Biology subjects

Magni, P.

Publications and source records attributed to Magni, P..

7 recordsLinked to original sources

bioRxiv2026

PhenoXtract: combining Large Language Model and Knowledge Graph embedding to extract phenotypes from clinical descriptions

MotivationStandardized phenotypic descriptions are essential for accurate diagnosis, yet clinicians and researchers face challenges in manually extracting and mapping phenotypes from scientific literature or patient clinical records to the Human Phenotype Ontology. Recent advances in deep learning offer new opportunities for automation. We developed PhenoXtract, a novel phenotype extraction approach that combines Large Language Models and Knowledge Graph embedding. PhenoXtract is a multistep pipeline that takes clinical descriptions as input, extracts candidate phenotype entities using large language models, and maps them to terms from an enriched version of the Human Phenotype Ontology, processed as a knowledge graph. ResultsEvaluation against expert-curated ground-truth datasets show a recall of 0.70 and precision of 0.85 for PhenoXtract, demonstrating concordance with manually extracted phenotypes, with a computation time of 10-20 seconds for each text analyzed. Moreover, PhenoXtract surpasses rule-based and deep learning-based state-of-the-art tools in two out of the three ground-truth datasets evaluated. These results suggest that hybrid approaches combining Large Language Models and Knowledge Graph embeddings represent a promising direction for automated clinical phenotyping at scale. Contactsberardelli@engenome.com

genomics↗

bioRxiv2025

Benchmarking generative AI tools for literature retrieval and summarization in genomic variant interpretation.

BackgroundGenerative AI is increasingly used to extract structured information across domains, but its reliability in academic and clinical research, where precision and accuracy are essential, remains largely unexplored. This study evaluates the ability of Large Language Models (LLMs)-based algorithms to generate accurate, literature-based summaries of human genomic variants, with a focus on real-world usability. ResultsWe benchmarked five open-access generative AI platforms--ChatGPT, MistralAI, VarChat, Perplexity, and ScholarAI--across 40 curated variants equally divided between somatic and germline settings. For each variant, summary reports were generated and blindly evaluated by domain experts using five defined metrics. VarChat emerged as the top-ranked tool, showing the highest summarization accuracy, citation relevance, and robustness against hallucinations. Gpt-4o consistently ranked second, showing particularly stable robustness in conditions where the literature was scarce. Perplexity and ScholarAI, despite being literature-focused, ranked lowest across most metrics. Tool performance was strongly influenced by the availability of peer-reviewed literature, confirming that current generative models remain sensitive to data scarcity. ConclusionsOur findings highlight the heterogeneity of current generative AI tools in genomic variant interpretation workflows. While some platforms already provide useful outputs, reliable integration into basic and clinical research requires expert validation and domain-related fine-tuning. This work provides for the first time a curated benchmark for assessing LLM-generated content in variant genomics and underscores the need for caution when using these tools to support variant interpretation.

bioinformatics↗

bioRxiv2025

Engineered Living Materials for Bioproduction Downstream Processing

De novo engineered living materials (ELMs) are cellular systems that self-assemble into macroscopic structures through genetically encoded interactions, offering a route to programmable materials grown directly from living cells. Despite their promise, the molecular design principles that enable scalable self-assembly in de novo ELMs remain poorly understood. Here, we engineer Escherichia coli to display elastin-like polypeptides (ELPs) on the outer membrane, transforming single cells into self-assembling living materials through weak intermolecular interactions. By tuning the polarity of ELP sequences, we generate assemblies spanning micrometer- to centimeter-length scales that sediment within a few hours while preserving cellular metabolic activity. We demonstrate the portability of this platform across genetic backgrounds and inducible expression systems, and deploy it in an ethanologenic E. coli strain as a proof of principle. In small-scale fermentation settings, ELP-based ELMs enable controllable flocculation, reduce filtration time by more than threefold, and maintain ethanol production performance comparable to that of the parental strain. Together, this work establishes ELP surface display as a modular strategy for constructing de novo engineered living materials and defines initial genetic design rules linking molecular-scale interactions to emergent macroscopic organization.

synthetic biology↗

bioRxiv2024

Harnessing CRISPR interference to re-sensitize laboratory strains and clinical isolates to last resort antibiotics

The global race against antimicrobial resistance requires novel antimicrobials that are not only effective in killing specific bacteria, but also minimize the emergence of new resistances. Recently, CRISPR/Cas-based antimicrobials were proposed to address killing specificity with encouraging results. However, the emergence of target sequence mutations triggered by Cas-cleavage was identified as an escape strategy, posing the risk of generating new antibiotic-resistance gene (ARG) variants. Here, we evaluated an antibiotic re-sensitization strategy based on CRISPR interference (CRISPRi), which inhibits gene expression without damaging target DNA. The resistance to four antibiotics, including last resort drugs, was significantly reduced by individual and multi-gene targeting of ARGs in low- to high-copy numbers in recombinant E. coli. Escaper analysis confirmed the absence of mutations in target sequence, corroborating the harmless role of CRISPRi in the selection of new resistances. E. coli clinical isolates carrying ARGs of severe clinical concern were then used to test the robustness of CRISPRi under different growth conditions. Meropenem, colistin and cefotaxime susceptibility was successfully increased in terms of MIC (up to >4-fold) and growth delay (up to 11-hours) in a medium-dependent fashion. To our knowledge, this is the first demonstration of CRISPRi-mediated re-sensitization to last-resort drugs in clinical isolates. This study laid the foundations for further leveraging CRISPRi as antimicrobial agent or research tool to selectively repress ARGs and investigate resistance mechanisms.

synthetic biology↗

bioRxiv2023

Digenic variant interpretation with hypothesis-driven explainable AI

MotivationThe digenic inheritance hypothesis holds the potential to enhance diagnostic yield in rare diseases. Computational approaches capable of accurately interpreting and prioritizing digenic combinations based on the probands phenotypic profiles and familial information can provide valuable assistance to clinicians during the diagnostic process. ResultsWe have developed diVas, a hypothesis-driven machine learning approach that can effectively interpret genomic variants across different gene pairs. DiVas demonstrates strong performance both in classifying and prioritizing causative pairs, consistently placing them within the top positions across 11 real cases (achieving 73% sensitivity and a median ranking of 3). Additionally, diVas exploits Explainable Artificial Intelligence (XAI) to dissect the digenic disease mechanism for predicted positive pairs. Availability and ImplementationPrediction results of the diVas method on a high-confidence, comprehensive, manually curated dataset of known digenic combinations are available at oliver.engenome.com.

bioinformatics↗

bioRxiv2020

dCas9 regulator to neutralize competition in CRISPRi circuits

CRISPRi-mediated gene repression allows simultaneous control of many genes. However, despite highly specific sgRNA-promoter binding, multiple sgRNAs still interfere with one another by competing for dCas9. We created a dCas9 regulator that adjusts dCas9 concentration based on sgRNAs demand, mitigating competition in CRISPRi-based logic gates. The regulators performance is demonstrated on both single-stage and layered CRISPRi logic gates and in two common E. coli strains. When a competitor sgRNA causes between two and ~25 fold-change in a logic gates input/output response without dCas9 regulator, the response is essentially unchanged when the regulator is used. The dCas9 regulator thus enables concurrent and independent operation of multiple sgRNAs, thereby supporting independent control of multiple genes.

synthetic biology↗

bioRxiv2020

CRISPR interference as low burden logic inverters in synthetic circuits: characterization and tuning

The rational design of complex biological systems through the interconnection of single functional building blocks is hampered by many unpredictability sources; this is mainly due to the tangled context-dependency behavior of those parts once placed into an intrinsically complex living system. Among others, the finite amount of translational resources in prokaryotic cells leads to load effects in heterologous protein expression. As a result, hidden interactions among protein synthesis rates arise, leading to unexpected and counterintuitive behaviors. To face this issue in rational design of synthetic circuits in bacterial cells, CRISPR interference is here evaluated as genetic logic inverters with low translational resource usage, compared with traditional transcriptional regulators. This system has been studied and characterized in several circuit configurations. Each module composing the circuit architecture has been optimized in order to meet the desired specifications, and its reduced metabolic load has been eventually demonstrated via in-vivo assays.

synthetic biology↗