• Research

Machine learning and systems biology approaches to study the gut microbiome

Developing a machine learning-based analytical pipeline using systems biology approaches to identify microbiome-related features implicated in ageing.

Project summary.

Led by: Korcsmáros Group

Start date: Oct 2018

Duration: Oct 2022

Grants: NPIF-BBSRC-CASE project

BB/S50743X/1

As life expectancy increases, developed countries have an increasingly elderly population. The gut microbiome primes the immune system during early development and contributes to lifelong homeostasis. Several disorders, such as dementia and Inflammatory Bowel Disease (IBD), are associated with gut microbiome alterations. Despite developments in generating microbiome data, the key obstacle remains - which is to extract predictive biomarkers. This problem not only consists of the processing of these large datasets, but also the fact that microbiome studies are often correlative, and therefore lack host-microbiota interaction mechanisms. Furthermore, there are a large number of unknown microbes (mostly commensals) whose health-contributing potential continues to be unclear. This, in turn, means these datasets remain very noisy, thus requiring complex analytical methods.

To address these issues, in partnership with BenevolentAI, this project aims to develop an integrated machine learning-based systems biology workflow, which can be applied to the gut microbiome data for identifying prognostic indicators of healthy ageing and age-related disorders. The approach is based on metagenomics and metatranscriptomics data, which captures the composition and functional potential of the microbiome in modulating host processes.

Details.

With an increase in metagenomic data, a standardised, powerful, robust and scalable method is required to process big datasets. Machine learning can efficiently model microbiome interactions by;

  1. learning novel features (by automatic discovery of “regularities” without relying on a priori knowledge);
  2. capturing multiple features (strains, proteins, pathways, etc) and model these for prediction;
  3. quickly learning complex patterns across large datasets.

Combining machine learning-based features with host-microbiome interactions and systems biology will improve our understanding of how the microbiota contributes to health using already-generated microbiome datasets.

To investigate the dynamic changes in the microbiome, this project will work in collaboration with The Motion Study. The Motion Study is a longitudinal study collecting diverse datasets of 360 elderly people with a special focus on the microbiome, ageing and dementia. To better understand the data collected within this study, the machine learning and systems biology pipeline developed in this project will be used to analyse and uncover discreet relationships within the patient cohort.

Collaborators.

Amir Saffari

BenevolentAI

Simon Carding

Quadram Institute

Impact statement.

This project will identify dynamic changes in the gut microbiome during healthy and unhealthy (disease-state) ageing. The gut microbiota is considered an essential companion of human cells, as microbes have been found to interact with nearly all human cells, therefore dynamic changes in these communities can play a fundamental role in the ageing process.

It follows, therefore, that understanding these complex interplay between the host and the microbial communities within the microbiota is an important step in understanding the ageing process.