Análisis de microarrays para la expresión diferencial del gen LRRK2 asociado a la enfermedad de Parkinson

Abstract

Neurodegenerative diseases are in a constant state of evolution; however, many of them have not been fully understood, leading to limited knowledge about their implications on the organism. This is notably observed in the case of Parkinson's disease. In this context, bioinformatics tools emerge as a hopeful avenue for understanding the biological functions affected by the disease. This is achieved through the application of algorithms such as differential expression analysis or machine learning for early telemonitoring. Data extracted from neural stem cells (NSCs) derived from embryonic stem cells (hESCs) carrying a pathogenic LRRK2 (G2019S) mutation were used for Omics data analysis. Functional enrichment of the samples allows for the determination of the transcriptomics of the mutation. Regarding the telemonitoring method, three machine learning algorithms were employed, utilizing a dataset with 52 variables and 252 observations, corresponding to voice samples from patients with and without Parkinson's disease. This is crucial as the disease tends to manifest dysphonia symptoms shortly after contraction. A total of 139 genes linked to this mutation and consequently to Parkinson's disease and its development were identified. The majority of these genes are present in cerebral processes, particularly in nerve impulses and structural components of the central nervous system. The developed predictive method boasts an accuracy level of 0.81 through the application of a robust and scientifically impactful machine learning algorithm known as Random Forest.