Open camult opened 6 years ago
@camult I'm thinking about your question again (initially sent via email). Is your aim to fit a Bayesian variable selection model (e.g., as implemented by varbvs), except that the samples are no longer independent, but correlated according to some relationship matrix? If so, this is the BSLMM model described in our PLoS Genetics paper. Alternatively, if you are thinking about a basic linear mixed model, you might want to consider other R packages, or software such as GEMMA or GCTA.
Perhaps you can give more details about your data and/or model.
Hi Peter,
I wonder if you could add an option to fit a linear mixed model, where we can consider the random effect of an individual, by passing a relationship matrix (Genomic relationship matrix or any other kernel).
Thank you