Network Medicine Framework for Identifying Drug Repurposing Opportunities for COVID-19 (2004.07229v2)

Abstract: The current pandemic has highlighted the need for methodologies that can quickly and reliably prioritize clinically approved compounds for their potential effectiveness for SARS-CoV-2 infections. In the past decade, network medicine has developed and validated multiple predictive algorithms for drug repurposing, exploiting the sub-cellular network-based relationship between a drug's targets and disease genes. Here, we deployed algorithms relying on artificial intelligence, network diffusion, and network proximity, tasking each of them to rank 6,340 drugs for their expected efficacy against SARS-CoV-2. To test the predictions, we used as ground truth 918 drugs that had been experimentally screened in VeroE6 cells, and the list of drugs under clinical trial, that capture the medical community's assessment of drugs with potential COVID-19 efficacy. We find that while most algorithms offer predictive power for these ground truth data, no single method offers consistently reliable outcomes across all datasets and metrics. This prompted us to develop a multimodal approach that fuses the predictions of all algorithms, showing that a consensus among the different predictive methods consistently exceeds the performance of the best individual pipelines. We find that 76 of the 77 drugs that successfully reduced viral infection do not bind the proteins targeted by SARS-CoV-2, indicating that these drugs rely on network-based actions that cannot be identified using docking-based strategies. These advances offer a methodological pathway to identify repurposable drugs for future pathogens and neglected diseases underserved by the costs and extended timeline of de novo drug development.

• The paper demonstrates a novel network medicine framework using AI and graph neural networks to predict drug-disease interactions for COVID-19.
• It integrates network diffusion and proximity metrics into a consensus multi-algorithm approach that outperforms individual models.
• The study screened 918 approved drugs, identifying 76 candidates with potential antiviral effects by targeting host interactions.

Network Medicine Framework for Identifying Drug Repurposing Opportunities for COVID-19

The paper presents a sophisticated methodology leveraging network medicine to identify drug repurposing opportunities against SARS-CoV-2, a crucial endeavor given the urgent need for effective treatments amidst the COVID-19 pandemic. The approach integrates advanced predictive algorithms based on artificial intelligence, network diffusion, and network proximity to rank a substantial number of clinically approved drugs for their potential efficacy against the COVID-19 pathogen.

Methodological Advancements

The core of this research rests on three diverse predictive approaches. The first employs artificial intelligence techniques within a graph neural network framework to process multimodal biological graphs. It effectively seeks potential new edges in these graphs, indicative of promising drug-disease interactions. The second, network diffusion, calculates similarities by evaluating the diffusion state distance vectors between drug targets and viral proteins. The third, network proximity, quantifies the shortest path lengths between the drug target and viral proteins, thus estimating how network proximity influences drug efficacy.

A notable realization from this paper is that no single predictive algorithm consistently excels across all evaluation metrics and datasets. This outcome led to the development of a consensus or multimodal approach aggregating the outputs from the diverse pipelines. This strategy demonstrated better predictive performance than any individual model, underscoring the hypothesis that pooling insights from multiple methodologies can harness complementary information.

Results and Implications

An extensive computational screen evaluated 918 drugs with known interactions and gathered clinical trial data. The paper found 76 effective drugs that did not directly interact with SARS-CoV-2 proteins, suggesting that these drugs exert their antiviral effects by influencing the host's subcellular interactions. These insights proffer significant implications for drug repurposing strategies; traditional docking-based approaches that focus primarily on direct interactions might overlook potentially efficacious treatments.

The consensus approach, specifically the CRank algorithm, showed superior performance in ranking potential anti-SARS-CoV-2 drugs. The algorithm manages drug listing prioritized through a Bayesian strategy that accounts for their consensus across the predictive pipelines, achieving notably higher hit rates in experimental validations compared to individual methods.

Potential Applications and Future Directions

This research opens promising avenues for identifying potential treatments, not only for COVID-19 but for other emerging and neglected diseases as well. The network medicine framework provides a scalable and adaptable template which could significantly democratize access to effective therapies through drug repurposing, circumventing the protracted timelines of traditional drug development.

In potential future directions, the field could benefit from employing enhanced network models that more accurately reflect the dynamic and multifaceted nature of protein interactions. Additionally, expanding this framework to multicellular and patient-derived experimental models may improve the translation of computational predictions to clinical efficacy. The evolving landscape of AI could also integrate richer datasets extending beyond the confines of current databases, incorporating real-world evidence such as genomics and proteomics data, further improving the precision of such predictive frameworks.

In conclusion, the insights provided are invaluable to the domain of network pharmacology and offer a robust platform for accelerated therapeutic discovery. Despite inherent challenges and the NP-hard nature of optimizing such complex systems, the methodological innovations and findings encourage further exploration into how network-based approaches can diversify and enhance biomedical research, fostering a more expedient pipeline from computational predictions to clinical applications.