IN SILICO APPROACH TOWARDS PARKINSON'S DISEASE PATHOPHYSIOLOGY, DRUG REPURPOSING AND POST TRANSLATIONAL MODIFICATIONS

Abstract

Parkinson’s disease (PD) is an enervating, debilitating and lethal neurodegenerative disorder marked by deterioration of neurons which produces dopamine in the central nervous system. PD is accompanied by a constellation of lethal motor and non-motor symptoms which are observed when the disease is progressed at an advanced stage. Hence, there is a great necessity of novel blood-based biomarkers which can help in early detection of the disease and can serve

as new therapeutic targets to impede the progression of disease. Herein, we performed blood- based differential gene expression analysis of Parkinson’s samples and healthy samples to look

for novel blood-based gene biomarkers and their target drugs. Herein firstly, we downloaded blood-based microarray gene expression omnibus (GEO) dataset to explore differentially expressed genes (DEGs) in PD samples compared to healthy control samples. We found 18 DEGs between PD and healthy samples, most of which were novel genes for PD. Further, we validated these DEGs via machine learning algorithms using their expression signature as input features. Validation with algorithms such as Artificial neural networks, Decision trees, Random Forest, Linear discriminant analysis and kernel principal component analysis (PCA) models resulted in accuracy of 92.8%, 78.5%, 92.8%, 100%, 92.8% respectively. Moreover, using hierarchical clustering based unsupervised machine learning we found that PD and healthy samples were well differentiated in two separate clusters. Furthermore, we used LINCS L1000 based drug repurposing search engine L1000CDS2

, and CoDReS tool to look for exemplar repurposed drugs which can reverse the expression of our obtained genes, thereby we obtained several drugs with neuroprotective properties. In addition, we looked for novel transcription factors regulating the dysregulation of genes targeted by the shortlisted drugs. Further, using in silico tools we found various post translational modifications involved in drug-gene pathway. Lastly, we searched for common drugs with can target both PD pathogenesis and ageing.