I am analyzing pedigrees to determine if families have a cancer susceptibility gene mutation based on their family history of cancer (various types). I have created a dataset for Google AutoML Tables with 250,000 generated pedigrees. I would like to use this to develop a classifier to predict if a given pedigree has a familial mutation.
What I am interested in, though, is not a yes/no answer, but rather the Bayesian posterior probability. In other words, what is the probability of this family carrying a mutation given a positive result. Is it possible to do this with AutoML Tables? If so, how would I go about this?
