Very early identification of high-risk children before they develop chronic diseases is extremely challenging because of the multitude of contributory genetic, environmental, and lifestyle factors. Recent foundational multiomic studies have begun to pioneer the approaches to identify disease and transitions and intervene early. In a landmark study, Price et al. followed 108 healthy adults for nine months, and collected biological samples (saliva, blood, urine, stool) every three months [26
]. Multiomic profiles were generated, which included whole genome sequencing, 16S rRNA gut microbiome sequencing, 218 clinical diagnostic tests, 262 proteins, and 643 metabolites. The multiomic assays were designed to assess five health domains (cardiovascular, diabetes, inflammation, nutrition and toxins, stress). The genome sequencing data was summarized into polygenic risk scores for 127 disease traits. These risk scores are a single variable that estimates the genome-wide risk for a given trait by summing the number of risk alleles for each individual, weighted by effect size estimates from large genome-wide association studies (GWAS) [27
]. The data were integrated by constructing an interomic correlation network, which captures pairwise interrelationships between the five omic layers. The α-diversity (species richness) of the gut microbiome was positively correlated with height and β-nerve growth factor levels, and negatively correlated with levels of CSF-1, IL-8, and FLT3 ligand. Clinical diagnostic tests identified deviations from wellness or test results outside of normal reference ranges [28
]. These insights were then leveraged to suggest evidence-based changes to diet (including supplements) and lifestyle (exercise, stress management) that resulted in significant improvements to biomarker levels across multiple health domains—type 2 diabetes (fasting glucose, HbA1c levels, insulin), cardiovascular disease (total cholesterol, LDL cholesterol), inflammation (IL-8, TNF), and toxins (mercury).
Chronic obstructive pulmonary disorder (COPD) is a highly heterogeneous, chronic, inflammatory lung disease which has early life origins in a subset of patients [29
]. In a proof-of-concept study, Li et al. evaluated the utility of multiomic data to differentiate between COPD patients, healthy non-smokers, and smokers with normal lung function [5
]. They found that the mean accuracy of subgroup prediction (healthy, smoker, COPD) was extremely poor when each of the omic data blocks were analyzed in isolation. However, combining data from multiple omic platforms increased the mean prediction accuracy to 100%, even when group sizes were limited to small numbers. These analyses highlight the potential for multiomic approaches to dramatically improve our understanding of highly complex and heterogeneous inflammatory diseases.
Very early identification of at-risk individuals from birth is now theoretically possible with polygenic risk scores [30
]. These risk scores combine information derived from variants across the entire genome [30
], and are able to identify segments of the population which are at heightened risk (more than three-fold) for a range of complex traits including inflammatory diseases [31
]. One caveat of polygenic risk scores is that information is collapsed across the genome without taking into account the cellular or biological context. To address this issue, cluster analysis of deep immunological and clinical data can be utilized to stratify subjects into distinct developmental trajectories [32
], and the polygenic risk scores can be leveraged to find those clusters enriched for subjects at heightened genetic risk. Extending this approach to multiomic layers will undoubtedly increase the resolution of these analyses and further refine the critical windows of opportunity for early intervention.