Multi-omic meta-analysis identifies functional signatures of airway microbiome in chronic obstructive pulmonary disease

Abstract

The interaction between airway microbiome and host in chronic obstructive pulmonary disease (COPD) is poorly understood. Here we used a multi-omic meta-analysis approach to characterize the functional signature of airway microbiome in COPD. We retrieved all public COPD sputum microbiome datasets, totaling 1640 samples from 16S rRNA gene datasets and 26 samples from metagenomic datasets from across the world. We identified microbial taxonomic shifts using random effect meta-analysis and established a global classifier for COPD using 12 microbial genera. We inferred the metabolic potentials for the airway microbiome, established their molecular links to host targets, and explored their effects in a separate meta-analysis on 1340 public human airway transcriptome samples for COPD. 29.6% of differentially expressed human pathways were predicted to be targeted by microbiome metabolism. For inferred metabolite-host interactions, the flux of disease-modifying metabolites as predicted from host transcriptome was generally concordant with their predicted metabolic turnover in microbiome, suggesting a synergistic response between microbiome and host in COPD. The metaanalysis results were further validated by a pilot multi-omic study on 18 COPD patients and 10 controls, in which airway metagenome, metabolome, and host transcriptome were simultaneously characterized. 69.9% of the proposed "microbiomemetabolite-host" interaction links were validated in the independent multi-omic data. Butyrate, homocysteine, and palmitate were the microbial metabolites showing strongest interactions with COPD-associated host genes. Our meta-analysis uncovered functional properties of airway microbiome that interacted with COPD host gene signatures, and demonstrated the possibility of leveraging public multi-omic data to interrogate disease biology.

Microbiota Emergencies in the Diagnosis of Lung Diseases: A Meta-Analysis

“…Similarly, Neisseria , Staphylococcus , and Dialister showed a higher level in lung cancerous lesions than in normal lung tissues ( Ran et al., 2020 ). However, different from the present study, one report found that Proteobacteria , Actinobacteria , and Firmicutes predominantly promoted the development of COPD by contributing the biosynthesis of palmitate, homocysteine, and urate ( Wang et al., 2020 ). Although several novel microbiology biomarkers had been described as having a diagnostic role in lung diseases, it still needs to be further confirmed by larger-scale clinical studies.…”

The composition of lung microbiome in lung cancer: a systematic review and meta-analysis

“…Unlike prior studies, which had reported decreased Bacteroidetes , their meta-analysis did not show any statistically significant difference in Bacteroidetes phylum between IBD subjects and healthy controls. In addition, considering inconsistencies in the findings of previous microbiome studies regarding the interaction between airway microbiome and host in COPD, Wang et al [ 28 ] analyzed COPD sputum sample microbiome using a total of 15 metagenomic datasets adopting a multi-omic meta-analysis approach. To identify taxonomic alterations in the airway microbiome in COPD versus controls, they limited their meta-analysis by combining the results across two 16S rRNA gene datasets due to the availability of two datasets with the case-control design.…”

A meta-analysis study of the robustness and universality of gut microbiome-metabolome associations

“…1A, Additional file 1: Table S1). Notably, these studies spanned various ages, geographies, health conditions, metagenomics/metabolomics platforms, and 16S rRNA gene hypervariable regions, all of which are expected to introduce heterogeneity between datasets, as demonstrated in previous microbiome and metabolome meta-analysis studies in various fields [44][45][46]. Importantly, in casecontrol studies, we treated healthy and disease subgroups separately (considering only study groups with ≥ 40 samples) to avoid the confounding impact of the disease state on both the microbiome and metabolome compositions.…”

“…Microbiome data was processed to obtain genus-level profiles, providing more comparable taxonomic profiles across 16S rRNA gene sequencing and whole genome shotgun sequencing (WGSS) datasets at the expense of sensitivity and resolution (Fig. 1B), as done in several recent microbiome-related meta-analysis studies [44,47]. Specifically, when possible, 16S rRNA gene sequencing raw data were re-processed using QIIME2 [48] to obtain genus-level relative abundances and MetaPhlAn2 tables were collapsed to genus-level profiles (see full details in Additional file 1: Table S2).…”