|Rapid evaporative ionization mass spectrometry for high throughput screening in food analysis: the case of boar taint
|Advanced Analysis of Food and Beverages
|Dr Sara Stead
|Ms Kaat VERPLANKEN
Dr Jella Wauters
Prof Lynn vanhaecke
Prof Zoltan Takats
Dr Nathaniel Martin
Abstract Information :
Increasing awareness on animal welfare has led to a European Treaty announcing a voluntary ban on the surgical castration of piglets by 2018. An alternative is to raise entire males however; the limitation is the possible occurrence of "boar taint" an off-odour caused by the accumulation of indole (IND), skatole (SK) and androstenone (AEON) in adipose tissue. IND and SK are indolic compounds derived from the degradation of L-tryptophan in the hindgut and their odour is often described as faecal-like. AEON is a pheromone produced in the Leydig cells of the testis having a urinary or sweaty like odour. To prevent adverse consumer reactions there is an urgent need for rapid methods for detecting "boar taint" at the slaughter line.
Rapid Evaporative Ionization Mass Spectrometry (REIMS) was used as an emerging direct analysis technique to train a predictive model for accurate high-throughput identification of boar taint above the odour threshold using representative samples of characterised pig adipose tissue. Adipose tissue was sampled using the iKnife handheld monopolar device, which was connected directly to a Xevo G2-XS Q-TOF mass spectrometer equipped with a REIMS source (Waters corporation, Manchester, UK). For each sample, 2 technical replicates were taken with a sampling time of 3-5 seconds. Untargeted mass spectrometric profiling in both negative and positive ionisation mode of pig neck fat samples enabled the construction of statistical models for the classification of pig carcasses in boar taint positive or negative groups and sows.
To demonstrate the classification potential of REIMS, blank (sow), boar taint positive and negative carcasses were included, in total of 150 samples. Both negative and positive ionization modes were investigated to increase the range of detected metabolites and were considered separately. In negative ionization, better classification accuracy (98%) was observed compared to positive ionization (94%). The OPLS-DA model showed good separation between sow and boar groups. The two boar groups showed some overlap, nevertheless, cross-validation demonstrated the model had a total correct classification rate of 99% and consequently could be used as an accurate predictive tool for boar taint. All blank and negative samples were correctly classified, whereas of the boar taint positive samples, 98% were correctly classified. The remaining 2% were classified as negative. The classification results obtained by chemical and sensory analysis, (which were used as Y-information for model building) could form the basis of the miss classification of these samples. Based on the sensory scores of the neck fat samples, these samples were severely tainted. The validity of the model was evaluated through R2(Y) and Q2(Y), CV-ANOVA testing and permutation tests. The results obtained in this study demonstrate that despite the lack of sensitivity for the boar taint compounds in targeted detection, tainted carcasses could be correctly classified by an untargeted approach. This makes REIMS suitable for discrimination between gender samples (sow versus boar) and for discrimination within gender (tainted versus untainted). This discrimination originates from alterations in lipid profiles, primarily in the fatty acid and phospholipid regions. As REIMS eliminates extensive sample pre-treatment procedures, analysis takes <10 seconds, it offers potential as the first technique enabling in-situ detection of boar taint combined with accurate classification. REIMS is a promising and powerful tool for other applications in food quality, whereby rapid characterization of food products is requisite.
Direct screening method for the classification of boar taint positive carcasses generating results within 10 seconds.