empirical_bayes.Rdempirical_bayes is a function which uses summary statistics to correct
for bias induced by Winner's Curse in SNP-trait association estimates,
obtained from a discovery GWAS. The function is strongly based on the method
originally detailed in
Ferguson et
al. (2013). However, the function also includes all potential adaptations to the empirical Bayes method
discussed in Forde et al. (2023).
empirical_bayes(summary_data, method = "AIC")A data frame containing summary statistics from the
discovery GWAS. It must have three columns with column names rsid,
beta and se, respectively, and columns beta and
se must contain numerical values. Each row must correspond to a
unique SNP, identified by rsid. The function requires that there must
be at least 50 SNPs as any less will result in very poor performance of the
method.
A string which allows the user to choose what modelling approach
to take for the purpose of estimating the log density function. The default
setting is method="AIC", which is the current published method. Other
options include method="fix_df", method="scam",
method="gam_nb" and method="gam_po". If
method="fix_df", the degrees of freedom is set to 7. The other three
options all enforce additional constraints on the shape of the estimated log
density function.
A data frame with the inputted summary data occupying the first three
columns. The new adjusted association estimates for each SNP are returned in
the fourth column, namely beta_EB. The SNPs are contained in this
data frame according to their significance, with the most significant SNP,
i.e. the SNP with the largest absolute \(z\)-statistic, now located in the
first row of the data frame.
Ferguson, J. P., Cho, J. H., Yang, C., & Zhao, H. (2013). Empirical Bayes correction for the Winner's Curse in genetic association studies. Genetic epidemiology, 37(1), 60\(-\)68.
Forde, A., Hemani, G., & Ferguson, J. (2023). Review and further developments in statistical corrections for Winner’s Curse in genetic association studies. PLoS Genetics, 19(9), e1010546.
https://amandaforde.github.io/winnerscurse/articles/winners_curse_methods.html
for illustration of the use of empirical_bayes with a toy data set and
further information regarding the computation of the adjusted SNP-trait
association estimates.