This repository contains code associated with the paper A generative nonparametric Bayesian model for whole genomes (2021) (Alan N. Amin*, Eli N. Weinstein*, Debora S. Marks), which proposes Bayesian embedded autoregressive (BEAR) models. The repository provides example BEAR models as well as tools for implementing new models. It enables building, training and evaluating BEAR models on large scale sequencing datasets, including whole genome, transcriptomic and metagenomic data.
For instructions on running examples and deploying the BEAR model, consult the documentation at https://bear-model.readthedocs.io/en/latest/.
This is a project of the Marks Lab in the Systems Biology Department at Harvard Medical School. It was developed by
- Alan Amin, <alanamin@g.harvard.edu>
- Eli Weinstein, <eweinstein@g.harvard.edu>
- Debora Marks, <debbie@hms.harvard.edu>
This project is available under the MIT license.
Alan N. Amin*, Eli N. Weinstein*, and Debora S. Marks. A generative nonparametric Bayesian model for whole genomes. Advances in Neural Information Processing Systems (NeurIPS). 2021. (* equal contribution) https://proceedings.neurips.cc/paper/2021/hash/e9dcb63ca828d0e00cd05b445099ed2e-Abstract.html