Dear all.
I am runing a mixed model, using "mixed solve" function in rrBLUP package (R studio), and it says that the matrix X of fixed effect must be full column rank (implies β is estimable).
in my case the X matrix is a desing matrix of 44 significant markers that I want to test as fixed effect.
can any one advises me how to build this matrix ?
best regards
Hi Hafssa Kabbaj
You may want to consider using package lme4 instead of rrBLUP. The advantage of lme4's lmer command over rrBLUP's mixed.solve is that the former allows you to setup a mixed model by specifying a formula instead of design metrics. For a gentle introduction to mixed modelling using lme4, please refer to following article.
Winter, B. (2013). A very basic tutorial for performing linear mixed effects analyses. arXiv preprint arXiv:1308.5499.
In your case, once you have setup the data in the long format, the specification of the model may be as simple as follows.
Model
Dear Usman,
thank you for this suggestion, but in the case of my study, I need to use rrBLUP package.
There is function in rrBLUP package called kin.blup (reference here: https://www.rdocumentation.org/packages/rrBLUP/versions/4.6/topics/kin.blup) which uses mixed.solve in the background but makes the matrix specification a bit easier.
Here are two more tutorials:
1. https://potatobreeding.cals.wisc.edu/wp-content/uploads/sites/161/2014/01/GS_tutorial.pdf
2. https://rdrr.io/cran/rrBLUP/man/kinship.BLUP.html
Hope it is helpful
- Badges
- Science topic
- Similar topics
- Mathematics
- Algebra
- Matrix