search

stephenslab / mashr

An R package for multivariate adaptive shrinkage.
https://stephenslab.github.io/mashr
Other
88 stars 19 forks source link

Data-driven priors based on betahat or z-scores #37

Open gaow opened 6 years ago

gaow commented 6 years ago

Currently this line gives a warning when it appears betahat is used to compute the data-driven matrices. But the message is ambiguous and does not read like any recommendation.

Should we choose for the users one way or another, and hide this message?

stephens999 commented 6 years ago

This was a note to myself to remind me to think harder about this issue.... Which I have not yet done! I agree it is confusing to users. Let's remove the warning and keep this issue open?

On Thu, Jun 14, 2018, 11:28 gaow notifications@github.com wrote:

Currently this line https://github.com/stephenslab/mashr/blob/c4e0063604175835619b312342517bb60a60abbb/R/data2cov.R#L16 gives a warning when it appears betahat is used to compute the data-driven matrices. But the message is ambiguous and does not read like any recommendation.

Should we choose for the users one way or another, and hide this message?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/stephenslab/mashr/issues/37, or mute the thread https://github.com/notifications/unsubscribe-auth/ABt4xRA51EKrO8PJSfp9J8aRdxXdyValks5t8o9IgaJpZM4UoRxS .