GFA Credible Intervals — gfa

Compute credible intervals for various estimated parameters.

Usage

gfa_credints(
  gfa_fit,
  nsamp = 500,
  level = 0.05,
  type = c("pve", "F", "L"),
  variant_ix = NULL
)

Arguments

gfa_fit: An object output by gfa_fit.
nsamp: Number of posterior samples.
level: Level for credible intervals (e.g. 0.05 results in 95% credible intervals).
type: A vector including some or all elements "pve", "F", or "L", see details.
variant_ix: Optional list of indices if type includes "L".

Details

This function will compute credible intervals for the PVE matrix, F, or L depending on the choice in "type". The returned object will include <type>\_lower and <type>\_upper for each element in type.

Please note that computing credible intervals for all elements of L can be very time consuming, since L is typically very large. Your needs may be better served by computing frequentist estimates and p-values for L using gfa_loadings_gls().