Define the class for the SuShiE inference results.

Variables:

priors : sushie.infer.Prior

The final prior parameter for the inference.

posteriors : sushie.infer.Posterior

The final posterior parameter for the inference.

pip_all : jax.Array

The PIP for each SNP across \(L\) credible sets.

pip_cs : jax.Array

The PIP across credible sets that are not pruned.

cs : pandas.core.frame.DataFrame

The credible sets output after filtering on purity.

alphas : pandas.core.frame.DataFrame

The full credible sets before filtering on purity.

sample_size : jax.Array

The sample size for each ancestry in the inference.

elbo : jax.Array

The final ELBO.

elbo_increase : bool

A boolean to indicate whether ELBO increases during the optimizations.

l_order : jax.Array

The original order that SuShiE infers. For example, if L=3 and it is 0,2,1, then the original SuShiE’s second effect (0-based index 1) is now third, and the original SuShiE’s third effect (0-based index 2) is now second after sorting use Frobenius norm.

Infer Public-members¶

static SushieResult(_cls, priors: Prior, posteriors: Posterior, ...)

Create new instance of SushieResult(priors, posteriors, pip_all, pip_cs, cs, alphas, sample_size, elbo, elbo_increase, l_order)

__repr__()

Return a nicely formatted representation string

__getnewargs__()

Return self as a plain tuple. Used by copy and pickle.