SparCC co-occurrence networking reveals intracommunity dynamics of the microbiome following colorectal surgery

The intestinal microbiota plays a critical role in post-surgical wound healing following bowel resection; however, perioperative, prophylactic antibiotic administration may deleteriously affect it. We previously used 16S rRNA amplicon sequencing of stool samples to assess perioperative and longitudinal changes in the microbiome through 6 months in patients undergoing (i) colonoscopy after mechanical bowel prep (MBP) alone, (ii) non-resectional colorectal surgery after MBP with oral antibiotics and prophylactic intravenous antibiotics no longer than 24 h post-operative (surgical bowel prep [SBP]), and (iii) resectional colorectal surgery with SBP. Our objective in this study was to investigate the translational utility of SparCC co-occurrence networking to uncover biologically relevant patterns. Network topological parameters and hub species were calculated using NetCoMi, and permutational statistical tests were used to compare parameters. Network similarity among cohorts and time points generally matched changes in beta diversity, except in the resectional cohort, where all networks could not be differentiated statistically. Similarity in centrality measures among hub species was frequently significantly less similar than expected by chance and corresponded to an increased edge density and modularity, suggesting the latter parameters may reflect re-stabilization of the microbiome following surgery. We further noted the infrequently reported genera Enterocloster and Ruthenibacter were hub species during time points associated with surgical recovery, suggesting potentially novel roles for these genera in wound healing. Streptococcus, frequently implicated in surgical site infections at our center, was also frequently positively associated with Blautia throughout all networks, suggesting an increasing abundance of commensal bacteria serves as a prophylactic strategy.IMPORTANCEThis study employs the emerging approach of co-occurrence networking to assess ecological dynamics in the microbiome following colonoscopy and colorectal surgery. We expand upon applications of this approach to determine hub species and investigate clinically translational interpretations of network topological parameters in the context of recovery across three different trajectories of perturbation. Our results provide a context in which to interpret these network parameters biologically and represent a foundational step in beginning to quantitatively leverage network-based approaches to study microbial ecology. Furthermore, we identify network hub taxa that may play previously unexplored roles in wound healing.