The influence of cell wall bound calcium, cell number and size on the development of mealiness in 'Forelle' pear. Evaluation of X-ray CT and NIR as non-destructive techniques for mealiness detection

Muziri, Tavagwisa (2016-03)

Thesis (PhD)--Stellenbosch University, 2016.

Thesis

ENGLISH ABSTRACT: ‘Forelle’ pear (Pyrus communis L.) is the second most planted pear and the second highest generator of foreign exchange for pears in South Africa. It is favoured for its red blush, melting texture, sweet taste and pear flavour. However, ‘Forelle’ develops mealiness, a floury, soft and dry texture with low extractable juice. Consumers dislike mealy fruit. ‘Forelle’ mealiness has been characterized by a loss of cell to cell binding during ripening in a previous study. This study aimed to further understand the role of cell wall bound and free Ca2+, as well as the cell size and cell number in the development of mealiness in ‘Forelle’. In addition, two non-destructive methods for the detection of mealiness in intact pears were examined. It was found that free Ca2+ constituted about 49-73% of the total cell Ca2+. Depending on farm origin, mealy fruit contained a lower free Ca2+ concentration compared to non-mealy pears. Plant growth regulators and selective blossom thinning that caused larger cells had a higher mealiness percentage. Scanning electron microscopy revealed larger intercellular spaces for treatments with a higher mealiness incidence. Macro X-ray computed tomography (X-ray CT) showed a higher percentage of defects in the neck of fruit that would become mealy after storage, and after softening. To our knowledge this is a first such finding. Micro (X-ray CT) found that cells of mealy fruit were larger and ellipsoidal in shape while non-mealy cells were smaller and more rounded. Mealiness was also associated with high fruit porosity. A further study described physicochemical measurements which relate to mealiness. Mealy fruit were mostly larger with a higher total soluble solids (TSS), TSS:TA ratio and lower juice area and juice weight obtained by a confined compression method. Fourier transform near-infrared absorbance spectroscopy (FT-NIR) was employed to determine if spectra could be used to distinguish between mealy and non-mealy fruit using sensory and TSS based schemes. Classification was done using orthogonal partial least squares discriminant analysis (OPLS-DA). This study showed that FT-NIR spectra can indeed be used to discriminate between mealy and non-mealy ‘Forelle’ pears. Two–class (mealy and non-mealy) discriminant analysis produced models with accuracies ranging from 51% to 95%. Mealiness caused an increase in transmittance in specific regions of the spectra. FT-NIR was then evaluated for the quantification of TSS using partial least squares (PLS) regression. Validated models had root mean squared error of prediction (RMSEP) = 0.76-0.94 and relative prediction deviation (RPD) = 1.53-2.17, with the equator blush consistently giving better performance for three farms making the model ideal for hand held FT-NIR applications. External validation results of farm location showed reduced model robustness. The decrease in prediction performance was attributed to the differing TSS ranges in locations and possibly seasons. It is recommended that future studies on FT-NIRs calibration models for ‘Forelle’ use fruit from wide origins with wide TSS ranges over various seasons.

AFRIKAANSE OPSOMMING: ‘Forelle’ pere (Pyrus Communis L.) is die tweede mees aangeplante peer en die tweede grootste inkomste genereerder van buitelandse valuta vir pere in Suid Afrika. Dit word verkies vir sy rooi blos, smeltende tekstuur, soet smaak en peer geur. ‘Forelle’ ontwikkel egter ʼn melerige, sagte en droë tekstuur met lae ekstraheerbare sap. Verbruikers hou nie van melerige vrugte nie. ‘Forelle’ melerigheid is in ʼn vorige studie verbind aan die verlies van sel tot sel verbindings gedurende rypwording. Die doel van hierdie studie was om die rol van selwand gebinde en vrye Ca2+, asook selgrootte en selgetal in die ontwikkeling van melerigheid in ‘Forelle’ te verstaan. Laastens is twee nie-destruktiewe metodes vir die opsporing van melerigheid in intakte pere ondersoek. Daar is gevind dat vrye Ca2+ omtrent 49-73% van die totale sel Ca2+ uitmaak. Afhangend van die boord van oorsprong, het melerige pere ʼn laer konsentrasie van vrye Ca2+ bevat, in vergelyking met nie melerige pere. Plant groeireguleerders en selektiewe bloeiseluitdunning wat groter selle veroorsaak het, het ʼn hoër persentasie van melerige vrugte gehad. Skanderings elektron mikroskopie openbaar dat groter intersellulêre spasies, in behandelings wat meer melerigheid gehad het, gevind is. Makro X-straal verwerkte tomografie (X-straal VT) het aangedui dat defekpersentasie hoër was in die nek van vrugte wat melerig sou word na opberging, asook na sagwording. Dit is volgens ons kennis die eerste diesulke bevinding. Mikro X-straal VT het gevind dat melerige selle groter en meer ellipsoïdaal in vorm was terwyl nie-melerige selle kleiner en meer rond was. Melerigheid was ook geassosieer met ʼn hoër vrugporositeit. ‘n Verdere studie het die fisikochemiese metings wat ʼn verwantskap toon met melerigheid bepaal. Melerige vrugte was meestal groter vrugte met ʼn hoër persentasie van totale oplosbare vastestowwe (TOVS), TOVS:titreerbare suur verhouding, en ʼn laer sap area en sap gewig wat bepaal was met die begrensde kompressie metode. Fourier getransformeerde naby infrarooi absorpsiespektroskopie (FT-NIR) om te bepaal of spektra benut kan word om tussen melerige en nie-melerige vrugte te onderskei d.m.v. sensoriese en TOVS klassifikasie is gedoen met behulp van ortogonale gedeeltelike laagste kwadraat diskriminant analise (OPLS-DA). Die studie toon dat FT-NIR spektra wel gebruik kan word om tussen melerige en nie-melerige ‘Forelle’ pere te onderskei. Twee–klas (melerige en nie-melerige) diskriminant analises het modelle geproduseer met ʼn akkuraatheid van 51% tot 95%. Melerigheid het ʼn verhoging in transmissie in sekere areas van spektra veroorsaak. FT-NIR is verder geëvalueer vir die kwantifisering van TOVS met die gebruik van gedeeltelike laagste kwadraat (PLS) regressie modelle. Die gevalideerde modelle het gemiddelde kwadraat fout van voorspelling (RMSEP) = 0.76-0.94 en relatiewe voorspellingsafwyking (RPD) = 1.53-2.17 gehad, met die ewenaar bloskant van die vrug wat konsekwent beter presteer het vir drie plase. Dit sou die model ideaal maak vir handgehoude tipe FT-NIR toepassings. Eksterne validasie resultate van plaas posisie toon verlaagte model robuustheid. Die verlaging in voorspellings prestasie was toegeskryf aan die TOVS verspreiding wat verskillend was vir plase en moontlik ook seisoene. Daar word aanbeveel dat toekomstige kalibrasie model studies op ‘Forelle’ vrugte van ʼn wye oorsprong en TOVS verspreidings insluit oor verskeie seisoene.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/98776
This item appears in the following collections: