Sensing the quality parameters of Chinese traditional Yao-meat by using a colorimetric sensor combined with genetic algorithm partial least squares regression

• A colorimetric sensor array was fabricated by printing 16 dyes on a silica-gel plate.
• The TVB-N, TVC and RN of Yao-meat were measured by conventional method.
• The color changes of sensors were correlated with TVB-N, TVC and RN by GA-PLS.
• The GA-PLS was used to select the most informative chemically responsive dyes.

Yao-meat is a traditional Chinese salted meat. Total volatile basic nitrogen content (TVB-N), total viable bacterial count (TVC), and residual nitrite (RN) level are important indexes of freshness for Yao-meat. This paper attempted the feasibility to determine TVB-N content, TVC and RN level in Yao-meat by a colorimetric sensor array chip. A color change profile for each sample was obtained by differentiating the image of sensor array before and after exposure to Yao-meat's volatile organic compounds. Genetic algorithm partial least squares regression (GA-PLS) was proposed to build the relationship between the TVB-N content, TVC, RN and the color change profiles of sensor array, and to select informative chemically responsive dyes for the three quality parameters. The GA-PLS models were achieved with RTVB-N = 0.812, RTVC = 0.856, RRN = 0.855, in prediction set. This study demonstrated that colorimetric sensory array with GA-PLS algorithm could be used successfully to analyze the quality of Chinese traditional Yao-meat.