• Simulation
  • 1 Intro
  • 2 Setting up
    • 2.1 Install packages
    • 2.2 Import Hapmap dataset
  • 3 Generate a sub-dataset
    • 3.1 Pick 2000 SNPs from 2 chromosomes as our small dataset
    • 3.2 Remove the monomorphic SNPs in this subset
  • 4 Make up a fake disease trait
    • 4.1 Choose four causal SNPs
    • 4.2 Generate disease status
  • 5 Conventional p-value procedure
    • 5.1 Conduct logistics regression on each SNP
    • 5.2 Get the most important 10 SNPs by ranking their p-values
  • 6 PLIS procedure
    • 6.1 Setting up
    • 6.2 Run the PLIS procedure
    • 6.3 Get the most important 10 SNPs by ranking LIS
  • 7 Compare the sensitivity of these two methods
    • 7.1 Get the sensitivity of PLIS procedure
      • 7.1.1 Get the sensitivity of PLIS method on chromosome a
      • 7.1.2 Get the sensitivity of PLIS method on chromosome b
    • 7.2 Get the sensitivity of conventional p-value method
      • 7.2.1 Get the sensitivity of conventional p-value method on chromosome a
      • 7.2.2 Get the sensitivity of conventional p-value method on chromosome b
    • 7.3 Plot the sensitivity trends of PLIS and p-value methods
      • 7.3.1 Stationary
      • 7.3.2 Animation
    • 7.4 Plot the sensitivity trend of PLIS and p-value methods – grouped by chromosome
      • 7.4.1 Stationary
      • 7.4.2 Animation
  • References
  • Published with bookdown

Using simulation to compare sensitivity of PLIS procedure and conventional p-value procedure

References

https://academic.oup.com/bioinformatics/article/25/21/2802/226040