Applying Machine Learning Algorithms to Analyze the Impact of Climate Change on Exosome Dysregulation and Their Role in Cancer Progression: A Multi-Dimensional Optimization Approach
Year: 2026
Type: article
Abstract: Climate change is considered to be a severe health issue on an international scale, and its impacts on cellular and molecular functions are being studied more extensively. Exosomes (nanoscale vesicles that mediate intercellular communication) are claimed to be dysregulated during environmental stressors and are a key contributor to cancer development. The limitations of the previous studies reside in the fact that study of climate variables is frequently performed separately, exosome dysregulation as a process is not well mechanistically understood, and machine learning methods are not applied to combine heterogeneous climate, molecular, and clinical data. The primary value associated with this study is the creation of a multi-dimensional machine learning model that correlates climate stressors, exosome biology, and cancer progression. Climate data (temperature, CO2, AQI and UV), exosomal biomarkers (mRNA, miRNA and proteins) and clinical outcomes are integrated and optimized through LASSO and dimensionality reduction techniques. The findings indicate natural distributions of environmental factors (mean temperature = 19.90 °C, mean CO2 = 379.95 ppm, mean AQI = 95.48, mean UV = 5.19). There are strong correlations between miRNA and mRNA exosomal biomarkers and temperature (r = 0.953; r = 0.894 respectively). Predictive models are very accurate (logistic regression = 0.94, SVM = 0.93, random forest = 0.90). Based on the conditions investigated, scenario analysis suggests that the progression probability is 1.00 at 1.5°C. The outcomes indicate that in the data set analyzed, there is sensitivity of cancer progression prediction to climate-related stressors.
Institutions: Thi Qar University, Al-Ayen Iraqi University
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: gold
OpenAlex ID: W7212374524