In the Journals

Gut microbiome marker linked to IBS severity

Researchers have identified an intestinal microbiota signature associated with severity of symptoms in irritable bowel syndrome.

Using a “machine learning approach allowed identification of a gut microbial signature for IBS severity, which also could be reproduced in a validation cohort,” the researchers wrote. “Importantly, because of its relatively low sensitivity, this microbiota signature cannot be used as a clinical predictor of IBS severity, but as a way to explore relevant features ... which deserve to be explored in future IBS microbiota studies.”

To better understand the role of the gut microbiome in IBS symptoms, investigators obtained fecal and mucosal samples from adults with IBS at an outpatient clinical in Sweden, and from healthy controls.

They used 16S ribosomal RNA targeted pyrosequencing to analyze 232 fecal samples and 59 mucosal biopsy samples from 110 IBS patients and 39 healthy controls in an exploratory set.

Then they analyzed 46 fecal samples from 29 IBS patients and 17 healthy controls in a validation set.

They also measured exhaled hydrogen (H2) and methane (CH4), oro-anal transit time, and severity of both psychological and GI symptoms in all participants, used quantitative polymerase chain reaction to measure fecal methanogens, and used numerical ecology analyses and a machine learning procedure — a “computational statistical technique” — to evaluate the data.

First, they found that while structurally distinct, fecal and mucosal microbiota strongly correlated with each other (P < .001).

Classic ecologic descriptive approaches showed no differences in microbial richness or variability between controls, IBS patients, IBS subtypes and patients stratified by symptom severity. However, the machine learning procedure allowed the researchers “to decrease the [operational taxonomic unit (OTU)] complexity by combining them into a consensus microbial signature from an ensemble of classifiers that discriminated patients with severe IBS from patients with moderate or mild IBS and healthy subjects.” The signature included 90 of 2,911 total bacterial OTUs.

They confirmed this microbial signature was associated with IBS severity in the validation set, with an AUC of 0.64, which suggested “the OTUs identified as a microbial signature for IBS severity were robust.”

Finally, by using this signature, the investigators determined that severity of IBS symptoms was negatively associated with microbial richness, exhaled CH4, methanogen presence and Clostridiales- or Prevotella-enriched enterotypes.

Diet and medications did not appear to significantly influence the microbial signature, the researchers noted.

“Our study highlights the heterogeneity of IBS patients, and the difficulty in stratifying patients based on a microbiota profile when using only classic ecologic approaches,” the researchers concluded. “The use of machine learning has allowed us to circumvent the issues related to large microbial data sets and to better explore the microbial data. We were able to identify several interesting links between gut microbiota and the clinical profile.” – by Adam Leitenberger

Disclosures: Several researchers report they are employees of Danone Research. Please see the full study for a list of all researchers’ relevant financial disclosures.

Researchers have identified an intestinal microbiota signature associated with severity of symptoms in irritable bowel syndrome.

Using a “machine learning approach allowed identification of a gut microbial signature for IBS severity, which also could be reproduced in a validation cohort,” the researchers wrote. “Importantly, because of its relatively low sensitivity, this microbiota signature cannot be used as a clinical predictor of IBS severity, but as a way to explore relevant features ... which deserve to be explored in future IBS microbiota studies.”

To better understand the role of the gut microbiome in IBS symptoms, investigators obtained fecal and mucosal samples from adults with IBS at an outpatient clinical in Sweden, and from healthy controls.

They used 16S ribosomal RNA targeted pyrosequencing to analyze 232 fecal samples and 59 mucosal biopsy samples from 110 IBS patients and 39 healthy controls in an exploratory set.

Then they analyzed 46 fecal samples from 29 IBS patients and 17 healthy controls in a validation set.

They also measured exhaled hydrogen (H2) and methane (CH4), oro-anal transit time, and severity of both psychological and GI symptoms in all participants, used quantitative polymerase chain reaction to measure fecal methanogens, and used numerical ecology analyses and a machine learning procedure — a “computational statistical technique” — to evaluate the data.

First, they found that while structurally distinct, fecal and mucosal microbiota strongly correlated with each other (P < .001).

Classic ecologic descriptive approaches showed no differences in microbial richness or variability between controls, IBS patients, IBS subtypes and patients stratified by symptom severity. However, the machine learning procedure allowed the researchers “to decrease the [operational taxonomic unit (OTU)] complexity by combining them into a consensus microbial signature from an ensemble of classifiers that discriminated patients with severe IBS from patients with moderate or mild IBS and healthy subjects.” The signature included 90 of 2,911 total bacterial OTUs.

They confirmed this microbial signature was associated with IBS severity in the validation set, with an AUC of 0.64, which suggested “the OTUs identified as a microbial signature for IBS severity were robust.”

Finally, by using this signature, the investigators determined that severity of IBS symptoms was negatively associated with microbial richness, exhaled CH4, methanogen presence and Clostridiales- or Prevotella-enriched enterotypes.

Diet and medications did not appear to significantly influence the microbial signature, the researchers noted.

“Our study highlights the heterogeneity of IBS patients, and the difficulty in stratifying patients based on a microbiota profile when using only classic ecologic approaches,” the researchers concluded. “The use of machine learning has allowed us to circumvent the issues related to large microbial data sets and to better explore the microbial data. We were able to identify several interesting links between gut microbiota and the clinical profile.” – by Adam Leitenberger

Disclosures: Several researchers report they are employees of Danone Research. Please see the full study for a list of all researchers’ relevant financial disclosures.