In NRM the relationship starts from 0 and hence skewed towards positive side. GRM has values on both positive and negative sides. In reality how come animals can be negatively related? Or is it just an expression of relationship that can be transformed to positive values by adding a constant? and if yes, what constant to be added?
I would say that thinking of relationships as a probability of identity by descent is probably misleading, and it is more useful to think of relationships as conditional correlations. Correlations can easily be negative--a negative correlation of genotypes means that detecting an allele in one individual makes it less likely to detect the allele in the other individual. When populations are structured, negative correlations of both genotypes and phenotypes may exist between individuals. Because phenotypes can be negatively correlated, the GRM values should not be transformed to be non-negative unless a similar transformation is performed on phenotypic values. Just my two cents.
Thanks for the answer.
However, how good it would be to look relationships as correlations?
Shouldn't be the values of gmatrix more than 1.0? eg. 6, 7, 9 like that?
- Badges
- Science topic
- Similar topics
- Biological Science
- Kinesiology
- Running