Integration of genomic and clinical data for predicting psoriasis onset

Psoriasis, an autoimmune chronic skin disorder, affects 2-3% of the world population and consequently might impose serious inconveniences of physical, psychological, and socioeconomic nature. Early prediction of psoriasis onset is necessary for preventive interventions and improved patient outcomes. This study develops a machine learning model to predict the onset of psoriasis using a random forest classifier and based on genomic markers (e.g., HLA-Cw6, IL12B SNPs) and clinical features (e.g., BMI, stress level). Developed in Python using the scikit-learn library, the model obtains a test accuracy of 50.8% and an AUC-ROC of 0.515 on the synthetic dataset of 5,000 patients with a test size of 500 samples on which 53.4% correspond to no onset, and 46.6% correspond to onset. The methodology, implementation, and performance evaluation have been detailed with seven tables and seven figures explaining everything from performance metrics through classification reports to confusion matrices. The limited results highlight the challenges, such as class balancing and artificial dataset constraints, as well as suggestions for using real data and improved modeling work. This lays the foundation for personalized dermatological care based on integrative data efforts.