Metadata-Version: 2.1
Name: bio-gopher
Version: 1.0.3
Summary: GOPHER: GenOmic Profile-model compreHensive EvaluatoR
Author-email: Shushan Toneyan <toneyan@cshl.edu>, Ziqi Tang <ztang@cshl.edu>, Peter Koo <pkoo@cshl.edu>
Project-URL: Homepage, https://github.com/shtoneyan/gopher
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Requires-Python: >=3.7
Description-Content-Type: text/markdown
License-File: LICENSE


<img src="./DALL·E 2022-10-05 14.40.17 - Constellation in a shape of groundhog. graphical art.png" width="100" height='100'>

**GOPHER**: **G**en**O**mic **P**rofile-model compre**H**ensive **E**valuato**R**

## Installation

```
$ pip install bio-gopher
```
Note that for proper installation, numpy needs to be installed before pyBigWig.

This repository contains scripts for data preprocessing, training deep learning models for DNA sequence to epigenetic function prediction and evaluation of models.

The repo contains a set of tutorial jupyter notebooks that illustrate these steps on a toy dataset. The two notebooks below are required prerequisites for the rest of tutorials:
- preprocessing/preprocessing/quant_dataset_tutorial.ipynb
- tutorials/train_model.ipynb


To replicate the results of the manuscript run the scripts in the analyzis directory. As a prerequisite download and unzip dataset.zip, trained_models.zip from zenodo https://doi.org/10.5281/zenodo.6464031 within the git repo. These contain test sets and pre-trained models. The analysis scripts can be ran in any order as long as paper_run_evaluate.py is ran first, in order to produce model evaluations which is required for further steps.
