Skip to content

irreversibly/xwas_with_nhanes_tutorial

 
 

Repository files navigation

Implement code to execute an X-wide association study (XWAS) in telomere length

  • X-wide association study (XWAS) is a data-driven method to find what variables in the set 'X' are associated with a phenotype (call it Y, e.g., telomere length)
  • Originally presented at ISEE/ISES 2018, Ottawa, Canada on August 26 2018

Contact information:

Prerequisites for the tutorial:

  • familiarity with programming and data structures
  • familiarity with R and regression

Software requirements

  • R (version 3.3.3 or greater)
  • Rstudio (version 1.01 or greater)
  • Packages: survey, tidyverse, knitr, and broom

Execution

Slide Downloads

Original paper:

XWAS in Telomere Length

Questions:

  1. What is your interpretation of the association sizes and pvalues for your XWAS?
  2. How much to the coefficients change for the adjustment variables for each of the correlations?
  3. Attempt to reproduce the findings in Patel et al., IJE 2016 (https://www.ncbi.nlm.nih.gov/pubmed/27059547) using more categories of exposure. How will you handle other X variables, such as self-reported variables?
  4. When increasing the number of variables, how would the pvalue threshold change to accommodate more tests? How would the FDR change, if at all?
  5. Execute the XWAS in another phenotype. What are the similarties and differences between your analysis in 1.
  6. How much variance explained in telomeres do the top factors explain? Is this to be expected?
  7. Implement the XWAS without using the for operator using the tidyverse suite of commands.

About

Tutorial on doing an XWAS in NHANES

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • HTML 100.0%