#### The support.bioconductor.org editor has been updated to markdown! Please see more info at: Tutorial: Updated Support Site Editor

Question: variancePartition - testing for individuals
0
24 days ago by
rina0
rina0 wrote:

Hi all!

I am analyzing the variance sources of TCGA expression data using variancePartition. I want to check among others the effect of individuals in the variance, but when I specify it at the formula, I get the following errors:

> form <- ~  submitter_id
> varPart <- fitExtractVarPartModel(Filt_EXP1, form, clin)
Error in checkModelStatus(fit, showWarnings = showWarnings, colinearityCutoff) :
Colinear score = 1 > 0.999
Covariates in the formula are so strongly correlated that the
parameter estimates from this model are not meaningful.
Dropping one or more of the covariates will fix this problem

> form <- ~ (1|submitter_id)
> varPart <- fitExtractVarPartModel(Filt_EXP1, form, clin)
Error: number of levels of each grouping factor must be < number of observations

> form <- ~ (0|submitter_id)
> varPart <- fitExtractVarPartModel(Filt_EXP1, form, clin)
Error in (function (cl, name, valueClass)  :
assignment of an object of class “numeric” is not valid for @‘Dim’ in an object of class “dgTMatrix”; is(value, "integer") is not TRUE


I can see why the above return an error, but at the vignette effect of individuals can be tested for. Is it a specific way I should specify it?

R.

expression formula variance • 54 views
modified 23 days ago by mikhael.manurung10 • written 24 days ago by rina0
Answer: variancePartition - testing for individuals
0
23 days ago by
mikhael.manurung10 wrote:

Dear Rina,

To identify the proportion of variance due to between-individual differences, you should provide the information in the metadata.

Your second version of the formula is the correct way to define it (the first formula will treat submitter_id as a continuous variable). And keep in mind to define the variable submitter_idas a factor.

Best,

Mikhael