lpi-prediction 0.0.1

A template library to build lncRNA-protein interaction prediction algorithms

LPI-Prediction: A template library to build lncRNA-protein interaction prediction algorithms

LPI-Prediction aims to be a simple and easy-to-use template library to develop lncRNA-protein interaction prediction algorithms, or any prediction algorithm in general. LPI-Prediction provides a set of abstractions to load and process data, and a set of abstractions to train, evaluate, and execute the interaction prediction algorithms. LPI-Prediction is designed to be easily extensible, and to greatly lower the cost of developing production-level lnRNA-protein interaction prediction algorithms.

Installation

LPI-Prediction is available on PyPI, and can be installed using pip as follows:

user@host$ python3 -m pip install lpi-prediction

Please note that LPI-Prediction requires Python3.7 or higher.

Examples

LPI-Prediction provides a set of examples to demonstrate how to use the library. The examples are available in the examples directory of the repository. The examples are:

• knn.py: A simple example that shows how to load and process data, and how to train and evaluate a K-Nearest Neighbors (KNN) prediction algorithm.
• svm.py: A simple example that shows how to load and process data, and how to train and evaluate a Support Vector Machine (SVM) prediction algorithm.
• ensemble.py: A simple example that shows how to load and process data, and how to train and evaluate an ensemble of prediction algorithms (KNN + SVM).

Moreover, MIRLO shows how to use LPI-Prediction to develop a production-level lncRNA-protein interaction prediction algorithm that does automatic hyperparameter tuning, and uses third party applications (Rscript) to extract features from the lncRNA and protein sequences.

License

LPI-Prediction is licensed under the MIT License. See the LICENSE file for more information.

Authors

LPI-Prediction was developed by Iñaki Amatria-Barral.