Investigation of Heterogeneity of Autoimmune Diseases with Single Cell Analysis and Integration of Machine Learning

Abstract

Autoimmune diseases are diverse conditions with complicated pathologies that show heterogeneity at the cellular level, affecting people of all ages with unidentified trigger factors associated with their development. The autoimmune diseases spread in various organs which pose challenges for diagnosis and treatment. However, more research is necessary, particularly at the cellular level, to identify the primary underlying cause and treat autoimmune diseases. Understanding the fundamental molecular mechanisms of autoimmune is the main problem at hand, with an emphasis on cellular heterogene ity, cell type identification, and gene regulatory dynamics among different autoimmune diseases. Single-cell RNA sequencing is an intriguing technique that has recently trans formed the science of genomics by making it possible to examine single-cell patterns of gene expression. The Single-cell RNA analysis is performed on healthy and diseased samples of juvenile idiopathic arthritis, immune thrombocytopenia, and inflammatory demyelinating disease of the autoimmune system. The analysis provided insightful in formation about cellular heterogeneity at the single-cell level. In the case of juvenile idiopathic arthritis CD4+ T cells and CD8+ T cells are present in greater abundance. In immune thrombocytopenia, the most prominent cell abundance are B cells, dendritic cells, and HSCs followed by erythroid cells. PBMC sample of inflammatory demyelinat ing disease has a greater abundance of B cells, CD8+ T cells, CD4+ T cells, NK cells, and monocytes while, in the CSF sample CD4+ T cells, CD8+ T cells, and monocytes are present in abundance. In the comparison of all diseases at the single-cell level, the common cell clusters are B cells, basophils, CD4+T cells, CD8+T cells, dendritic cells, granulocytes, monocytes, and NK cells. Along with enhanced biological pathways and functional categories connected to those genes, DGE and enrichment analyses also shed light on the frequency of gene expression within various cell populations. Cell-cell communication provides interaction strength within the cell and also with the environ ment. Moreover, based on recognized cell types, machine-learning classification models are constructed to differentiate between different diseases. The accuracy of the XG Boost model is 89%, which is better than the accuracy of the Random Forest, which is 83%. Both models predict the same labels with high count values. These results greatly advance our knowledge of the complex molecular mechanisms behind several autoimmune diseases. In the Future, include specialized and efficient therapies, which will eventually improve the lives of those suffering from these diseases.