Nutritional profiling of horse gram through NIRS-based multi-trait prediction modelling

Area of Science:

• Agricultural Science
• Food Science
• Biotechnology

Background:

• Horse gram (Macrotyloma uniflorum) is an underutilized, nutritious legume crucial for human nutrition in developing countries.
• Traditional methods for assessing nutritional traits in horse gram are time-consuming and labor-intensive, hindering large-scale germplasm screening.
• There is a need for rapid and efficient techniques to evaluate the nutritional components of horse gram germplasm.

Purpose of the Study:

• To develop and validate Near-Infrared Spectroscopy (NIRS) prediction models for key nutritional components in horse gram.
• To establish a rapid, high-throughput screening method for horse gram germplasm evaluation.
• To identify trait-specific horse gram germplasm for developing healthy food products and improving crop production.

Main Methods:

• Utilized a diverse collection of 139 horse gram accessions for reference data generation.
• Developed prediction models for protein, starch, TSS, phenols, and phytic acid using Modified Partial Least Squares (MPLS) regression.
• Applied spectral preprocessing (SNV-DT) and optimized models for derivatives, gap selection, and smoothing, evaluating with RSQ and RPD statistics.

Main Results:

• NIRS models demonstrated good predictive accuracy for protein (RSQ=0.701, RPD=1.85), starch (RSQ=0.987, RPD=4.03), TSS (RSQ=0.800, RPD=4.06), phenols (RSQ=0.778, RPD=2.15), and phytic acid (RSQ=0.730, RPD=1.88).
• Statistical analyses confirmed the reliability and strength of the developed NIRS models.
• The study achieved high predictive accuracy for pre-breeding applications through a rapid, multi-trait evaluation approach.

Conclusions:

• NIRS provides a rapid and precise alternative to conventional methods for nutritional analysis of horse gram.
• The developed NIRS models are suitable for high-throughput screening of large horse gram germplasm collections and market samples.
• This approach facilitates the identification of superior horse gram germplasm for enhanced food products and agricultural development.