From: Comparison of analyses of the QTLMAS XIV common dataset. I: genomic selection
Feature Model | BayesA | BayesB | BayesC (=SSVS stochastic search variable selection) | BayesCpi |
---|---|---|---|---|
Probability for a locus to be a QTL | 1 | 1-p | 1-p | 1-p |
QTL-specific effect variance (variance heterogeneity) | Yes | Yes | No | No |
Modelling of no-QTL | Not aplicable | Null variance | Tiny variance | Null variance |
Estimated parameter | Â | Â | Â | p(uniform prior) |
Hyperparameters (assumed known) | df1, S2 | df, S, p | df, S, p | df, S |
Use Metropolis-Hastings sampler? | No | Yes | No | No |