Research On Rapid Methods For Mold And Deoxynivalenol Contamination Detection In Wheat And Its Products

Freshly-harvested wheat is susceptible to mold infection during transportation and storage,resulting in harmful mycotoxins,which can be harmful to human health.This paper mainly studies the spectral information and odor substance changes of wheat which contaminated by harmful fungi and vomit toxin.First,the spectral information of samples were obtained by near infrared spectroscopy(NIR)and mid infrared spectroscopy(MIR).Quantitative analysis models between infrared spectra and DON concentrations were established by partial least squares regression analysis(PLSR)and stepwise multiple linear regression(SMLR).Secondly,the spectral information and odor information of the samples were obtained by near infrared spectroscopy and electronic nose.Finally,a qualitative and quantitative detection model of wheat infected by harmful molds based on near-infrared spectroscopy was established.Electronic nose combining with GC-MS technology were used to determine the chemical composition and content of harmful mold wheat volatiles.The change of volatile components with storage time was then analyzed to compare the differences in wheat volatiles caused by different harmful molds.The main conclusions of this study are as follows:1.A quantitative detection model of DON pollution in wheat and its products was established based on NIR and MIR.(1)The NIR results showed that the absorption values of samples with different contents of DON were different at 6790 cm-1,5753 cm-1,5145 cm-1,4727 cm-1 and 4322 cm-1.In PLSR,the coefficient of determination of prediction set(RP2),root mean square error prediction(RMSEP)and residual predictive deviation(RPD)were 0.9004,0.412mg/kg and 3.06 respectively.In SMLR,the results showed that the modeling effect of wave number 11 was optimal,the coefficient of determination of prediction set(Rp2),root mean square error f prediction(RMSEP)and residual predictive deviation(RPD)were respectively as 0.8778,0.438mg/kg and 2.79.(2)The MIR results showed that the absorption values of samples with different contents of DON were different at 1740 cm-1,1648cm-1,1549cm-1 and 900-1300 cm-1.The quantitative results of PLSR model for DON toxin content in wheat are as follows:Rp2,RMSEP and RPD were 0.8652,0.438 mg/kg and 2.67 respectively.In SMLR,the results showed that the modeling effect of wave number 9 was optimal,RP2,RMSEP,RPD were respectively as 0.8592,0.426 mg/kg,2.63.2.Qualitative and quantitative detection models for the wheat suffering from harmful fungal infection was established based on near infrared spectroscopy.(1)Near-infrared spectroscopy results showed that near-infrared spectroscopy can distinguish wheat infected with different molds with an average correct rate of 84%.It can better distinguish infections of single mold wheat with an average correct rate of 86%.The correct rate of distinguishing all samples was determined at 82%.The PLSR model of total number of colonies in wheat quantified the Rp2,RMSEP and RPD values as 0.8979,0.376 Log CFU/g and 2.58,respectively.(2)The online NIR results showed that near-infrared spectroscopy can effectively distinguish wheat infected with a single mold when the wheat moving speed was different,and the best discrimination effect was at 0.05 ms-1(low speed)moving speed with an average correct rate of 94%.The correct rate of identification of all samples was 84%.The PLSR model of the total number of colonies in wheat quantified the RP2,RMSEP and RPD values as 0.8687,0.334 Log CFU/g and 2.76,respectively.3.Qualitati've and quantitative detection models for the wheat suffering from harmful fungal infection was established based on electronic nose technology.(1)The electronic nose results showed that the electronic nose technology can effectively distinguish wheat infected with different molds,and the average correct rate was 88%.The average correct rate of electronic nose in infection with a single mold wheat was 95%,and the correct rate of discriminating for all samples was 90%.The PLSR model for the total number of colonies in wheat quantified the results as follow:Rp2 was 0.7977,RMSEP was 0.509 Log CFU/g,and RPD was 2.20.(2)The results of GC-MS showed that in the later period of storage,as the storage time increased,the content of wheat olefins inoculated with 5 kinds of harmful molds decreased,the content of aldehydes decreased and the content of esters as a whole displayed a downward trend.There were certain differences in the volatile components of wheat infected by different molds,and there were also great differences in the different storage stages of the same strain.The results showed that the infrared spectroscopy and the electronic nose method could be used as a rapid and non-destructive method to determine the infection status of moldy wheat,so as to ensure the safety of wheat storage and transportation.