Background <p>Liver metastasis is the leading cause of mortality in colorectal cancer (CRC) patients, and there is an urgent need for biomarkers to improve the diagnosis and prognosis of CRC. However, the molecular mechanisms driving metastatic progression remain incompletely understood, necessitating the discovery of novel therapeutic targets.</p> Methods <p>GEO datasets and machine learning algorithms (Lasso regression, SVM, and RF) were employed to screen for upregulated genes in CRC liver metastasis. Protein-protein interaction (PPI) networks were constructed using STRING. Single-cell RNA sequencing (scRNA-seq) and ATAC-seq were performed to analyze immune infiltration and chromatin accessibility in metastatic tissues. Functional enrichment analysis of RNA-seq data was conducted to explore DAPK1-related pathways. Finally, CRISPR/Cas9-mediated knockout of DAPK1 in LoVo cells was used to validate its role in invasion and migration.</p> Results <p>In this study, we identified DAPK1 as a metastasis-specific upregulated gene in CRC. scRNA-seq of CRC metastases highlighted a significant correlation between DAPK1 expression and macrophage infiltration. Furthermore, ATAC-seq analysis demonstrated increased chromatin accessibility of DAPK1 in CRC liver metastasis. CRISPR/Cas9-mediated knockout of DAPK1 in LoVo cells markedly suppressed their invasion and migration.</p> Conclusions <p>By combining machine learning and multi-omics bioinformatics approaches, we identified DAPK1 as a novel biomarker for CRC liver metastasis, which has potential implications for prognosis and therapeutic targeting. These findings highlight DAPK1 as a critical driver of metastatic progression and provide a theoretical foundation for the treatment of CRC.</p> Graphical Abstract <p></p>