Nutrient Estimation from 24-Hour Food Recalls Using Machine Learning and Database Mapping: A Case Study with Lactose

Abstract

The Automated Self-Administered 24-Hour Dietary Assessment Tool (ASA24) is a free dietary recall system that outputs fewer nutrients than the Nutrition Data System for Research (NDSR). NDSR uses the Nutrition Coordinating Center (NCC) Food and Nutrient Database, both of which require a license. Manual lookup of ASA24 foods into NDSR is time-consuming but currently the only way to acquire NCC-exclusive nutrients. Using lactose as an example, we evaluated machine learning and database matching methods to estimate this NCC-exclusive nutrient from ASA24 reports. ASA24-reported foods were manually looked up into NDSR to obtain lactose estimates and split into training (n = 378) and test (n = 189) datasets. Nine machine learning models were developed to predict lactose from the nutrients common between ASA24 and the NCC database. Database matching algorithms were developed to match NCC foods to an ASA24 food using only nutrients ("Nutrient-Only") or the nutrient and food descriptions ("Nutrient + Text"). For both methods, the lactose values were compared to the manual curation. Among machine learning models, the XGB-Regressor model performed best on held-out test data (R 2 = 0.33). For the database matching method, Nutrient + Text matching yielded the best lactose estimates (R 2 = 0.76), a vast improvement over the status quo of no estimate. These results suggest that computational methods can successfully estimate an NCC-exclusive nutrient for foods reported in ASA24. . Both databases have comparable nutrient completeness, with FNDDS at 100% and NCC at 92-100% completeness, but the databases differ in the number of nutrients reported. The ASA24 output includes 65 nutrients [5] and licensed 2018 NCC Database files include 166 nutrients and food components. Sixty-two nutrients are shared between the ASA24 output and NCC database. The NCC database also outputs nutrients and food components such as lactose, soluble and insoluble fiber, sugar alcohols, and individual amino acids while ASA24 does not [6]. While both ASA24 and NDSR/NCC database have widespread use and contain thousands of foods, there is no unique identifier to match each food on a one-to-one basis. Only a small number of foods in ASA24 have an exact known counterpart in NCC. Manual lookup of foods reported in ASA24 into NDSR based on text descriptions and nutrient profiles is time-consuming but is currently the only method to obtain values of NDSR-exclusive nutrients. This presents a major hurdle for investigating nutrient intake when using ASA24 if the research question requires assessment of a nutrient absent in the underlying database.Our research group is investigating a series of questions that require assessment of a nutrient that is not reported in the ASA24 output: lactose. Most adults worldwide are unable to digest lactose, the primary carbohydrate in milk. Some populations, however, are able to digest lactose into adulthood in a heritable trait known as lactase persistence (LP) [7]. LP genotypes may influence dairy and more specifically, lactose, ...

Rhamnan sulphate from green algae Monostroma nitidum improves constipation with gut microbiome alteration in double-blind placebo-controlled trial

“…Of these, fucoidan and carrageenan are beneficial sulphated polysaccharides, which exhibit anti-obesity properties through altering the composition of gut microbiota 2 . These polysaccharides were extensively studied using high fat diet-induced obese mice and clinical trials with obese population, and these studies proved that their anti-obesity properties were involved in the reduction of Bacteroidetes / Firmicutes ratio 3 , which is a relevant marker of gut dysbiosis in obese population, increased diversity of microbiota 4 , 5 , and increase in specific gut bacteria 5 , 6 .…”

Artificial Intelligence in Nutrients Science Research: A Review

“…Chin et al tested the Automated Self-Administered 24-Hour Dietary Assessment Tool (ASA24) on the example of lactose with regard to the Nutrition Data System for Research (NDSR) [ 46 ]. ASA24, also known as food diaries, is a web-based tool that enables multiple, automatically coded, self-administered 24-h diet recalls.…”

“…An important issue in this research area is the assessment of the reliability and credibility of the test results obtained using AI techniques. Another essential issue is the modification of the dietician–patient relationship in the case of replacing, in whole or in part, the work of a medical professional by AI systems [ 43 , 44 , 45 , 46 , 47 , 48 , 49 , 50 , 51 , 52 , 53 ]. The problem of trust in AI-based systems, especially in the elderly, remains open.…”

An Innovative Machine Learning Approach to Predict the Dietary Fiber Content of Packaged Foods

“…These methods have been applied in many areas including antioxidant protein identification [22], transcriptome subtype classification [23], malnutrition screening [24], and food item recognition [25]. In the context of nutrient prediction, Chin et al (2019) have demonstrated the feasibility of using gradient boosting to predict the lactose content of dietary recall items given known nutrient information [26]. Additionally, Ma et al (2021) have shown the possibility of using neural networks to predict the carbohydrate, protein, and sodium content of branded foods based on their ingredients [27].…”