Uncertainty Quantification for AMPL Point Predictions

Overview

This repository contains the research work on applying machine learning models to predict the binding or inhibition activity of NimA-related kinases. The project is part of a collaboration with the Accelerating Therapeutics for Opportunities in Medicine (ATOM) Consortium. It focuses on addressing challenges posed by imbalanced and scarce experimental data through uncertainty quantification techniques.

Installation

To set up this project locally, follow the steps below:

git clone https://github.com/TsukiTiger/NEK9_Final.git
cd NEK9_Final
pip install -r requirements.txt

Usage

To run the main analysis notebook, use the following command:

code .
cd notebooks

Ensure you have the necessary data files in the appropriate directories as expected by the scripts.

Data

Due to confidentiality agreements, raw data files are not included in this repository. Data used in this project are part of the ATOM Consortium's private datasets. Please ensure you have the correct permissions to access the data.

Models

This project includes several machine learning models aimed at classifying drug compounds. The models are saved in the models/ directory after training and can be loaded for further analysis or prediction as follows:

import joblib
model = joblib.load('models/model_name.pkl')

Contributing

We welcome contributions from the community. Please fork the repository and submit a pull request with your proposed changes. For major changes, please open an issue first to discuss what you would like to change.

Contact

• Chongye Feng - chongyef@gmail.com
• Ya Ju Fan, PhD - Mentor's Email Here
• Amanda Paulson, PhD - Mentor's Email Here