Fatemeh Farhangnia - Student Research Committee, Tabriz University of Medical Sciences, Tabriz, Iran
Morteza Ghojazadeh - Neuroscience Research Center, Tabriz University of Medical Sciences, Tabriz
Alireza Lotfi - Student Research Committee, Tabriz University of Medical Sciences, Tabriz, Iran
Fatemeh Amirnaseri - Student Research Committee, Tabriz University of Medical Sciences, Tabriz, Iran

Background and aims: Parkinson (PD) is the second-most common neurodegenerative diseaseafter Alzheimer’s disease that is a degenerative condition of the brain associated with motorsymptoms (slow movement, tremor, rigidity and imbalance) and other complications includingcognitive impairment, mental health disorders, sleep disorders and pain and sensory disturbances.Currently, PD has no cure and no early diagnostics methods exist. Mitochondrial dysfunction ispresented in the early stages of PD, and it is considered an important pathophysiology component.The supreme point is that reinforcing a positive compound effect in mitochondrial can occurusing the machine learning model, confirming the platforms for mitochondria-based drug targetinteraction (DTI).Method: A comprehensive systematic search by using the terms such as “Artificial Intelligence”,“machine learning”, “Parkinson’s disease”, “Drug target interaction” as keywords, was conductedin four Online Databases: Web of Science, Scopus, PubMed, and Embase up to February ۲۰۲۳.Also, for screening and data extraction, some applications such as “Rayyan” were used. Reviewsand studies that did not use artificial intelligence for Parkinson’s disease DTI were excluded.Studies that met our inclusion criteria were then critically appraised by two authors independently.Results: We retrieved ۹۵۰ relevant publications from online databases. After a thorough examinationof the titles and abstracts and the removal of duplicate publications (n=۷۳), ۵۰۵ studieswere eliminated. In ۳۸ cases of disagreement between two authors, the opinion of the third authorwas the determiner. The full texts of ninety- four papers were reviewed. Eventually, thirteenstudies met our inclusion criteria and included in our study. About ۶۲ percent of studies used ML(machine learning) algorithm to improve drug target interaction and modeling of new drug targetsfor patients with PD. In some studies, the QSAR model developed with artificial intelligence wasused to identify drug targets.Conclusion: Based on the results of the studies, artificial intelligence approach can be useful inidentifying drug targets and developing them. A logical target in the drug treatment of Parkinson’sdisease is leucine-rich repeat kinase ۲ (LRRK۲). It is also related to the treatment or reductionof symptoms. Structure-based and ligand-based approaches can be used. Using artificial intelligence,QSAR models have been developed with the aim of using them for pharmaceutical purposes.This model can be used in virtual screening to identify inhibitory proteins. There are severalsoftware available for QSAR development that are either commercial or free to use.

Parkinson, Drug development, Artificial Intelligence, Drug target interaction, Deep Learning