Integration of Machine Learning Algorithms and Single-Cell Sequencing Analysis Reveals the Efferocytosis-Related Molecular Subtype and Prognostic Scoring Index in Colon Adenocarcinoma.

Abstract

Background: Colon adenocarcinoma (COAD) is a leading cause of cancer-related mortality, with limited therapies for advanced stages. Efferocytosis, the clearance of apoptotic cells, modulates tumor immunity and progression. We investigated efferocytosis-related genes (ERRGs) in COAD through multiomics integration.

Methods: We analyzed multiomics data from public databases to identify differentially expressed ERRGs and their molecular subtypes. An ERRG score index was developed using integrated machine learning algorithms to evaluate its predictive capacity. Single-cell sequencing and in vitro functional assays were performed to validate key findings.

Results: Among 162 ERRGs, 22 were dysregulated in COAD. Three molecular subtypes exhibited distinct prognoses, immune profiles, and therapy responses. The ERRG score system accurately predicted clinical outcomes, with low scores correlating with improved survival and sensitivity to certain drugs. Single-cell analysis highlighted TIMP1 as a key regulator, confirmed by its knockdown suppressing tumor proliferation and migration in vitro.

Conclusion: ERRGs demonstrate prognostic and therapeutic relevance in COAD, providing insights into molecular subtyping and immunotherapy prediction. TIMP1 emerges as a potential therapeutic target, warranting further clinical validation.