Approaches to the Algorithm of Analyzing the Results of Laboratory Testing of Micro- and Macronutrient Content of Bakery Products: Part 1

Abstract

Introduction: Data on the chemical composition of food products are important for solving many problems in medical and social spheres. The development of mechanisms for updating existing databases of the chemical composition of foodstuffs, including the need to change approaches to obtaining primary data and develop algorithms of their processing, is in demand. Objective: To develop an algorithm of obtaining statistically correct values of average concentrations and variability of the main micro– and macronutrients in bakery products. Materials and methods: To develop and test the algorithm, we used the results of testing bakery products obtained in 2020 within the Federal Project on Public Health Strengthening by the laboratories of the Russian Federal Service for Surveillance on Consumer Rights Protection and Human Wellbeing (Rospotrebnadzor). Results: A good separating power was demonstrated by k-means clustering into two groups by the fat content. An algorithm for generalization of data obtained from different laboratories is proposed due to impossibility to assess the whole aggregate of potential errors related to testing, laboratory personnel, data entry, etc. To assess the effectiveness of each stage and the algorithm as a whole, we used the value of the deviation of the resulting variability from the initial one. As a result of processing, this indicator ranged from 5 % for the carbohydrate content to 72 % for the fat content. For the contents of carbohydrates, ash, dietary fiber, vitamin B1, sodium and moisture in both clusters, statistically significant differences were obtained between the processed and original data. This result and the comparability of the obtained values of the mean and variability with the reference ones may indicate the correctness of the algorithm. There were no statistically significant differences between the obtained values of fat and protein content, but the consistency of the order of values with the reference ones was also recorded. Conclusion: The developed algorithm made it possible to obtain up-to-date information about the chemical composition of bakery products. Further research should be aimed at testing and, if necessary, adjusting the algorithm for all major food groups.