Machine Learning Integrated with Acoustic Sensors Enables an Efficient and Deep Analysis of the Quality and Safety of Milk

Machine Learning Integrated with Acoustic Sensors Enables an Efficient and Deep Analysis of the Quality and Safety of Milk

Scientific Achievement
Figure illustrates self- assembly of aggregates consisting of 50 individual casein molecules on the surface of a piezoelectric crystal. The graph on the back shows changes in the frequency of the piezoelectric crystal as the self-assembled casein layer is removed through interaction with trypsin (black) and plasmin (red).  (hi-res image)

Developed an efficient low-cost method to detect protease (enzyme) activity in milk that is 25 times better than commercial protease fluorescent detection.

Significance and Impact

The ability to predict protease activity in milk is essential for its safe use. The simple casein-coated acoustic sensor developed in our work allows very rapid and cost-effective sub-nanomolar detection and classification of protease activity in milk.

Research Details

– The plasmin is a bovine milk protease and its activity controls milk texture and flavor.

– Conventional methods to detect plasmin are time consuming and expensive.

– Integration of machine learning to analyze casein –enzyme reaction  enabled a high sensitivity low cost technique for plasmin detection.

 

M. Tatarko, E. S. Muckley, V. Subjakova, M. Goswami, B. G. Sumpter, T. Hianik, and I. N. Ivanov, "Machine learning enabled acoustic detection of sub-nanomolar concentration of trypsin and plasmin in solution,"  Sensors & Actuators: B. Chem272, 282 (2018).  DOI: 10.1016/j.snb.2018.05.100

CNMS Researchers

Share