index.Rmd

---
title: "Home"
site: workflowr::wflow_site
output:
  workflowr::wflow_html:
    toc: false
---

[Introduction.](intro.html) A vignette.

[MASH v FLASH simulation results.](MASHvFLASHsims.html) Compares FLASH fits to MASH fits in a number of settings. Uses various methods to fit the FLASH object. The "OHL" method is best.

[MASH v FLASH on GTEx.](MASHvFLASHgtex.html) Compares FLASH to MASH using data from the GTEx project. Results are similar, but FLASH is more conservative.

[MASH v FLASH detailed simulation study.](MASHvFLASHsims2.html) Using the "OHL" method, compares the performance of FLASH-pn, FLASH-ash, and MASH in a single simulation setting. FLASH and MASH perform differently on different covariance structures. There is little or no advantage to using ash priors rather than point-normals.

[MASH v FLASH detailed GTEx study.](MASHvFLASHgtex2.html) Uses an improved workflow to examine differences in MASH and FLASH fits to GTEx data. Particularly strong effects in a single condition might cause MASH to miss smaller effects that are shared among conditions. But FLASH might miss very small effects that are shared across all conditions.


## Archived:

[A first attempt.](mashvflash.html) Stowed and forgotten.