Last updated: 2018-09-21

workflowr checks: (Click a bullet for more information)
Expand here to see past versions:


An introductory vignette.

MASH v FLASH simulation results. Compares MASH and FLASH fits to data simulated from MASH and FLASH models.

MASH v FLASH detailed simulation study. Compares the performance of MASH and FLASH on different covariance structures in a single simulation setting.

The next two analyses compare MASH and FLASH fits using datasets from the GTEx project.

MASH v FLASH workflows for GTEx. Proposes improved workflows to obtain MASH and FLASH fits to GTEx data. Describes the MASH fit and five different FLASH fits (“OHF”, “Top 5”, “Top 10”, “Top 20”, and “Zero”) in detail.

Next up is a series of analyses that examine individual tests in an effort to identify salient differences among fits:

Next, I experiment with putting nonnegative priors on loadings:

FLASH v CorShrink. Here I respond to Dey and Stephen’s claim that FLASH “grossly distorts” correlation estimates.


This reproducible R Markdown analysis was created with workflowr 1.0.1