Near-infrared reflectance spectroscopy-based fast versicolorin A detection in maize for early aflatoxin warning and safety sorting.

Abstract:

:Aflatoxins (AFs) are potent carcinogens present in numerous crops. Access to accurate methods for evaluating contamination is a critical factor in aflatoxin risk assessment. Versicolorin A (Ver A), a precursor of aflatoxin B1 (AFB1), can be used as an indicator for the presence of AFB1, even when the AF is not yet detectable. Currently employed Ver A detection methods are expensive, time consuming, and difficult to apply to numerous samples. Herein, Ver A was detected via near-infrared spectroscopy. Both quantitative and two-grade sorting methods were set-up using the extreme gradient boosting algorithm coupled with a support vector machine. This two-tiered method obtained a root-mean-square error of prediction value of 3.57 μg/kg for the quantitative model, and an accuracy rate of 90.32% for the sorting approach. This novel method is rapid, accurate, solvent free, requires no sample pretreatment, and detects Ver A in maize, making it convenient for practical use.

journal_name

Food Chem

journal_title

Food chemistry

authors

Zheng SY,Wei ZS,Li S,Zhang SJ,Xie CF,Yao DS,Liu DL

doi

10.1016/j.foodchem.2020.127419

subject

Has Abstract

pub_date

2020-12-01 00:00:00

pages

127419

eissn

0308-8146

issn

1873-7072

pii

S0308-8146(20)31281-4

journal_volume

332

pub_type

杂志文章