Closed sofiapuvogelvittini closed 2 years ago
Try zlm(..., method ='glmer', fitArgsC=list (REML=FALSE))
. You can pass
other arguments with fitArgsC
and fitArgsD
.
On Wed, Apr 21, 2021, 5:05 AM sofiapuvogelvittini @.***> wrote:
Dear MAST team, I would like to ask you a question.. I am using MAST for modelling gene expression in a snRNAseq data, with a random intercept for subjects identity. I am wondering if in this case, the parameters are still estimated with LM ( as mentioned in the paper) or does it automatically switch to RELM? I would like to use ML as I want to select the best model based on the deviance of the residuals.
My formula is the following zlmCond<-MAST::zlm(~condition + cenegeneson + test_var1 + test_var2 + test_varx + (1|donor_id), data, method=´glmer´,ebayes=FALSE)
Thanks for your help!
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/RGLab/MAST/issues/157, or unsubscribe https://github.com/notifications/unsubscribe-auth/AALLAHXK2ZNLLDZM22AG5DDTJ2INRANCNFSM43J5T7KA .
Hello,
I tried using the solution mentioned by amcdavid, but it results in the error:
Error in (function (formula, data = NULL, REML = TRUE, control = lmerControl(), :
formal argument "REML" matched by multiple actual arguments
I made a code snippet to reproduce the error:
data(maits, package='MAST')
sca <- FromMatrix(t(maits$expressionmat), maits$cdat, maits$fdat)
conditional_model <- zlm(~(PercentToHuman:condition)+ (1|condition), sca, exprs_value='et', method = 'glmer', ebayes = FALSE, fitArgsC=list(REML=FALSE))
I am using MAST version 1.20 on R version 4.1.1
Am I using fitArgsC wrong, or could it be something else entirely?
Thanks in advance.
@whreitsma @sofiapuvogelvittini
Hmm, actually it looks like I had set REML = FALSE
all along.
This now can be overridden with fitArgsC
in the latest development version 31603fe.
Dear MAST team, I would like to ask you a question.. I am using MAST for modelling gene expression in a snRNAseq data, with a random intercept for subjects identity. I am wondering if in this case, the parameters are still estimated with LM ( as mentioned in the paper) or does it automatically switch to RELM? I would like to use ML as I want to select the best model based on the deviance of the residuals.
My formula is the following zlmCond<-MAST::zlm(~condition + cenegeneson + test_var1 + test_var2 + test_varx + (1|donor_id), data, method=´glmer´,ebayes=FALSE)
Thanks for your help!