Genotype-phenotype modeling with longitudinal imaging data


I am trying to use PLSR for genotype-phenotype modeling using longitudinal imaging data from 5 different time points. My matrix of predictors (X) is the numerical SNP data that is repeated 5 times for each individual to match the dimension of the response matrix (Y) and the design vector is the list of samples ID in the same order as both matrices. But when I run the code I get the following error:

Error in if (max(sapply(1:J, function(x) { :
missing value where TRUE/FALSE needed

When I include a continuous variable with variation across different time points in the matrix of predictors, I get the following warning:

Warning message:
In cor(A[[k]], variates.A[[k]]) : the standard deviation is zero

Is there a way to resolve this? If MixOmics does not support this type of modeling, do you know any other package that allows this?

I really appreciate your help.


Dear @Hengi,

the problem comes from repeating the values in the SNP dataset which create variances = 0. Probably, the best thing to do is first to ask yourself the question you want to answer with the time-course imaging data, in relation with the SNP data.
One solution would be to focus on a particular time point, the second would be to model first the imaging longitudinal data to summarise their expression across time (I dont have answers for the latter by the way! our lmms and timeomics frameworks, both on CRAN and Bioconductor summarise the values at the individual level, not the time level).