Paolo Magni

This study benchmarks generative AI tools for summarizing genomic variant literature, finding VarChat and Gpt-4o perform best with expert validation needed for reliability.

PhenoXtract extracts phenotypes from clinical descriptions using a hybrid approach of Large Language Models and Knowledge Graph embeddings, achieving high recall and precision.

This paper engineers Escherichia coli to display elastin-like polypeptides, enabling self-assembling living materials with controllable flocculation and ethanol production.

OncoVI is a Python-based tool that automates oncogenicity classification of somatic variants using guidelines, achieving 80% accuracy on test data and 79% agreement with expert assessments.

This paper develops a fully-automated workflow to streamline NGS-based analyses in Molecular Tumour Boards by integrating variant interpretation and biomarker estimation.

diVas interprets digenic variants using hypothesis-driven machine learning and XAI, achieving 73% sensitivity in prioritizing causative gene pairs across 11 cases.

This study uses CRISPR interference to reduce antibiotic resistance in E. coli by targeting resistance genes without causing mutations.

The paper evaluates variant prioritization methods for rare disease diagnosis using data from the Rare Genomes Project, finding that top models identify causal variants in most solved families and uncover potential candidates in unsolved ones.

This paper evaluates CRISPR interference as low burden logic inverters in synthetic circuits, demonstrating reduced metabolic load through in-vivo assays.

This paper introduces a dCas9 regulator that reduces competition among sgRNAs in CRISPRi circuits, enabling independent gene repression control.