Unsupervised Machine Learning for Adaptive Immune Receptors with immuneML: use case 1

Résumé fourni par la source

This file contains all the results from use case 1 of the manuscript Unsupervised Machine Learning for Adaptive Immune Receptors with immuneML. Specifically, it includes all the analysis specification files for immuneML and the immuneML output. The analyses are: simulation of the fully labeled TRBbeta dataset with known ground truth using LIgO training of generative models on the simulated dataset (LSTM, VAE, PWM) visualization analysis of generated sequences summary analysis of "true motifs" - how many of the generated sequences from each of the models contain the ground truth simulated motifs that were the starting point supervised classification of generated sequences against test data (from the simulated dataset produced in the first point, but not used for training). For full information on how to reproduce this result, see the GitHub repository: https://github.com/immuneML/immuneML-unsupervisedML-usecases. immuneML GitHub repository: https://github.com/uio-bmi/immuneML; The analyses are compatible with the latest immuneML version at the time of submission (3.0.21).

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.