Dataset for Phenotyping and Discovery in Systemic Lupus Erythematosus Omani Population (Preprint)

Abstract

Systemic lupus Erythematosus (SLE) is an autoimmune disease characterized by autoantibodies directed against self-antigens, immune complex formation, and immune deregulation, resulting in damage to any organ. Moreover, the natural history of SLE is unpredictable. Therefore there is a need for a data-driven insight for a better understanding of the disease to drive therapeutics strategies.

Our objectives are (a) to explore the prevalence of the symptoms perceived by Omani SLE patients; (b) to identify the symptom clusters occurring in patients with SLE; and (c) to examine the correlation of disease activity index (SLEDAI) and physician global assessment (PGA) with each subgroup

Our method consists of collecting a broad spectrum of data from Sultan Qaboos University Hospital (SQUH). The data included: demographic, clinical, laboratory, and therapy. After its collection, the data goes through several stages—first, data cleaning and feature extraction. The next step was to explore data analysis to identify the data types and explore the data distribution to get a full view and understand the dataset. Then, three clustering methods were used, hierarchical agglomerative clustering, K-means clustering, and spectral clustering to cluster our dataset. Next, the clustering results evaluated using correlation with SLEDAI and PGA

The exploratory data analysis show that joint pain is the most common symptoms in Omani SLE patients, followed by positive anti-dsDNA antibody, low complement (C3, C4), acute cutaneous lupus (ACL), renal disorder, and hemolytic anemia.The clustering analysis results showed two separate patients clusters, mild clus- ter and severe. Patients in the severe cluster have a higher prevalence of the re- nal disorder, hemolytic anemia, anti-dsDNA antibody, and low complements (C3, C4). As a result of analyzing cumulative manifestations and treatment, the se- vere cluster patients suffer from malar rash and proteinuria with higher use of cyclophosphamide, mycophenolate mofetil, and azathioprine. The second cluster is mild disease activity, and it is associated with joint pain, low complements (C3, C4), and a positive anti-dsDNA antibody.

In this research, we explored the hierarchical clustering in order to identify severity clusters in Omani patients with SLE. The results of clustering experiments were validated using two types of cluster validation, which are internal cluster validation and external cluster validation. Two separate patients clusters were identified, severe cluster and mild cluster. The results demonstrated the existing relation between symptoms prevalence associated with disease activity. Our findings showed the symptoms (renal disorder, hemolytic anemia, low complement (C3, C4), and positive anti-dsDNA ) are associated with severe cases of SLE disease activity. These important findings can be helpful for research purposes and patients management.