GPmeth is a modeling framework based on GPflow. It detects differential DNA methylation / chromatin accessibility in single-cell multimodal datasets that study continuous biological processes. The idea is to generate temporal trajectories based on one modality (typically RNA-seq) and then model methylation/accessibility along these continuous axes with a Gaussian process model. GPmeth models methylation at base resolution and can automatically refine boundaries of differentially methylated/accessible regions.
GPmeth is built with GPflow, which dependso on TensorFlow and TensorFlow Probability. I recommend installing these first.
# Install tensorflow and tensorflow-probability
pip install gpflow tensorflow~=2.10.0 tensorflow-probability~=0.18.0
# Install gpmeth
git clone git@github.com:mffrank/gpmeth.git
cd gpmeth
pip install .