Session Time: 1:45pm-3:15pm
Location: Agora 3 West, Level 3
Objective: We applied IBM Watson for Drug Discovery (WDD), a cognitive computing platform that uses natural language processing, to identify compounds with the potential to enhance Parkin-mediated mitophagy.
Background: Repurposing compounds with regulatory approval is an attractive method to accelerate the availability of therapeutic options for Parkinson’s disease (PD). However, identifying drugs with suitable efficacy and prioritizing for development represents a major challenge. WDD uses natural language processing and machine learning to extract domain specific text features from published literature and infer new connections between drugs of interest. We used WDD to analyze published abstracts of a set of training compounds known to induce mitophagy and applied machine learning to rank a set of candidate drugs according to similarity of linguistic context.
Method: 7 training compounds able to induce mitophagy in experimental paradigms were identified via literature search. 3231 candidate drugs were filtered from the entire Drugbank database (www.drugbank.ca). WDD analyzed ~1.3 million Medline abstracts and ranked all candidate drugs based on semantic similarity to the training compounds. Leave-one-out cross-validation demonstrated a strong predictive power of training entities over each other compared to candidates (area under the Receive-Operator Characteristics curve 0.9513). The top 87/3231 candidates were prioritized for validation studies in cell and fly based models of mitophagy.
Results: 79 candidates were screened in a high throughput in vitro assay that quantifies Parkin recruitment to mitochondria following induction of mitophagy with Carbonyl cyanide m-chlorophenyl hydrazone. 3 hits were identified that significantly increased Parkin recruitment to the damaged mitochondrial membrane and subsequent mitophagy. In vitro testing of the 3 hit compounds has yielded a single drug that provided rescue of locomotor dysfunction in Drosophila with paraquat-induced mitochondrial dysfunction.
Conclusion: We provide proof of concept that based on natural language processing and machine learning, IBM WDD can predict novel treatments to enhance mitophagy pathways associated with PD. Ongoing studies will determine the potential for repurposing our lead compound to treat mitochondrial dysfunction in PD.
References: Acknowledgements The authors would like to acknowledge: the Ontario Brain Institute and the Government of Ontario for providing access and training to the IBM Watson Drug Discovery platform.
To cite this abstract in AMA style:N. Visanji, N. Moskal, G. Shi, A. Lacoste, S. Spangler, P. Lewis, A. Mcquibban. Using cognitive computing to identify existing drugs with potential to stabilize mitophagy pathways associated with Parkinson’s Disease [abstract]. Mov Disord. 2019; 34 (suppl 2). https://www.mdsabstracts.org/abstract/using-cognitive-computing-to-identify-existing-drugs-with-potential-to-stabilize-mitophagy-pathways-associated-with-parkinsons-disease/. Accessed December 7, 2023.
« Back to 2019 International Congress
MDS Abstracts - https://www.mdsabstracts.org/abstract/using-cognitive-computing-to-identify-existing-drugs-with-potential-to-stabilize-mitophagy-pathways-associated-with-parkinsons-disease/