Researchers at Huazhong Agricultural University (HZAU) have made significant progress in applying mid-infrared spectroscopy (MIRS) to rapidly and accurately predict amino acid content in milk, paving the way for high-throughput analysis in dairy cow breeding and milk quality enhancement.

Led by professor Zhang Shujun, the research team published two key studies in Food Research International and the Journal of Dairy Science. The studies established MIRS-based models for predicting total and free amino acid content in milk from Chinese Holstein cows, and examined how milk storage time affects MIRS predictions.

Milk is a major dietary protein source and rich in essential amino acids vital to human development. Yet traditional methods for measuring amino acids, such as high-performance liquid chromatography, are costly and time-consuming – limiting large-scale data collection.

By combining MIRS technology with machine learning algorithms, the team developed accurate, fast, and low-cost prediction models that could be integrated into China's dairy performance recording system.

"Our model has ended dependence on foreign amino acid detection systems and provides a key solution to a long-standing bottleneck in large-scale milk analysis," Zhang said. The research also determined optimal milk storage conditions for accurate testing, ideally within three days and not exceeding six.

The project, supported by the National Key Research and Development Program and central university research funds, offers theoretical and technical support for intelligent phenotyping and genetic evaluation of dairy cattle in China and beyond.

Predicted vs. actual amino acid values in milk using MIRS-based model on the validation dataset. [Photo/news.hzau.edu.cn]