shortcomingsin this regard,as discussed earlier.However, theycould havea
rolein evaluating the effectsof processing,for instance,on the protein of
interest. Suchassays will essentiallybe highlyspecific,and will probablyalso
be sensitive. However, theywill be usedalmostexclusivelyin the laboratory,
and thereforesimplicity of operationwill not be a primary design consideration.
23.4.5 Commonlimitations
Mostof the assaysdiscussed abovehavea numberof limitations,some of which
are inherent in the methodology, while others result from particular
combinationsof methodologyand the substrate in whichthe analyteis sought.
The mostsignificant limitationsrelateto extraction of the analytefromthe food
for analysis, interferenceby othercomponents of the matrix, whichcannotbe
readilyseparated,and changes in the analyteitself,whichreducethe abilityof
the methodto detectit.
Variabilityin extractingthe analytefromthe food
Mostcommonmethods,including many of thosefor totalproteinanalysis (e.g.
Bradford, bicinchoninicacid(BCA))and enzyme-linkedimmunosorbentassay
(ELISA)methods, operatein an aqueous environmentand require extraction of
the proteinprior to analysis.The efficiencyof this extractionwill dependon the
solubility of the protein(s)of interest in the aqueousbufferusedfor extraction.
Manyfoodsand foodproducts includelipidsas part of theirformulation,and
manyproteins,includingallergens,are associated in the foodor productwiththe
lipidcomponent.In experiments to measure the totalresidual proteincontentof
edibleoils,we consistentlyrecoveredonly50%by extraction into phosphate-
buffered saline, basedon a comparisonwiththe content measured by excited
nitrogen analysis,whichdoesnot requireextraction. Thiseffect can be difficult
to detect.In the examplequoted,recoveryof protein spikedintooils was
virtually quantitative, presumably because of differences in their physico-
chemical properties,comparedwiththe proteins remaining in oils afterrefining.
In a differentcontext, Keck-Gassenmeieret al.^18 foundverylow recoveries(2±
3%) of peanutproteinaddedto chocolate products, but wereable to improve this
to near-quantitative recoveryby the addition of fish gelatineto the extraction
buffer.Theseexperiencesindicatethe needfor a thoroughknowledgeof the
physicochemical characteristics of boththe matrixand the protein(s)of interest
in orderto obtain reliableresults.
Matrix interference
As wellas interfering withthe recoveryof the analyte(s)of interest, the food
matrix,or some of its components, mayactuallyinterfere withthe subsequent
assay,if those components are co-extracted in sufficient amounts.For instance,
we have found thaton occasions, solutions withveryhigh sugarcontent
(although within the rangeusedin severalfoods) reducedconsiderablythe
recoveryof -lactoglobulin(unpublishedresults).Othermaterials commonly
Managingrisksfromallergenicresidues 373