Skip to content

Interpretable genotype-phenotype landscape modeling

License

Notifications You must be signed in to change notification settings

usnistgov/lantern

Repository files navigation

LANTERN: an interpretable genotype-phenotype landscape model

DISCLAIMER: LANTERN is currently in alpha and the primary interface is subject to change

Documentation

The official documentation with usage is available at: https://lantern-gpl.readthedocs.io. Documentation covers instructions for running lantern on GPL data, including a demonstration on an example dataset. On a typical desktop, examples should run in under ten minutes.

System Requirements

Hardware requirements

lantern requires only a standard computer with enough RAM to support the in-memory operations. Runtime of model training on large-scale GPL datasets ($n \geq 10,000$) will benefit from GPU hardware but is not strictly necessary.

Software requirements

This package is supported for macOS and Linux. The package has been tested on the following systems:

  • macOS: Catalina (10.15)
  • Linux: Ubuntu 18.04

Python Dependencies

lantern depends primarily on pytorch and gpytorch, as well as the components of the Python scientific stack:

  • pandas
  • numpy

Installation guide

Installation time for lantern is typically less than one minute.

Install from github

LANTERN currently must be installed from this repository directly, either through pip:

python -m pip install git+https://github.com/usnistgov/lantern.git

of by cloning this repository:

git clone https://github.com/usnistgov/lantern.git
cd lantern
python setup.py install

In either case, it is recommended to install inside a virtual environment.

Manuscript reproduction

Source code to reproduce the analysis of the LANTERN manuscript are available at github.com/usnistgov/lantern/tree/master/manuscript.

License

This project is covered under the NIST Software License