NEMoE: A nutrition aware regularized mixture of experts model addressing diet-cohort heterogeneity of gut microbiota in Parkinson’s disease

Abstract

The microbiome plays a fundamental role in human health and diet is one of the strongest modulators of the gut microbiome. However, interactions between microbiota and host health are complex and diverse. Understanding the interplay between diet, the microbiome and health state could enable the design of personalized intervention strategies and improve the health and wellbeing of affected individuals. A common approach to this is to divide the study population into smaller cohorts based on dietary preferences in the hope of identifying specific microbial signatures. However, classification of patients based solely on diet is unlikely to reflect the microbiome-host health relationship or the taxonomic microbiome makeup. To this end, we present a novel approach, the Nutrition-Ecotype Mixture of Experts (NEMoE) model, for establishing associations between gut microbiota and health state that accounts for diet-specific cohort variability using a regularized mixture of experts model framework with an integrated parameter sharing strategy to ensure data driven diet-cohort identification consistency across taxonomic levels. The success of our approach was demonstrated through a series of simulation studies, in which NEMoE showed robustness with regard to parameter selection and varying degrees of data heterogeneity. Further application to real-world microbiome data from a Parkinson’s disease cohort revealed that NEMoE is capable of not only improving predictive performance for Parkinson’s Disease but also for identifying diet-specific microbiome markers of disease. Our results indicate that NEMoE can be used to uncover diet-specific relationships between nutritional-ecotype and patient health and to contextualize precision nutrition for different diseases.

Nutritional Intake and Gut Microbiome Composition Predict Parkinson’s Disease

“…These inferences require validation in a larger PD GM study and/or meta-analysis. The approach of using the global microbiome signature paired with important clinical measures to identify disease has been developed into a methodological approach pairing nutritional intake with gut enterotypes to define health state ( Xu et al, 2021 ).…”

The Gut Microbiome in Parkinson’s Disease: A Longitudinal Study of the Impacts on Disease Progression and the Use of Device-Assisted Therapies

“…This was further supported by the newly developed two-stage model, which showed that the incorporation of nutritional data, specifically protein contribution to total energy intake, was able to stratify the cohort into a subpopulation where the association between GM and PD progression was improved. The interpretability of why certain nutritional markers assist in characterizing disease progression is poorly understood and warrants further study ( Xu et al, 2021 ). The recognized preference for increased carbohydrate intake in PD ( Palavra et al, 2021 ) may lead to potential restrictive consumption of protein and its inadvertent contribution to total energy, which could also be further influenced by patient preference to minimize potential levodopa malabsorption, particularly with higher and frequent regiments of levodopa in more advanced disease.…”

Overcoming cohort heterogeneity for the prediction of subclinical cardiovascular disease risk

“… 60 https://cran.r-project.org/web/packages/glmnet/index.html NEMoE Xu et al. 33 https://github.com/SydneyBioX/NEMoE poLCA Linzer et al. 63 https://cran.r-project.org/web/packages/poLCA/index.html rsubgroup Atzmueller et al.…”

“…rMoE was performed using the implementation from the NEMoE package. 33 rMoE comprises a mixture of experts model 66 with added elastic-net penalties to both the gating and experts networks. Clinical variables were used in the gating network and the metabolite ratios were used in the experts network with the penalty regulariser values of 0.08 and 0.012 respectively.…”

“… 26 , 27 Data-driven approaches for uncovering cohort heterogeneity can be divided into unsupervised techniques which ignore the outcome being studied, and supervised techniques, which include information about the outcome. Unsupervised techniques include clustering 10 , 28 , 29 and latent class analysis 30 , 31 of clinical phenotypes; supervised techniques include mixture of experts (MoE) 32 , 33 , 34 , 35 and subgroup discovery algorithms 17 , 36 to subdivide large cohorts. Following stratification of individuals and identification of heterogeneity present, techniques including hierarchical modeling of subcohorts and ensemble learning can then be employed to improve the prediction of CAD in the whole cohort.…”