CHIP-NMC: An Application for Corn Hybrid and Inbred Prediction of Nixtamalization Moisture Content

Abstract

Background and Objectives

The quantity of water absorbed during the nixtamalization of maize greatly influences the final product's taste, nutritional profile, and machinability. A machine learning model that uses near-infrared spectroscopy to predict the moisture content of nixtamalized maize inbred lines was previously developed. Inbred and hybrid maize differ in many ways including shape, size, and composition of kernels, which can all affect nixtamalization moisture content.

Findings

The inbred model was assessed for application with hybrid germplasm, the primary input for most industrial uses, and a low Spearman correlation coefficient of 0.539 was observed. A new model trained on diverse hybrid maize was developed and validated. The hybrid model achieved a Spearman's rank correlation coefficient of 0.815 across five populations of food-grade and nonfood-grade maize.

Conclusions

The hybrid model was accurate and used to assess relationships between grain compositional properties and nixtamalization moisture content and significant relationships with fat and fiber content were found.

Significance and Novelty

The hybrid model developed here and the previous inbred model have been incorporated into a Shiny R application called CHIP-NMC, which can be incorporated into various stages in the masa-based product development chain including breeding, elevator acceptance, and manufacturing.