Exploring Co-occurrence patterns and microbial diversity in the lung microbiome of patients with non-small cell lung cancer

Abstract

Abstract Background It has been demonstrated in the literature that a dysbiotic microbiome could have a negative impact on the host immune system and promote disease onset or exacerbation. Co-occurrence networks have been widely adopted to identify biomarkers and keystone taxa in the pathogenesis of microbiome-related diseases. Despite the promising results that network-driven approaches have led to in various human diseases, there is a dearth of research pertaining to key taxa that contribute to the pathogenesis of lung cancer. Therefore, our primary goal in this study is to explore co-existing relationships among members of the lung microbial community and any potential gained or lost interactions in lung cancer. Results Using integrative and network-based approaches, we integrated four studies assessing the microbiome of lung biopsies of cancer patients. Differential abundance analyses showed that several bacterial taxa are different between tumor and tumor-adjacent normal tissues (FDR adjusted p-value < 0.05). Four, fifteen, and twelve significantly different associations were found at phylum, family, and genus levels. Diversity analyses suggested reduced alpha diversity in the tumor microbiome. However, beta diversity analysis did not show any discernible pattern between groups. In addition, four distinct modules of bacterial families were detected by the DBSCAN clustering method. Finally, in the co-occurrence network context, Actinobacteria, Firmicutes, Bacteroidetes, and Chloroflexi at the phylum level and Bifidobacterium, Massilia, Sphingobacterium, and Ochrobactrum at the genus level showed the highest degree of rewiring. Conclusions Despite the absence of statistically significant differences in the relative abundance of certain taxa between groups, it is imperative not to overlook them for further exploration. This is because they may hold pivotal central roles in the broader network of bacterial taxa (e.g., Bifidobacterium and Massilia). These findings emphasize the importance of a network analysis approach for studying the lung microbiome since it could facilitate identifying key microbial taxa in lung cancer pathogenesis. Relying exclusively on differentially abundant taxa may not be enough to fully grasp the complex interplay between lung cancer and the microbiome. Therefore, a network-based approach can offer deeper insights and a more comprehensive understanding of the underlying mechanisms.

Keywords

• Lung cancer
• Lung microbiome
• 16S rRNA gene sequencing
• Microbiome diversity
• Co-occurrence network
• Microbiome data integration