Exploring NPM1P51 as a key biomarker for glaucoma by integrating differential gene expression analysis and artificial neural network models.

Glaucoma is a group of optic neuropathic conditions in which the progressive degeneration of retinal ganglion cells (RGC) occurs, ultimately leading to irreversible vision loss. Despite our understanding of pathophysiology, potential biomarkers and early detection are still very limited. Hence, there is an urgent need to find the potent molecular biomarker of Glaucoma so that early diagnosis and personalised glaucoma treatment can be possible. In this study, we performed comparative transcriptomic analyses to discern differentially expressed genes (DEGs) linked with Glaucoma. The aim is to find a robust candidate biomarker to facilitate early diagnosis and characterisation of Glaucoma. Advanced feature selections are performed by employing DEG identification and machine learning algorithms, enabling the selection of highly prognostic relevance genes. Furthermore, Artificial neural network (ANN) models were developed to recognise the complicated nonlinear patterns within high-dimensional data and identify the most promising genes having a potent role in glaucoma progression and risk prediction. After the analysis, NPM1P51 turned out to be the most significant contributor after performing multiple analyses, suggesting it to be a strong pseudogene biomarker. Further, CACNG7 showed moderate results in Garson and Olden plots, but better in Generalised weight plots. It encodes a subunit of the voltage-gated calcium channel complex, aligning with already established roles in the mechanisms of neurodegeneration in glaucoma, particularly in apoptosis of retinal ganglion cells through calcium dysregulation. These results provide a foundation for further experimental validations and therapeutic interventions.