MissenseHMM

MissenseHMM: state-based annotations for missense variants through joint modeling of pathogenicity scores. The pre-generated annotations can be found in the annotations_v1.0 folder and also at https://public.hoffman2.idre.ucla.edu/ernst/8WTU7/.

Requirements

First, download dbNSFP data from https://www.dbnsfp.org/. Modify the IN_LOC and VERSON variables in config.yaml accordingly.

Next, download the ChromHMM software from https://compbio.mit.edu/ChromHMM/. Modify the CHROMHMM_LOC variable in config.yaml to point to the path to the ChromHMM.jar file.

Make sure you have Java >= 1.8.0_281 installed and the Python packages listed in requirements.txt.

Training and applying the model

Use ./submit to run the full Snakemake pipeline.