Inflammatory bowel disease genomics, transcriptomics, proteomics and metagenomics meet artificial intelligence

Abstract

Various extrinsic and intrinsic factors such as drug exposures, antibiotic treatments, smoking, lifestyle, genetics, immune responses, and the gut microbiome characterize ulcerative colitis and Crohn's disease, collectively called inflammatory bowel disease (IBD). All these factors contribute to the complexity and heterogeneity of the disease etiology and pathogenesis leading to major challenges for the scientific community in improving management, medical treatments, genetic risk, and exposome impact. Understanding the interaction(s) among these factors and their effects on the immune system in IBD patients has prompted advances in multi-omics research, the development of new tools as part of system biology, and more recently, artificial intelligence (AI) approaches. These innovative approaches, supported by the availability of big data and large volumes of digital medical datasets, hold promise in better understanding the natural histories, predictors of disease development, severity, complications and treatment outcomes in complex diseases, providing decision support to doctors, and promising to bring us closer to the realization of the "precision medicine" paradigm. This review aims to provide an overview of current IBD omics based on both individual (genomics, transcriptomics, proteomics, metagenomics) and multi-omics levels, highlighting how AI can facilitate the integration of heterogeneous data to summarize our current understanding of the disease and to identify current gaps in knowledge to inform upcoming research in this field.

Keywords: Crohn's disease; artificial intelligence; deep learning; genes; genetics; inflammatory bowel disease; machine learning; omics; pathogenesis; ulcerative colitis.

FIGURE 1

Typical graphical representation of primary concepts of artificial intelligence and their historical appearance. AI, artificial intelligence; ML, machine learning.

FIGURE 2

Classification of machine learning approaches and their common algorithms.

FIGURE 3

The five steps in the ML workflow: data acquisition, data cleaning, model construction and training, model evaluation, and deployment. During the construction of the model, an optimal ML algorithm is selected based on the training dataset and the problem to resolve. This algorithm allows the model to learn and predict behaviors. After the model is trained, its performance is evaluated to test and validate the model itself. If the predictive outcomes are not satisfactory, the model should be further improved by giving feedback in the step “model construction and training”, which allows adjustments of some parameters/features. In some cases, it may be required to return to the “data acquisition” step to modify the data entering the ML process. ML, machine learning.

FIGURE 4

A possible structure for an IBD neural network consists of a multi‐layer stack: the first layer handles input data from different sources, such as IBD omics elements, endoscopic images, and histological data; one or more hidden layers perform calculations complexes, interactions and combinations of parameters; and finally, an output layer which, after receiving the signals processed by the hidden layers, provides the results. For example, an ANN algorithm could be used to train and instruct an ML model to distinguish IBD patients from non‐IBD subjects (exploratory cohort), and subsequently be tested and validated on a separate cohort to evaluate the accuracy of the model in classifying IBD patients from non‐IBD subjects. IBD, inflammatory bowel disease.

FIGURE 5

Data integration at individual OMICs (genomics, transcriptomics, proteomics, and metabolomics) with artificial intelligence tools could provide multi‐omics analysis to decipher the complex labyrinth of inflammatory bowel disease.

FIGURE 6

Data‐integration and analysis with artificial intelligence in inflammatory bowel disease: a combination of heterogeneous information, including omics data, clinical variables, exposomes, and biometric data.