Connecting the salivary microbiome and blood epigenome with a machine learning interaction approach in children with moderate-to-severe asthma

<bold>Introduction:</bold> Microbial dysbiosis and epigenetic changes are both associated with asthma severity and asthma control in children. We aimed to discover interactions between the salivary microbiome and blood epigenome in children with moderate-to-severe asthma in order to better understand host-microbiome interactions associated with asthma control. <bold>Methods:</bold> We performed a cross-sectional analysis of the 16S salivary microbiome and blood epigenome data of 116 moderate-to-severe asthmatic children (6-17 years) included in the pan-European SysPharmPediA study. Patients were classified as either uncontrolled or controlled based on exacerbations and the (childhood) asthma control test. Taxa present in at least 20% of samples were selected, while differentially methylated regions (DMRs) were identified in regard to asthma control. The predictive value of each taxon was determined for all DMRs using random forest regression. <bold>Results:</bold> 105 DMRs were identified between uncontrolled and controlled asthmatics and compared with 163 taxa. In addition to spurious interactions between taxa and DMRs, <italic>Streptococcus sp.</italic> and <italic>TM7x</italic> showed significant predictive capabilities for 33 and 37 DMRs respectively. These DMRs contained genes, such as <italic>IL-15</italic>, <italic>WNT1</italic> and <italic>TRIM46</italic>, and were associated with macrophage and neutrophil recruitment, Th2 cell expansion and TNFα production. <bold>Conclusion:</bold> We identified two taxa, <italic>Streptococcus sp.</italic> And <italic>TM7x</italic>, with predictive power for the methylation of DMRs associated with asthma control in children, affecting multiple facets of asthma pathophysiology. Further research is needed to investigate the role of these two taxa in asthma and the epigenome.