This project used dataset from The Cancer Genome Atlas Network (TCGA) with 462 CRC samples and 33379 gene expression features. We aim to evaluate the effectiveness of specific machine-learning models, namely K-Nearest Neighbors (KNN), Random Forest, and Support Vector Machines (SVM) in predicting MSI status and identify the genes that plays a critical role in enhancing the predictive accuracy of the model for MSI status in colorectal cancer.