THE FIRST SYSTEMS BIOLOGY APPROACH TO THE MICROBIOME OF SIBO BY HIGH-LEVEL CO-OCCURRENCE COMMUNITY NETWORK ANALYSES IDENTIFIES TWO DISTINCT PROTEOBACTERIA-RELATED CLUSTERS
DDW ePoster Library. Pimentel M. 05/21/21; 319393; Fr243
Abstract
Discussion Forum (0)
Engage with the presenter here during ePoster Session: Diarrheal Disorders: Bacterial Overgrowth - Drug Induced and Other Enterocolitides (Microscopic, Enteropathy, Check Point Inhibitors, etc.)
On Friday, May 21, 2021 12:15 - 1 p.m. EDT

Number: Fr243
THE FIRST SYSTEMS BIOLOGY APPROACH TO THE MICROBIOME OF SIBO BY HIGH-LEVEL CO-OCCURRENCE COMMUNITY NETWORK ANALYSES IDENTIFIES TWO DISTINCT PROTEOBACTERIA-RELATED CLUSTERS

Society: AGA
Track: Stomach and Small Bowel Disorders
Category: Basic & Clinical Intestinal Disorders

Author(s): Leopoldo Valiente-Banuet1, Gabriela Leite1, Ruchi Mathur1, Ali Rezaie1, Mark Pimentel11 Cedars-Sinai Medical Center, Los Angeles, California, United States

Introduction: Small intestine Bacterial Overgrowth (SIBO) is characterized by an overabundance of bacteria in the small bowel, defined as >103 CFU/mL coliforms from a duodenal aspirate(North American Consensus). Recent studies revealed SIBO was driven by shifts in proteobacteria and that the genera Klebsiella and Escherichia had a pronounced effect on the small intestinal microbiome. Here, a systems biology approach was used to define whether co-occurrence networks could determine characteristics of SIBO and if they could predict SIBO in an independent dataset. Method: SIBO subjects used for network co-occurrence analysis were identified from the REIMAGINE study. REIMAGINE is an ongoing prospective collection of duodenal aspirates in patients undergoing upper endoscopy using validated sequencing methodologies for small intestine microbiome sequencing (Leite, et al). For comparison, non-SIBO subjects matched for BMI, gender, age and 16S rRNA sequencing depth >10,000 were identified [n=32 per condition]. The co-occurrence inference method corresponded to a probabilistic framework for presence – absence datasets through which the associations are positive, negative, or non-associated, and the inferences were obtained from permutations of sample sets for different sample sizes taken from folds of the data. This scheme allowed to validate and obtain the combination of parameters that best represented both conditions. In the networks, a node corresponded to a genus, an edge to the association between genera, and a triple to the simultaneous association of three genera. After identifying co-occurrence characteristics of SIBO vs non-SIBO, the ability to predict SIBO based on the findings were applied to an independent set of 21 SIBO samples. Results: All triples along folds and simulations corresponded to network subunits in stable configuration with three positive edges, or 1 positive edge and two negative edges. Based on this structural result and focusing on the triples with 1 positive and 2 negative associations for the SIBO and non-SIBO representation, consistent clusters of genera between the two groups were identified. Examples of these triples are shown in Figure 1 for Selenomonas, Streptococcus, and Ruminococcaceae_gen in non-SIBO (A-C), and (D) for Escherichia in SIBO. Table 1 shows identified clusters of genera for both conditions. In SIBO, one cluster was dominated by Campylobacter and the other by proteobacteria like Escherichia, Erwinia, Salmonella, and Trabulsiella. Using all triple-based associations of the microbiome on the external validation set, SIBO was predicted with a >80% accuracy. Conclusion: In this first systems biology approach to SIBO using co-occurrence networks, there are 2 Proteobacteria genera clusters related to SIBO. The characteristics of the microbiome associations are predictive of SIBO in new cases with high accuracy.
Engage with the presenter here during ePoster Session: Diarrheal Disorders: Bacterial Overgrowth - Drug Induced and Other Enterocolitides (Microscopic, Enteropathy, Check Point Inhibitors, etc.)
On Friday, May 21, 2021 12:15 - 1 p.m. EDT

Number: Fr243
THE FIRST SYSTEMS BIOLOGY APPROACH TO THE MICROBIOME OF SIBO BY HIGH-LEVEL CO-OCCURRENCE COMMUNITY NETWORK ANALYSES IDENTIFIES TWO DISTINCT PROTEOBACTERIA-RELATED CLUSTERS

Society: AGA
Track: Stomach and Small Bowel Disorders
Category: Basic & Clinical Intestinal Disorders

Author(s): Leopoldo Valiente-Banuet1, Gabriela Leite1, Ruchi Mathur1, Ali Rezaie1, Mark Pimentel11 Cedars-Sinai Medical Center, Los Angeles, California, United States

Introduction: Small intestine Bacterial Overgrowth (SIBO) is characterized by an overabundance of bacteria in the small bowel, defined as >103 CFU/mL coliforms from a duodenal aspirate(North American Consensus). Recent studies revealed SIBO was driven by shifts in proteobacteria and that the genera Klebsiella and Escherichia had a pronounced effect on the small intestinal microbiome. Here, a systems biology approach was used to define whether co-occurrence networks could determine characteristics of SIBO and if they could predict SIBO in an independent dataset. Method: SIBO subjects used for network co-occurrence analysis were identified from the REIMAGINE study. REIMAGINE is an ongoing prospective collection of duodenal aspirates in patients undergoing upper endoscopy using validated sequencing methodologies for small intestine microbiome sequencing (Leite, et al). For comparison, non-SIBO subjects matched for BMI, gender, age and 16S rRNA sequencing depth >10,000 were identified [n=32 per condition]. The co-occurrence inference method corresponded to a probabilistic framework for presence – absence datasets through which the associations are positive, negative, or non-associated, and the inferences were obtained from permutations of sample sets for different sample sizes taken from folds of the data. This scheme allowed to validate and obtain the combination of parameters that best represented both conditions. In the networks, a node corresponded to a genus, an edge to the association between genera, and a triple to the simultaneous association of three genera. After identifying co-occurrence characteristics of SIBO vs non-SIBO, the ability to predict SIBO based on the findings were applied to an independent set of 21 SIBO samples. Results: All triples along folds and simulations corresponded to network subunits in stable configuration with three positive edges, or 1 positive edge and two negative edges. Based on this structural result and focusing on the triples with 1 positive and 2 negative associations for the SIBO and non-SIBO representation, consistent clusters of genera between the two groups were identified. Examples of these triples are shown in Figure 1 for Selenomonas, Streptococcus, and Ruminococcaceae_gen in non-SIBO (A-C), and (D) for Escherichia in SIBO. Table 1 shows identified clusters of genera for both conditions. In SIBO, one cluster was dominated by Campylobacter and the other by proteobacteria like Escherichia, Erwinia, Salmonella, and Trabulsiella. Using all triple-based associations of the microbiome on the external validation set, SIBO was predicted with a >80% accuracy. Conclusion: In this first systems biology approach to SIBO using co-occurrence networks, there are 2 Proteobacteria genera clusters related to SIBO. The characteristics of the microbiome associations are predictive of SIBO in new cases with high accuracy.
Code of conduct/disclaimer available in General Terms & Conditions

By clicking “Accept Terms & all Cookies” or by continuing to browse, you agree to the storing of third-party cookies on your device to enhance your user experience and agree to the user terms and conditions of this learning management system (LMS).

Cookie Settings
Accept Terms & all Cookies