fcfdr is an all-encompassing R package to implement the cFDR approach
for a variety of auxiliary covariates. As input it requires GWAS
p-values for a trait of interest and auxiliary data values. It then
outputs “v-values” which can be interpreted as GWAS p-values that have
been adjusted using the auxiliary data values.
If you have any questions please do not hesitate to contact me:
GitHub repository: https://github.com/annahutch/fcfdr
Hutchinson A, Reales G, Willis T, Wallace C (2021) Leveraging auxiliary data from arbitrary distributions to boost GWAS discovery with Flexible cFDR. PLoS Genet 17(10): e1009853. https://doi.org/10.1371/journal.pgen.1009853
Hutchinson A, Liley J, Wallace C (2021) fcfdr: an R package to leverage continuous and binary functional genomic data in GWAS. bioRxiv 2021.10.21.465274. https://doi.org/10.1101/2021.10.21.465274
You can install
fcfdr from GitHub with:
# install.packages("devtools") devtools::install_github("annahutch/fcfdr")
See the vignettes for examples of usage.
Brief history of cFDR
The cFDR framework was first introduced by Andreassen and colleagues in 2013. The group also applied the approach to uncover new genetic associations for blood pressure and multiple sclerosis by leveraging GWAS test statistics for related traits.
In 2015, Liley and Wallace showed that the conventional cFDR approach does not control the frequentist FDR and described an extension of the approach to ensure that FDR is controlled, however their approach is rather conservative. In 2021, Liley and Wallace described an elegant cFDR framework that takes as input GWAS p-values for related traits and returns “v-values” which can be interpreted as GWAS p-values that have been adjusted using the auxiliary related trait data, and thus can be used directly in any error-rate controlling procedure (e.g. Benjamini-Hochberg).
Up until this point, the cFDR framework was designed for a very specific setting, that is to increase GWAS discovery (in the “principal trait”) by leveraging GWAS test statistics from a genetically related (“conditional”) trait. In 2021 we described an extension of the cFDR framework called Flexible cFDR that allows for auxiliary covariates sampled from arbitrary continuous distributions to be leveraged with GWAS test statistics - thus enabling broader applicability. We also describe an extension for binary covariates, called Binary cFDR. In practice, Flexible cFDR and Binary cFDR can be used to leverage functional genomic data with GWAS results to boost power for GWAS discovery whilst controlling the FDR.