Bulk and single-cell RNA-sequencing analyses along with machine learning methods identify macrophage signatures in colorectal cancer

Tumour-associated macrophages (TAM) are an important component of the tumour microenvironment (TME) and stimulate colorectal cancer (CRC) cells in different ways. Bulk and single-cell RNA sequencing (scRNA-seq) data on CRC were downloaded from multiple public databases to dissect the TME of CRC at high resolution. The Scissor algorithm was used to correlate the bulk and scRNA-seq data with genomic and phenotypic data. TAM were divided into Scissor + and Scissor – subtypes. The characteristic genes of TAM subsets closely related to prognosis were identified. The Scissor + genetic signature was applied for risk stratification of patients with CRC. Machine learning techniques were employed to determine the most accurate TAM gene set to develop a TAM-related risk signature for CRC. Finally, colony formation assay, scratch assay and invasion assay were performed to elucidate the biological function of the key gene SPP1 within this prognostic signature. The Scissor + subgroup of TAM was strongly associated with a poor prognosis. Using multiple machine learning algorithms, we identified nine TAM signature genes associated with CRC progression. TAM signature genes were linked to clinicopathological features to construct a nomogram that provided a personalised tool for evaluating prognosis. This study analysed the potential association between TAM and CRC prognosis, and developed a prognostic model associated with TAM that is expected to become an important tool for the management of patients with CRC.