Comprehensive Tumor Immune Landscape Analysis

Gene ontology (GO) and Kyoto encyclopaedia of genes and genomes (KEGG) analyses were performed to view the differential potential bio function pathways between cluster 1 and 2. GO, and KEGG analysis was conducted by employing gene set variation analysis (GSVA) and gene set enrichment analysis (GSEA) with the R package ‘GSVA’²⁹ and ‘clusterProfiler’,³⁰ respectively.
Two types of immunogram were introduced to evaluate tumour immunogenicity, as previously reported.³¹, ³² Pathways including T cell immunity, absence of inhibitory molecules, absence of checkpoint expression, absence of inhibitory cells, recognition of tumour cells, trafficking and infiltration, priming and activation, innate immunity, interferon‐gamma response, proliferation, and glycolysis were evaluated by using ssGSEA with R package ‘GSVA’.
The gene set of immune escape‐related genes was obtained from the previous study.³³ As for single‐cell RNA‐seq analysis, genes that were not detected in over 80% of all samples were excluded.
The tumour immune landscape was depicted by the ESTIMATE algorithm, CIBERSORT algorithm, and xCELL algorithm. The ESTIMATE algorithm and xCELL algorithm were performed with R package ‘ESTIMATE’³⁴ and ‘xCell’,³⁵ respectively. The CIBEERSORT algorithm was performed as guided in https://cibersortx.stanford.edu/.³⁶, ³⁷R package ‘CellChat’ was introduced to predict cell–cell communication based on ligand‐receptor pairs between tumour cells and immunocytes.³⁸ Significant differential ligand‐receptors pairs between cluster 1 and cluster 2 were selected.