CancerSP: Cancer Stage Progression

CancerSP is a web-based platform developed for the analysis of high-throughput genomics data and the prediction of cancer stages. The platform utilizes machine learning models trained on TCGA Level 3 genomics data to distinguish between Early and Late stages across six different types of cancer. Web Server: https://webs.iiitd.edu.in/raghava/cancersp/ About the Platform Accurately predicting the stage of cancer is essential for understanding the mechanisms behind metastasis and for selecting appropriate therapeutic strategies. CancerSP focuses on classifying the pathological stage of cancer patients by analyzing gene expression levels (RSEM values). Data Source: Models were trained on genomic profiles of cancer patients from six cancer types available in The Cancer Genome Atlas (TCGA). Methodology: The platform primarily employs Random Forest algorithms to classify patients into early or late stages. Feature Reduction: The system successfully reduced the feature space from approximately 17,000 genes to a signature of less than 100 genes that effectively delineate cancer stages.