Edi Prifti

CurateMake is a workflow for harmonizing and validating ITS reference databases, improving taxonomic coherence and error correction through nomenclatural standardization and phylogenetic validation.

CroCoDeEL detects and quantifies cross-sample contamination in metagenomic data without negative controls or prior knowledge, showing high accuracy even at low contamination rates.

MetagenBERT uses Transformer models to embed DNA reads for disease classification, achieving comparable performance to existing methods for cirrhosis and better results for type 2 diabetes.

The paper develops a polygenic risk score for baseline QTc and shows it predicts sotalol-induced QT prolongation in a cohort of healthy subjects.

This paper reviews deep learning methods in metagenomics, focusing on their application to human gut microbiome analysis and their potential to improve disease prediction and patient stratification.

This paper introduces metagenome2vec, an end-to-end deep learning approach that classifies diseases from raw metagenomic reads, achieving performance comparable to conventional pipelines.

The paper develops interpretable metagenomics prediction models based on microbial taxon abundance ratios, additions, and subtractions, achieving performance comparable to or better than traditional machine learning methods.