Anaplastic thyroid cancer, or ATC, is characterized by highly malignant tumors and has historically been associated with the process of epithelial-mesenchymal transition (EMT). Despite this association, key proteins that support this relationship have not yet been identified. This study aimed to establish a connection between specific target proteins and the progression of EMT in ATC by use of differentially expressed gene analysis (DEG), weighted gene co-expression network (WGCNA), and machine learning models (LASSO, SVM-RFE). The Gene Expression Omnibus (GEO) and Molecular Signatures Database (MSigDB) were used to obtain genomic data from ATC patients and identify EMT-related genes respectively, while machine learning models were used to identify key genes within the obtained data. The WT1 protein was isolated and found to promote tumor growth in nude mice, suggesting its role as a key protein within EMT in ATC. Upon WT1 knockdown in the mice models, tumor growth was shown to be stunted, with volume and weight also being affected. Statistical analysis consistently yielded statistically significant p-values less than 0.05.