 Metastatic propagation is the leading cause of death for most cancers, making prediction and understanding of this process essential for effective treatment. Somatic mutations have been linked to both tumoragesis and metastasis, but less attention has been paid to whether metastatic events can be identified through genomic mutational signatures which are concise descriptions of the mutational processes. To address this issue, researchers developed MetaWise, a deep neural network, DNN, model, by applying mutational signatures as input features calculated from whole exome sequencing, whereas data of TCGA and other metastatic cohorts. The model was able to accurately distinguish between metastatic tumors and their corresponding primary tumors, outperforming traditional machine learning, ML, models and a deep learning, DL, model called Deodor L. Additionally, the model identified several mutational signatures associated with metastatic spread in cancers, such as those related to a pobeck mutagenesis, UV, induced signatures, and DNA damage response deficiencies. This article was authored by Wei Xingsheng, Mengchen Pu, Xiaorong Li, and others. We are article.tv, links in the description below.