Computational Analysis to Elucidate Parkinson's Disease Pathophysiology and Therapeutics

Parkinson's disease (PD) is a deadly disorder of the brain characterized by the deterioration of dopamine engendering neurons in the nervous system. The onset of PD is marked with a myriad of severe motor and non-motor symptoms, typically evident during advanced stages of the disorder. This underscores the urgent need for identification of druggable novel blood-based biomarkers for early detection and as potential therapeutic targets. In this study, we conducted a comprehensive analysis of blood-based transcriptome in both PD and healthy samples, aiming to identify novel transcriptomic biomarkers. Utilizing a blood-based microarray dataset from the NCBI's Gene Expression Omnibus, we pinpointed genes that are differentially expressed among Parkinson's and healthy control samples. Subsequently, we identified hub genes from these DEGs and assessed their significance through machine learning approach, employing their expression signatures as input features. Our machine learning analyses, utilizing artificial neural network and random forest algorithms, demonstrated that the expression signatures of these hub genes can effectively differentiate between PD and healthy conditions with accuracies of 71.42% and 78.57%, respectively. Finally, employing a molecular docking approach, we sought to identify novel natural compounds with potential therapeutic efficacy in PD treatment. In summary, our study contributes insights into blood-based biomarkers and potential treatment avenues for PD.