Aroma Volatiles - Fruit Ripening: Physiology, Signalling and Genomics - page 65

applicable to the large and multivariate
data derived from SPME-GC-MS analysis,
which can be regarded as a high-through-
put metabolomic tool. Principal component
analysis (PCA) and partial least squares
projection to latent structure (PLS), together
with PLS-discriminant analysis (PLS-DA),
are popular and effi cient statistical methods
to explore and extract data sets from
volatiles in ripening fruit (Aprea et al. ,
2011). In addition, an increased demand for
spectral data annotation and mining
promoted the occurrence of a web-based
database. The Golm Metabolome Database
(http://gmd.mpimp-golm.mpg.de) has been
established to allow researchers to compare
spectral data and generate hypotheses
about the changes in volatiles for mutants
of genes of unknown function.
Fruit aroma volatiles are traditionally
analysed using GC-MS. However, GC
analysis is not a particularly practical
technology because of the complexity of
the volatile compounds.
Because of this fact and the need to link
aroma analysis with human perception,
electronic nose (E-nose) technology has
been developed over the past decade to
investigate the sensory properties of many
fruits (Brezmes et al. , 2005). Such studies
have included the characterization of
different fruit cultivars and maturity stages
with fruits such as mangoes (Lebrun et al. ,
2008), mandarins (Gómez et al. , 2006) and
peaches (Zhang et al. , 2011). In addition,
the E-nose has also been used to determine
fruit ripeness during the postharvest
storage of fruits such as apples (Saevels et
al. , 2004), apricots (Defi lippi et al. , 2009b),
bananas (Llobet et al. , 1999) and peaches
(Infante et al. , 2008). Four peach fruit
cultivars could be separated into cor-
responding clusters using an E-nose on the
day of harvest, and the fruit from each
cultivar exhibited a clear distribution
according to ripening stage (Benedetti et
al. , 2008). In another study, an E-nose was
used to evaluate the aroma quality of peach
fruit during cold storage, and fruits with
chilling injury (CI) symptoms after 21 days
of storage at a chilling-inducing tem-
perature (5°C) were separated from those

without the disorders at 0 and 8°C (Zhang
et al. , 2011).

5.2.2 Fruits with different genetic
backgrounds exhibit differences in aroma
volatiles

As mentioned above, fruit aroma volatiles
are represented by complex mixtures of
compounds, and the volatile profi les are
highly dependent on fruit species and
varieties. Table 5.1 outlines the volatile
compounds that are responsible for some
fruit aroma qualities.
Analysis of fruit aroma volatiles has
been used as a tool to characterize and
classify varieties. The production of vola-
tile compounds by fi ve ripening tomato
fruit mutants ( rin , nor , Nr , Cnr and hp1 )
and the isogenic wild-type 'Ailsa Craig'
were compared. Distinct modifi cations of
the volatile profi les were found in
ripening mutants, and the differences
were most dramatically caused by fatty
acid-derived volatiles such as hexanal and
hexenal (Kovács et al. , 2009). In apple
fruit, the developed PLS-DA model and
the calculated variable importance values
were successfully applied to identify
characteristic volatiles of the four
varieties (Aprea et al. , 2011). Among these
volatile compounds, 3-methylpentanol,
propyl 2-methylbutanoate and isobutyl
2-methylbutanoate are characteristic peaks
for the 'Golden Delicious' variety, and
N -phenylaniline is a marker for 'Granny
Smith' apples. 'Pinova' apples show higher
levels of 2-ethylphenol and hexyl
2-methylbutanoate, whilst 'Stark Delicious'
fruit has higher emissions of ethyl
butanoate, 5-hexenyl acetate and butyl
hexanoate. According to volatile profi les,
50 peach and nectarine fruits with different
genotypic backgrounds could be separated
into four groups, in which Chinese wild
peaches and 'Wutao' had high levels of
terpenoids and esters, cultivars of
American and European origin showed
high linalool, 'Ruipan 14' and 'Babygold 7'
had high lactones, and the remaining
cultivars were without characteristic