Manage Abstracts
View, approve, edit, or delete Abstracts using the table below.
Entry ID
228
Entry Date
September 14, 2022 4:01 pm
Reference Id
WgPQTLz1hlAaemoZ
Approve Entries
CFDR Reviewer
Document Status
Published
Click to Edit
Name
Email
Language
English
Year
2018
Category
Presentation Preference
Title
Introduction of a web-based application to generate personalized nutrition recommendations based on metabolomic, proteomic and genetic data.
Names, Organizational Affiliations, and Locations of all Authors
Introduction/Purpose
Analysis of molecular measures, i.e. genetics, metabolomics, and proteomics, have enabled a more preventive and personalized approach to healthcare. Personalized healthcare may mitigate disease risk through early diagnosis and the potential for personalized lifestyle, e.g. nutrition, interventions based on molecular data. Advanced tools combining and visualizing multiple molecular data and their association with dietary intake are required to generate personalized nutrition (PN) recommendations.
Objective(s)/Process or Summary of Content
We aimed to develop a web-based application for generating PN recommendations based on genetic, metabolomic, and proteomic data.
Method(s)/Systemic Approach Used
We employed text mining tools (Ovid for Medline, PolySearch2 for PubMed) to search the scientific literature on associations between nutrient intake or dietary patterns and genetic variants (>100,000), or blood levels of metabolites (>150) or proteins (>200) in human studies alone. Nutrients were linked to specific foods using the Canadian Nutrient File. Machine learning algorithms were developed for ranking based on scientific evidence, nutrient content and across multiple molecular measures. We used interactive data visualization strategies to visualize PN recommendations for healthcare professionals and their clients.
Results/Conclusions
We curated knowledge databases on the associations between abnormal levels of molecular measures and nutrient intakes or dietary patterns, e.g. omega-3 fatty acids or Mediterranean diet. Employing machine learning ranked food items across abnormal levels of multiple molecular measures and grouped them into food categories, to efficiently determine PN recommendations with the potentially highest benefit. Our custom webbased application displayed links between abnormal molecular measures and nutrient intakes or dietary patterns for easy communication of PN recommendations. The smart-phone tracking app allowed clients to easily see PN recommendations and record their actions to monitor and capture adherence longitudinally.
Conclusions(s)/Recommendations
We developed a user-friendly, web-based tool to determine PN recommendations, based on metabolomic, proteomic and genetic data, which may be most beneficial to a client.
Significance to Dietetics
Our tool may aid dietitians in providing PN recommendations based on multiple molecular data.
Funded by
Original Work
Transfer of Copyright
Co-Authors Review and Notification
Newsletter
Delete Entry Link
