to a disease doesn't mean it can accurately tell who will get it.

Methodology

The study used logistic regression and ROC curve analysis on genetic data related to age-related macular degeneration and other diseases.

Potential Biases

Potential overinterpretation of genetic findings in personalized medicine.

Limitations

The study highlights that strong associations do not ensure good classification ability and that the effectiveness of genetic markers needs to be formally established.

Participant Demographics

The study discusses various age groups, particularly focusing on those 40 years and older for AMD.

Statistical Information

P-Value

10−13

Confidence Interval

0.74–0.83

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1371/journal.pgen.1000337

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication