ReGenAI

Drug repositioning for Retinitis Pigmentosa by an artificial intelligence application to single-cell transcriptomics.

IT EN

The Challenge & Our Mission

The Challenge

Retinitis Pigmentosa (RP) is a group of inherited retinal diseases affecting approximately 1 in 4,000 people. It causes progressive photoreceptor cell loss, leading to irreversible vision impairment and often complete blindness. Despite decades of research, there is still no generalized drug therapy that can halt or reverse the disease's progression.

Our Mission

The ReGenAI project proposes an innovative approach to address this challenge. Our mission is to leverage cutting-edge artificial intelligence and single-cell genomics to identify existing, approved drugs that can be repurposed to treat Retinitis Pigmentosa, with the goal of significantly accelerating the journey from scientific discovery to clinical application.

A Three-Phase Strategy

1

Decoding the Disease

We generate a high-resolution "genetic fingerprint" of RP using single-cell RNA sequencing on the RD10 mouse model.

2

AI-Powered Discovery

Our AI model screens thousands of drugs from the LINCS database to find molecules that can "reverse" the disease fingerprint.

3

Pre-clinical Validation

The top drug candidates are tested on RD10 mice to confirm their efficacy in improving retinal function and structure.

How Our AI Works

Input 1: Disease Data

scRNAseq profiles from RD10 mouse retinas

Input 2: Drug Data

LINCS database of drug-induced gene expression

Adversarial AI Model

The model uses a shared encoder and an adversarial classifier to learn the true biological signature of RP, ignoring dataset-specific noise, and predicts which drugs can reverse it.

Meet the Team

This project is driven by a synergistic collaboration between two highly specialized research units, led by principal investigators with complementary expertise.

Biomedical Unit — UNINA

This unit conducts the core biological experiments of the project. Leveraging its state-of-the-art Genomic and Animal Facilities (TIGEM), the team manages the animal models, generates all the single-cell transcriptomic data, and performs the in-vivo validation of the drug candidates identified by the AI, using advanced functional and morphological assays (ERG, OCT).

Computational Unit — UNISANNIO

This unit manages the project's computational research. With access to a high-performance Superdome server, the team develops and trains the sophisticated AI model. They are responsible for analyzing the complex single-cell data, prioritizing drug candidates through large-scale virtual screening, and performing the final bioinformatic analysis of treated samples.

Principal Investigators

Prof. Michele Pinelli

Associate Professor of Medical Genetics

As the project's PI, Prof. Pinelli coordinates the biomedical and computational efforts. His research focuses on the transcriptomics of the human retina and the diagnosis of rare genetic diseases. His expertise is fundamental for generating high-quality biological data and ensuring the clinical relevance of the project's findings.

Prof. Francesco Napolitano

Associate Professor of Bioinformatics

Prof. Napolitano leads the development of the project's computational strategy. As a Computer Scientist, his work centers on creating innovative machine learning and deep learning models for computational drug discovery. He develops the novel AI tools and analytical pipelines that are essential for identifying therapeutic candidates from complex genomic data.

Latest News

AI Identifies 6 Potential Drug Candidates

The ReGenAI computational model has screened thousands of compounds and identified a shortlist of 6 molecules with high potential to reverse the Retinitis Pigmentosa signature at a cellular level. These candidates are now moving to in vitro testing to validate their efficacy.

Advanced AI Model for Drug Prioritization is Ready

After months of refinement, the advanced version of our AI model is now complete. It uses a domain adaptation approach to compare the disease signature with a large database of drug-induced genetic profiles, enhancing its predictive power.

Foundational AI Model Development is Complete

The computational team at UNISANNIO has successfully developed the foundational version of our deep learning model. It can process scRNAseq data and learn the core features of the disease signature.

Research Team is Complete

We are pleased to announce that the ReGenAI research team is now fully assembled. The research fellows for both biomedical and computational units have been successfully recruited.

Key Milestone: Single-Cell Data Generation Complete

A major milestone achieved: our biomedical unit at UNINA has completed the generation of high-resolution single-cell RNA sequencing data from RD10 mouse retinas. This dataset forms the foundation for AI model training.

Project ReGenAI Kick-off

We are thrilled to announce the official start of the ReGenAI project, funded by the Italian Ministry of University and Research under PRIN 2022. The project unites experts in genetics, computational biology, and AI to discover repurposable drugs for Retinitis Pigmentosa.

Funding & Support

Italian Ministry of University and Research (MUR)

Funded under the PRIN 2022 program (Prot. 2022ENM9AZ)

European Union — NextGenerationEU

Supported through the Italian PNRR program