Using network analysis and machine learning to identify genes implicated in Spinal Muscular Atrophy

Nofal, Eslam Mahmoud Ibrahim Imam; Hafez, Elsayed Elsayed; Haikal, Amira Yassein Mohamed; Elhosseini, Mostafa Abdelkhalik

doi:10.21608/jcbr.2022.173808.1283

Using network analysis and machine learning to identify genes implicated in Spinal Muscular Atrophy

Document Type : Original Article

Authors

¹ Egyptian Armed Forces

² Institute, City Of Scientific Research and Technological Applications (SRTA)

³ Head of Computers and Control System dept./ Engineering/ Mansoura University

⁴ Mansoura University – Faculty of Engineering

10.21608/jcbr.2022.173808.1283

Abstract

Background: Spinal Muscular Atrophy (SMA) is a genetic disease that causes the loss of a survival motor neuron (SMN), leading to vital muscle atrophy. Aim: Despite numerous studies to find a cure for this disease, the best of these treatments is still suffering from some limitations and difficulties. It was found that treatments that focus on just one gene are not usually effective. Consequently, the current study investigates gene impacts and interactions by gathering an appropriate microarray dataset for various human SMA instances. In addition, embryonic stem cell samples, which are anticipated to play a significant role in the future treatment of the majority of incurable diseases. Materials and Methods: By using linear models for microarray data analysis (LIMMA), highly differentially expressed genes (DEG) were identified. Then, cluster these genes into modules using machine learning and weighted gene co-expression network analysis (WGCNA) algorithms. Results: By using the preservation methods, the foundation of interesting modules was evaluated between the collected cases. Moreover, the results of previous studies on SMN1, SMN2, NAIP, DYNC1H1, and PLS3 genes have proved that they are direct causes or modifiers of SMA disease severity. However, the change in the expression of these genes did not come at the forefront of the changed genes, which is the exact opposite of what is expected. Accordingly, other interesting modules were determined here as highly correlated modules with these genes. These modules’ genes were imported into Cytoscape for generating SMA networks, and finding their hub genes. Conclusion: These genes can be used as key genes for better analysis, diagnosis, and therapy development, such as BCL2, Cntn1, TYRP1, N4Bp2, and PFDN2.

Keywords