Cristian Padrón-Manrique

This study uses machine learning and explainable AI to identify metabolomic signatures and biomarkers in COVID-19 and Post-COVID-19 patients, revealing heterogeneous metabolic responses.

mb-PHENIX denoises microbiota data by combining diffusion and supervised UMAP, improving differential abundance analysis in sparse and noisy count matrices.

The paper introduces sc-PHENIX, an imputation method for scRNA-seq data that addresses over-smoothing issues in diffusion-based approaches by using PCA-UMAP initialization.