8D Yeasts and Wine Flavour 373
tions evaluated by chemical and/or sensory techniques. Nevertheless, the process
of screening and evaluation of new yeasts remains a relatively inefficient process,
since various fermentation characteristics are also assessed in addition to the target
flavour property. Basic fermentation properties include vigour, attenuation, nutrient
requirements, flocculation, and foaming, and freedom from off-flavour formation,
such as H 2 S, acetic acids, high ethyl acetate and high SO 2. Use of robotic technolo-
gies will greatly facilitatehigh throughput screening; however key developments
are still required for multiple characterisations of large numbers of strains. High
throughput metabolomic approaches, with which the volatile fraction of fermenta-
tion metabolites produced can be determined by GC techniques (Howell et al. 2006)
or the non-volatiles by NIR techniques (Cozzolino et al. 2006) coupled to powerful
chemometric tools, are still not suitable for routine work or are inaccessible to many
researchers.
Advanced yeast genetic techniques will bring about the greatest developments
in yeast and wine flavour evolution (Pretorius 2000; Pretorius and Bauer 2002).
Screening yeast using gene arrays techniques, for appropriate gene markers which
are associated with flavour metabolism could reduce the need for extensive metabo-
lite screening (see for example Marullo etal. 2007). Several key genes relating to
ester and long-chain polyfunctional thiol formation, degradation and transforma-
tion have recently been discovered and manipulated to improve the flavour mod-
ulation properties of yeasts (Howell et al. 2005; Subileau et al. 2008; Swiegers
et al. 2006, 2007; Thibon et al. 2008). More appropriate regulation of gene expres-
sion, or modulation of a metabolic pathway,will further facilitate creation of strains
with optimised flavour profiles. Controlling production of off-flavours such as H 2 S
and undesirable mercaptans are important targets (Linderholm et al. 2008; Suther-
land et al. 2003), as is the control of acetic acid production (Pigeau and Inglis 2005;
Saint-Prix et al. 2004). Disconnecting environmental factors from flavour metabolic
pathways will produce strains with more stable flavour properties in musts of vari-
able nutrient content, and lower the risk of nutrient linked off-flavour formation.
Such targets are likely to require application of systems biology approaches that
can integrate genome expression profiles and enzyme catalytic reactions (Borneman
et al. 2007). Further ahead, the introduction of new metabolic pathways into yeast
could be envisaged, leading to the production of compounds which are otherwise
grape-derived. Such an example is the cloning of key genes into yeast to produce
resveratrol (Becker et al. 2003), although this particular compound has no flavour
significance. Strains developed by geneticengineering approaches will have limited
impact in those countries in which consumers and governments regulate their use
in grape and wine production; those countries continue to develop strains for wine
production by conventional genetic breeding techniques (Chambers et al. 2008).
Another recent growth area has been the selection of autochthonous yeasts, which
can promote the regional characteristics of wines in order to overcome the apparent
problem of homogenous or industrial wines. A recent focus has been to select yeasts
on the basis of several yeast volatile aroma compounds but with improved under-
standing of the key odorants of wine the focus will move to those grape-derived
compounds that drive varietal character, as summarised in Table 8D.1. These strains