Driving a better tomato

Even simple foods and beverages are composed of sometimes dozens, but often hundreds, of different chemical compounds. For the most part, these are volatiles, that is, compounds that in the first instance contribute to the aroma of the food. Together with the non-volatile taste compounds – the sugars, acids, salts, amino acids and so on – the aroma compounds make up the characteristic flavour of the food. We know that not all aroma volatiles contribute equally to flavour: some may have no impact, others are crucial to defining the flavour of the food. A subset of these important volatiles may also be the key compounds that determine whether or not you like that flavour.

Take the tomato …. or the strawberry. These are two fruits about which most of us will have strong opinions in regards to flavour quality. We all tend to have an idealized tomato or strawberry flavour memory, somewhere from our past, with occasional reinforcements from holidays in countries where the great flavour of these fruits seemed to have been somehow preserved. But the everyday fruit? It is easy to be impressed by the technical skills that must have been employed to eliminate flavour from our supermarket tomatoes and strawberries.

When let down by the flavour of regular produce, some consumers turn to organic versions in the belief that this approach produces better flavours. Even if this is true (and there is little evidence as yet), then access to sufficient quantities at local supermarkets is likely to be a problem for some time. Breeding programs are a potentially more fruitful approach to producing better flavour. After all, certain varieties of fruit are already selectively bred to have better resistance to disease, better handling qualities, and in some cases, flavour. Hence, Heirloom tomatoes are seen (in the USA at least) as a variety that typically has better flavour appeal.

So far, so good. But in producing a better tomato, what would the breeders breed for? Or, if we wanted to genetically modify a tomato to produce better flavour, which genes would we target? The answer comes down to knowing which volatiles are the most crucial ones that best deliver the most acceptable flavour. To date, an assumption has been made for tomatoes and other fruits and vegetables that those volatiles present in the greatest amount and which are perceived most easily (that is, have the lowest detection thresholds), are the most important contributors to flavour and therefore to consumer acceptability.

The relationship between aspects of flavour and preferences is a key question throughout the food industry. Companies often wish to know which characteristics of their products (or indeed their rival’s products) is ‘driving’ preferences. This is, what qualities are most responsible for variations, up or down, in liking, and conversely what characteristics are irrelevant. A traditional way of doing this has been to have consumers rate liking for products, while specially trained panels rate the intensity of sensory descriptors. The outcome could potentially be simple - your product has too much acid and not enough sugar – but seldom is. Terms such as
grassy, lemony, medicinal, or cherry might be used quite consistently by the trained panel, and also may link well to consumer preferences. However, quite a large extra step is required before a company could then alter their product to maximize the good attributes and minimize the bad ones. A major issue is that terms used by trained panels may reflect the perception of one compound in the food or a combination of many. When we turn to horticultural products, it becomes even more complex due both to variations across seasons and sample, and also the fact that the grower requires either a breeding program or gene technologies to implement changes.

Now, a broad multidisciplinary group at University of Florida (USA), including prominent sensory/consumer scientists Linda Bartoshuk and Howard Moskowitz, has addressed the question of what drives tomato preferences by powerfully combining genetic, sensory and chemical profiling of a wide range of tomato cultivars [
1]. Their results are surprising on a number of fronts, not least from what they did not find. The levels of compounds thought to be important for tomato flavour, or their thresholds, had little consistent impact on whether or not a tomato was liked or disliked in the sensory testing. Part of the study included tomatoes from transgenic plants in which a gene that regulated the levels of so-called C6 volatiles (such as the grassy-smelling compound cis-3-hexenal) was “switched off”. While these tomatoes were certainly different to the tomatoes with normal high levels of C6 volatiles, their flavour was liked equally.

So what
does contribute most to tomato preferences? The study found this to be a complex question, not only because of potentially dozens of volatile compounds involved but also because the levels of the compounds differed by up to 3000-fold across the tomato samples tested. In addition, some varieties did not perform as expected in the blind sensory tests - supermarket tomatoes were sometimes preferred to Heirloom varieties.

We already know that a strong tomato flavour and sweetness are important influences on tomato liking. The researchers were able to pin down a group of 12 compounds that had a large impact on tomato flavour. Intriguingly, Sweetness was also determined by multiple volatile compounds, some of which overlapped with those important to flavour. In particular, three compounds –
geranial (with a rose-like aroma), and 2-methylbutanal and 3-methyl-1-butanol, both malty-smelling (and contributors to the flavour of a diverse range of other foods/beverages, including dairy products, coffee, and wines) – determined the perception of sweetness,
independent of actual sugars in the tomatoes

There has been considerable recent research interest regarding the ability of such retronasal (orally-presented) odours to enhance tastes [
2]. Some studies have, for example, indicated that a tastant such as salt can be reduced without loss of acceptability in those foods in which salty-smelling odours are enhanced. What the data from the Univ. Florida study show is that the consumer’s perception of, and liking for, sweetness in tomatoes is not simply a matter of the glucose or fructose levels in the fruit. Rather, sweetness is more than a taste; it is a flavour, a combination of sweet tastes from the carbohydrates and sweet smells from the volatile compounds.

The crucial lesson from this study is that the powerful combination of sensory/psychophysical, chemical and genetic tools can provide more reliable clues to the sources of consumer preferences. Without these different kinds of data, a finding that non-preferred tomato varieties had low sweetness levels might have led to a conclusion that a higher sugar content was required. In a context of promoting fruit and vegetables as a healthier and tastier alternative to fast foods, a breeding strategy to ‘correct’ the apparent defect might be counter-productive. Instead, demonstrating that volatiles can be important determinants of sweetness and acceptability could justify breeding varieties perhaps with lower sugar levels but enhanced sweet-smelling volatiles.


  1. Tieman, D., et al., The Chemical Interactions Underlying Tomato Flavor Preferences. Current Biology, 2012, 22(11) pp. 1035 - 1039.
  2. Prescott, J., Psychological processes in flavour perception, in Flavour Perception, A.J. Taylor and D. Roberts, Editors. 2004, Blackwell Publishing: London. p. 256-277.