The use of collision cross-section (CCS) values obtained by ion mobility high-resolution mass spectrometry has added a third dimension (alongside retention time and exact mass) to aid in the identification of compounds. However, its utility is limited by the number of experimental CCS values currently available. This work demonstrates the potential of artificial neural networks (ANNs) for the prediction of CCS values of pesticides. The predictor, based on eight software-chosen molecular descriptors, was optimized using CCS values of 205 small molecules and validated using a set of 131 pesticides. The relative error was within 6% for 95% of all CCS values for protonated molecules, resulting in a median relative error less than 2%. In order to demonstrate the potential of CCS prediction, the strategy was applied to spinach samples. It notably improved the confidence in the tentative identification of suspect and nontarget pesticides.

Collision cross section (CCS) databases based on single-laboratory measurements must be cross-validated to extend their use in peak annotation. This work addresses the validation of the first comprehensive TWCCSN2 database for steroids. First, its long-term robustness was evaluated (i.e., a year and a half after database generation; Synapt G2-S instrument; bias within ±1.0% for 157 ions, 95.7% of the total ions). It was further cross-validated by three external laboratories, including two different TWIMS platforms (i.e., Synapt G2-Si and two Vion IMS QToF; bias within the threshold of ±2.0% for 98.8, 79.9, and 94.0% of the total ions detected by each instrument, respectively). Finally, a cross-laboratory TWCCSN2 database was built for 87 steroids (142 ions). The cross-laboratory database consists of average TWCCSN2 values obtained by the four TWIMS instruments in triplicate measurements. In general, lower deviations were observed between TWCCSN2 measurements and reference values when the cross-laboratory database was applied as a reference instead of the single-laboratory database. Relative standard deviations below 1.5% were observed for interlaboratory measurements (<1.0% for 85.2% of ions) and bias between average values and TWCCSN2 measurements was within the range of ±1.5% for 96.8% of all cases. In the context of this interlaboratory study, this threshold was also suitable for TWCCSN2 measurements of steroid metabolites in calf urine. Greater deviations were observed for steroid sulfates in complex urine samples of adult bovines, showing a slight matrix effect. The implementation of a scoring system for the application of the CCS descriptor in peak annotation is also discussed.

ABSTRACT. The increasingly abundant food fraud cases have brought food authenticity and 1 safety into major focus. In this study, we present a fast and effective way to identify meat 2 products using rapid evaporative ionization mass spectrometry (REIMS). The experimental setup 3 was demonstrated to be able to record a mass spectrometric profile of meat specimens in a time 4 frame of less than 5 seconds. A multivariate statistical algorithm was developed and successfully 5 tested for the identification of animal tissue with different anatomical origin, breed and species 6 with 100% accuracy at species and 97% accuracy at breed level. Detection of the presence of 7 meat originating from a different species (horse, cattle, and venison) has also been demonstrated 8 with high accuracy using mixed patties with a 5% detection limit. REIMS technology was found 9to be a promising tool in food safety applications providing a reliable and simple method for the 10 rapid characterization of food products. 11 12

IntroductionFish fraud detection is mainly carried out using a genomic profiling approach requiring long and complex sample preparations and assay running times. Rapid evaporative ionisation mass spectrometry (REIMS) can circumvent these issues without sacrificing a loss in the quality of results.ObjectivesTo demonstrate that REIMS can be used as a fast profiling technique capable of achieving accurate species identification without the need for any sample preparation. Additionally, we wanted to demonstrate that other aspects of fish fraud other than speciation are detectable using REIMS.Methods478 samples of five different white fish species were subjected to REIMS analysis using an electrosurgical knife. Each sample was cut 8–12 times with each one lasting 3–5 s and chemometric models were generated based on the mass range m/z 600–950 of each sample.ResultsThe identification of 99 validation samples provided a 98.99% correct classification in which species identification was obtained near-instantaneously (≈ 2 s) unlike any other form of food fraud analysis. Significant time comparisons between REIMS and polymerase chain reaction (PCR) were observed when analysing 6 mislabelled samples demonstrating how REIMS can be used as a complimentary technique to detect fish fraud. Additionally, we have demonstrated that the catch method of fish products is capable of detection using REIMS, a concept never previously reported.ConclusionsREIMS has been proven to be an innovative technique to help aid the detection of fish fraud and has the potential to be utilised by fisheries to conduct their own quality control (QC) checks for fast accurate results.Electronic supplementary materialThe online version of this article (10.1007/s11306-017-1291-y) contains supplementary material, which is available to authorized users.

Boar taint is a contemporary off-odor present in meat of uncastrated male pigs. As European Member States intend to abandon surgical castration of pigs by 2018, this off-odor has gained a lot of research interest. In this study, rapid evaporative ionization mass spectrometry (REIMS) was explored for the rapid detection of boar taint in neck fat. Untargeted screening of samples (n=150) enabled discrimination between sow, tainted and untainted boars. The obtained OPLS-DA models showed excellent classification accuracy, i.e. 99% and 100% for sow and boar samples or solely boar samples, respectively. Furthermore, the obtained models demonstrated excellent validation characteristics (R(Y)=0.872-0.969; Q(Y)=0.756-0.917), which were confirmed by CV-ANOVA (p<0.001) and permutation testing. In conclusion, in this work for the first time highly accurate and high-throughput (<10s) classification of tainted and untainted boar samples was achieved, rendering REIMS a promising technique for predictive modelling in food safety and quality applications.