search

stephenslab / susieR

R package for "sum of single effects" regression.
https://stephenslab.github.io/susieR
Other
176 stars 46 forks source link

XtX + t(XtX)) #152

Open MoMaz123 opened 2 years ago

MoMaz123 commented 2 years ago

Hi, am trying to run the code, but I keep having the same error even after doing as requested, please advise. thanks,Mo error: XtX is not symmetric; forcing XtX to be symmetric by replacing XtX with (XtX + t(XtX))/2

pcarbo commented 2 years ago

Hi @MoMaz123 this should be a message and not an error; you are most likely seeing this because there are small differences between XtX and t(XtX). You may be able to ignore this message. Let us know if you have further questions.

MoMaz123 commented 2 years ago

Hi Pcarbo, thanks for the reply, the problem is that I can see anything when I run

summary(fit)$cs susie_plot(fit, y="PIP") print(fit$sets$cs)

all of them are "null". please advise, Thanks, mo

pcarbo commented 2 years ago

It sounds like susie is estimating from your data that there are no effects (hence no credible sets). Could you tell us a bit more about your data, and how you called susie_suff_stat or susie_rss?

privefl commented 1 year ago

I have a similar warning about XtX not being symmetric, whereas isSymmetric(R) is TRUE.

pcarbo commented 1 year ago

Thanks, @privefl. Do you have the Rfast R package installed?

privefl commented 1 year ago

Yes, I do.

pcarbo commented 1 year ago

Okay, my expectation is that Rfast::is.symmetric(x) will return false

privefl commented 1 year ago

You could maybe simply use base R function isSymmetric()

pcarbo commented 1 year ago

My guess is that the Rfast function is has a lower tolerance. Either way, I would just ignore the warning.