Developed and maintained:
-
CARMA :
We propose a novel Bayesian model, CARMA (CAusal Robust mapping method in Meta-Analysis studies), for fine-mapping in order to identify putative causal variants at GWAS loci. The main features of CARMA are
-Modeling jointly summary statistics and high-dimensional functional annotations -Introducing a novel Bayesian hypothesis testing approach to account for discrepancies between summary statistics and LD from external reference panels -flexible specification of the prior distribution of effect sizes including a heavy-tail Cauchy distribution that is more adaptive to the possibly different signal-to-noise ratios across loci
-
PO.EN :
Presence-only model with Elastic Net penalty is a regularized generalized linear model training on the presence-absence response. This package provides functions for tuning and fitting the presence-only model. The presence-only model can be used to predict regulatory effects of genetic variants at sequence-level resolution by integrating a large number of epigenetic features and massively parallel reporter assays (MPRAs).