Near infrared reflectance spectroscopy was used to quantify the cordycepic acid and adenosine content in Paecilomyces tenuipes mycelia. Quantification of cordycepic acid and adenosine content was achieved by the conventional method. Models were built using near infrared reflectance for determinating the contents of cordycepic acid and adenosine. Monte Carlo partial least square (MCPLS) and Moving window partial least square (MWPLS) was applied to optimize the models. Finally, the optimum models for determination of cordycepic acid and adenosine contents in Paecilomyces tenuipes powder samples. The correlation between actual and predictive values of calibration samples (Rc) were 0.8851 and 0.9185, the root mean square error of prediction set (RMSEP) were 16.6171 and 0.5949, the root mean square error of calibration set RMSEC were 15.5724 and 0.5844, respectively. These results demonstrate that the robustness, fit and predictive capability of these models were satisfied. This method should be popular in determination the key parameters during fermentation processes.
展开▼
