Study On The Lateral Flow Immunoassay For Detection Of Pesticide Multi-residues Based On Smartphone-based Image Recognition System

Pesticides,mainly including insecticides,fungicides,herbicides and plant growth regulators,are essential chemical inputs for the agricultural production.Different types of pesticides may be used during the seed treatment,sowing,growth and storage.With the excessive,over-range or illegal use of pesticides,there may be pesticide multi-residues in agro-products.Most fresh edible agro-products have timeliness,and their entry into the market needs safety supervision,and rapid detection technology is an important means to find problems the first time and ensure the expense security.In recent years,the rapid detection technology based on highly sensitive immunochromatographic test strips and smartphones has become a research hotspot.However,previous reported smartphone-based lateral flow immunoassay（LFIA）technologies mainly face two inevitable challenges of cumbersome image acquiring devices（dark boxes with light sources）for avoiding the influence of the ambient light,and poor universal applicability for different types of smartphones due to the fixed relative location between smartphones’built-in cameras and test strips.Most smartphones were only used to obtain images of test strips,and separately open-source data analysis software（such as Image J）was used for the image processing.Besides,highly sensitive and high-throughput detection also limits the practicablity of LFIA.This study selected five fruits and vegetables（cowpea,celery,tomato,apple and cucumber）,and the pesticides（imidacloprid,acetamiprid,carbofuran,triadimefon and paclobutrazol）as the study objects,and a smartphone-based lateral flow image recognition rapid detection method was developed.It could realize the goal of rapid quantitative,high-throughput,highly sensitive and intelligent detection of pesticide multi-residues under ambient light conditions.The developed method has good universality and is applicable to different types of smartphones.It significantly improves the practicability and intelligent level of the LFIA-based rapid detection technology.It can provide a method reference for realizing the on-site rapid screening of pesticide multi-residues in complex fruit and vegetable matrices.The main contents and results of this study are as follows:1.In this study,the effect of several key parameters including the illuminance and color temperature of light source,different types of smartphones,and photography distance between the cameras and LFISs on the self-developed smartphone-based image recognition system was evaluated under the condition of standard light source.Results showed that the illuminance and color temperature of light source had little impact on the results obtained from the developed system（RSDs<8%）.The developed system had good stability and universality,which could be applicable to different types of smartphones without fixing the relative location between the cameras and CGISs.Besides,smartphones used for quantitative detection could produce results consistent with the CCD and CMOS cameras.In order to further evaluate the applicability of the developed system,the effect of different types of smartphones and different scenes was investigated in ambient light.Results showed that the system colud also exhibit great stability and accuracy in ambient light,which demonstrated a great consistency with the results obtained by the professional strip reader with the R²=0.9947（the slope was 1.029）.2.On the basis of the traditional colloidal gold immunochromatographic strip（CGIS）,the sensitivity of LFIA was improved only by lowering the conjugate ratio（CR）between the hapten and the carrier protein（BSA）in the capture antigen,without modifying the composition and structure of the test strip in any form.The enzyme-linked immunosorbent assay（ELISA）procedure using NH₄SCN as the chaotropic agent was used to determin the relative avidity between monoclonal antibodies（m Abs）and different coating antigens with different CRs,which figured out the working principle of lowering the CRs to improve the sensitivity of LFIA.Combined with the way of lowering the original T/C values obtained by testing with blank samples,the sensitivity of LFIA for IDP,ATP and CBF was totally increased for 1.5-,1.0-and 3.3 fold,respectively.The developed smartphone-based lateral flow image recognition system was finally applicable for the simultaneous quantitative determination of IDP,ATP and CBF in cowpea,celery and tomato matrices under ambient light.The method detection limits（MDLs）for IDP,ATP and CBF were 0.54-0.96,13.8-22.2 and 5.6-12.6μg/kg in the tested three vegetable matrices.The recoveries of the smartphone-based LFIA technique for three analytes ranged from 71.7%to 98.4%with RSDs<18%,indicating the good accuracy and precision.The practical applicability of the proposed method was evaluated by testing real contaminated vegetable samples,and was confirmed by the liquid chromatography-tandem mass spectrometry（LC-MS/MS）.3.On the basis of the avidity adjustment strategy,this study also adopted the multiline design and the broad-specific antibody strategy to realize the simultaneous determination of multiplex targets on the same CGIS.Under the optimum p H value and the amount of labeled antibody,specific-antibodies of TDF and CBF,and m Abs of ATP were labeled with colloidal gold,respectively.They were mixed and coated on the conjugate pad in the ratio of 1:1:1,and three T lines（T1,T2 and T3）were set on the membrane to coat capture antigens of TDF,CBF and ATP,respectively.The CGIS with three T lines was finally preparaed for the simultaneous determination of TDF,PBT,CBF,3-OH-CBF and ATP.Combined with the smartphone-based image recognition system,a rapid detection method for the simultaneous quantitative detection of these five pesticide residues in apple,tomato and cucumber matrices under ambient light was successfully developed.The method displayed good accuracy,precision,sensitivity and practicability.The average recoveries were 72.3-112.8%with RSDs<18%and the MDLs were 2.9-18.0μg/kg.The detection results obtained from the developed method were in good agreement with those of LC-MS/MS.The developed multiline CGIS combined with the six-channel image recognition system can realize the simultaneous determination of five targets in six samples in 15min,which can significantly improve the LFIA with high-throughput and intelligent detection.