Exploratory integrated analysis of the microbiome and host transcriptome reveals correlations between the intratumoral microbiota and esophageal squamous cell carcinoma

Background The intratumoral microbiota can affect the occurrence, development, and prognosis of esophageal squamous cell carcinoma (ESCC) through various mechanisms. This study explored the relationship between key genes and key microorganisms in ESCC. Methods A systematic multiomic integration pipeline was applied to analyze the transcriptomic and 16S rRNA sequencing data of cancerous and paracancerous tissues obtained from patients with ESCC. Differential expression analysis, protein–protein interaction networks, and machine learning algorithms were successively used to identify key host genes in ESCC. Notably, gene set enrichment analysis (GSEA) and immune infiltration analysis were employed to investigate relevant mechanisms. Subsequently, differential abundance analysis and machine learning were used to elucidate key intratumoral microorganisms. Finally, a combined multiomic analysis using the Data Integration Analysis for Biomarker discovery using Latent cOmponents (DIABLO) framework and Spearman correlation networks (CIRCOS) was employed to explore synergistic host–microbe associations. Results TNNC2 , CTHRC1 , MMP10 , CGAS , IL18 , SERPINE1 , CCNA1 , and POSTN were identified as key genes. GSEA revealed that key genes were mainly involved in the cell cycle, indicating that cell cycle disorders might affect the progression of ESCC. The enrichment scores of seven types of immune cells significantly changed during ESCC progression. Additionally, 14 differentially abundant microorganisms were detected, and four key microorganisms were identified: Lachnoanaerobaculum , Fusobacterium , Actinomyces , and Capnocytophaga . Among them, Lachnoanaerobaculum was associated with multiple key genes, including POSTN , MMP10 , CTHRC1 , and CGAS . Exploratory DIABLO analysis highlighted concordant variation of the selected transcriptomic and microbiomic features between ESCC and adjacent nontumor tissue tissues. Conclusion Eight host genes and four microbial genera were identified as candidate ESCC-associated features. Given the small cohort and lack of independent validation, these findings should be considered preliminary and confirmation in larger independent cohorts is required.