Hi everyone,
I am trying to estimate the narrow-sens heritability of dispersal (propensity) in a natural population of lizard for wich I have a pedigree dataset. I used a bayesian approch, by fitting an animal model (function : brm ; package : brms) with three random effects (additive genetic, maternal effect, environnement). Using the posterior distribution of this effects, I was abble to estimate h² and the credibility interval, IC (function : HPDinterval ; package : coda).
My issue is that IC is very large, with the lower limit beeing very close to zero (e. g. for an estimate of h² = 0.30, IC [2e-12 ; 0.5]). How can I by sure that my estimation is robuste and that h² is different of zero ?
For now I fitted the animal model without the additive genetic effet and used loo and looic methods on both model. Comparing the output, I am sure that the additive genetic effect increase the quality of the model, help it to fitt best the data.
Is that enough ? Or do you see anything else I can do to assert / evaluate the robustness of heritability estimate ?
Thanks in advance.
- Badges
- Science topic
- Similar topics
- Agricultural Science
- Agronomy
- Crop Science