Functionalities

Model Fitting based on Markov chain Monte Carlo

• rss_bvsr.m
  Fit the Bayesian model that consists of the RSS likelihood and the "Bayesian variable selection regression" (BVSR; Guan and Stephens, 2011) prior.
• rss_bslmm.m
  Fit the Bayesian model that consists of the RSS likelihood and the "Bayesian sparse linear mixed model" (BSLMM; Zhou, Carbonetto and Stephens, 2013) prior.
• rss_ash.m
  Fit the Bayesian model that consists of the RSS likelihood and the "Adaptive shrinkage" (ASH; Stephens, 2016) prior.

The details of the model fitting algorithms are available here.

Model Fitting based on Variational Bayes

• rss_varbvsr.m
  Fit the Bayesian model that consists of the RSS likelihood and the BVSR prior. This can be viewed as an extension of Carbonetto and Stephens (2012) for the analysis of summary-level data.

Miscellaneous

• compute_pve.m
  compute the estimated PVE (or SNP heritability, defined in Guan and Stephens (2011)) based on GWAS summary data.

• band_storage.m
  convert a symmetric, banded matrix to a compact matrix in such a way that only the main diagonal and the nonzero superdiagonals are stored.