We collected a large amount of data from our questionnaire and DNA analysis. Our basic material consisted of 419 questionnaires with corresponding DNA. In the laboratory, however, there were unfortunately also mistakes - no wonder with so many samples and scientists ;-). Once, the wrong tube was disposed of; another time, the DNA probes did not work very well and one of the three SNPs was not clearly assigned. Finally, we were able to enter 296 reliable DNA analysis for all 3 SNPs into the data bank. This data then needed to be compared to the information from the questionnaire (insofar as these were completely filled out; for example, the questions on the taste perception of stevia were only filled out by 409 of the subjects that we evaluated)

We met during the Christmas holidays to enter the values into a self-created Microsoft Access data bank that a father had actively helped to develop. In order to ensure the professionalism of our analysis, we looked for assistance. University professor Dr. Karl Entacher of the University of Applied Sciences Salzburg Urstein (Fachhochschule Salzburg Urstein), finally agreed to conduct the analysis. After creating the first chart, he showed us the program IBM SPSS Statistics 19 (licensed by the FH Urstein) in an evening course and also patiently corrected our initial errors. In the following charts and diagrams, we have summarized the most important results. The reliability of the results was checked with the chi-square test. Further results will be available as of June 2011 on the project homepage (http://projekte.ursprung.at). The most important result of the project is summarized under point 6.

1.) Distribution of the Testers Based on Age and Sex

(399 valid questionnaires)

Fig. 1

2.) Distribution of Genotypes

(296 valid results)

Tab. 1 Frequency of the Genotypes
Fig. 2 Frequency of the Genotypes

As expected, the genotype PAC/AVI was the most common with 39 percent, followed by AVI/AVI with 31 percent and PAV/PAV with 7 percent. The other genotypes are rare and summarized further in the group "Other". When comparing these results with scientific literature, it shows that PAV/AVI and AVI/AVI are distributed as expected. Striking, however, is the low proportion of PAV/PAV. According to a study of 1,252 individuals in Erfurt, Germany, this percentage should be significantly higher, around 14 percent (Sausenthaler 2009). Another study in the USA of 500 individuals reached 14 percent (Cannon 2005). Our deviation could be explained by the smaller sample size, as well as the fact that the one of the three DNA probes did not work as well as the others and the failure rate in the laboratory was not equally distributed across all three SNPs. Furthermore, we observed a higher proportion of rare genotypes ("Other"), which we would have expected to be lower in Austria. However, this result adheres to the scientific literature, given that we did our sampling at the House of Nature, which attracted a higher proportion of foreign tourists among the test subjects. The genotype is dependent on ethnicity; thus, people from sub-Saharan Africa are more likely to have variations that are rare here in Austria.

Tab. 2 Frequency of Genotypes, grouped
Fig. 2 Frequency of Genotypes, grouped
Fig. 3 Frequency of Genotypes, grouped with Error Indicators at the Data Points


Sausenthaler S. et al. 2009, Lack of Relation Between Bitter Taste Receptor TAS2R38 and BMI in Adults.Obesity 17 5, 937-938
Cannon D.S. et al. 2005, Associations between phenylthiocarbamide gene polymorphisms and cigarette smoking. Nicotine & Tobacco Research 7, 6 853–858

3.) Distribution of Sex and Age Group by Genotype

(296 valid results)

Figures 4 and 5 show the distribution of sex and age groups "children and youth", as well as "adults" by different genotypes.

Fig. 4 Sex distribution by Genotype
Fig. 5 Age Group Distribution by Genotype

4.) The Taste Perception Concerning PROP by Genotypes

(296 valid results)

All subjects tasted different concentrations of PROP solutions (see Table 1, Chapter 5), beginning with the lowest concentration (number 1). As soon as they tasted the bitterness, they crossed off the corresponding number on the questionnaire. Those who did not taste any bitterness in any of the tests, crossed off number 7. PROP is a quasi-standard for sensory tests of the bitterness gene TAS2R38. In figures 6 and 7, the results are represented in different graphs.

Fig. 6 PROP-Perception by Genotypes
Fig. 7 PROP-Perception by Genotypes

In earlier literature, people with the genotype PAV/PAV were considered Super-Tasters, because they could perceive PROP even in the smallest doses. In our trial, the type PAV/PAV could taste PROP only at the maximum concentration number 3. If a higher concentration was tasted, the taster sometimes had to run to the sink, because such a horrible taste was perceived. Earlier, people with the genotype AVI/AVI were designated as Non-tasters, because 80 percent of the time, they could taste PROP only at the highest concentration or not at all. In our results, AVI/AVI people most frequently crossed off number 7, the solution with the highest proportion of PROP. These people did not perceive any of the solutions as bitter.

Our study also shows why the terms Supertaster and Nontaster in relation to PROP are hardly used anymore in recent scientific literature. As shown in Figure 6, there are also people with the genotype AVI/AVI (alleged Nontasters), who perceive small amounts of bitter substances. The reason is that the disposition to Supertasters is localized on more than one gene, as shown in Hayes et. al. 2008. Our distribution of taste perception that we elicited in relation to PROP by the genotype AVI/AVI is consistent with Hayes's work.

Our sensory tests and the corresponding DNA analysis coincide perfectly with recent scientific literature. We are very proud of this. From time to time over the course of the project, we feared that mistakes could be hiding. The first DNA analysis results of the AVI/AVI type appeared in the monitor and showed that the testers had obviously tasted the bitterness of the concentration number 1. Something did not seem to fit. When we also had the number 7 testers, the people who could not taste PROP at all, entered alongside the AVI/AVI types, we breathed easier.


Hayes E.J. et al 2008, Supertasting and PROP Bitterness Depends on More Than the TAS2R38 Gene, Chem. Senses 33: 255–265.

5.) The Questionnaire Feedback on the Taste of Stevia

(409 valid questionnaires)

All of the testers tasted food products that were sweetened with stevia, including various teas, cakes and puddings. The taste perception was rated in part by multiple choice questions, as well as descriptively ("tastes like medicine", "artificial", "chemical", "bitter"). Even the aftertaste a couple of minutes later was classified into the categories "pleasant", "neutral" and "unpleasant". The following are the evaluations.


Total Sample
409 Subjects

unpleasant: 105; 25,7%

neutral: 122; 29,8%

pleasant: 182; 44,5%


Male: 223 Subjects

unpleasant: 41; 8,4%

neutral: 80; 35,9%

pleasant: 102; 45,7%


Female: 186 Subjects

unpleasant: 64; 34,4%

neutral: 42; 22,6%

pleasan: 80; 43%


Younger than 17 Years Old: 109 Subjects

unpleasant: 20; 18,3%

neutral: 39; 35,8 %

pleasant: 50; 45,9%


Older than 22 Years Old: 191 Subjects

unpleasant: 56; 29,3%

neutral: 50; 26,1 %

pleasant: 85; 44,5%


Older than 30 Years Old: 136 Subjects

unpleasant: 39; 28,7%

neutral: 36; 26,5 %

pleasant: 61; 44,8%


Older than 50 Years Old: 47 Subjects

unpleasant:10; 21,3%

neutral: 9; 19,1 %

pleasant: 28; 59,6%

Tab. 3: Taste Perception of Stevia, Evaluation of the Questionnaire Regardless of Genotype

6.) Taste Perception in Relation to Stevia by Genotype

(296 valid results)

In the next step, we analyzed the taste perception of stevia combined with the three major genotypes (AVI/AVI, AVI/PAV, PAV/PAV) and discovered something astounding:

as shown in Figure 8, no one with the genotype PAV/PAV rated stevia as negative. Everyone found stevia to taste good or at least neutral. More people with the heterozygote genotype PAV/AVI (i.e. those who inherited different variants of the gene from their parents), rated stevia as pleasant than those with genotype AVI/AVI. It can easily be said that the more PAV is present, the better stevia tastes.

This is surprising. We had assumed that the group PAV/PAV with their "Supertaster" ability to perceive bitterness better, would probably not like stevia. And now our result suggests that the opposite is true.

The correlation between the genetic variant of the bitter receptor gene TAS2R38 and the rating of the sweetener of the stevia plant entirely contradicts our working hypothesis. Research can really be exciting! The result is "highly significant", meaning with the probability of p=99% true. (In professional terms, through a chi-square test, the hypothesis or decision that the feature "Stevia General" is distributed in the same way in the different basic populations (AVI/AVI, PAV/AVI and PAV/PAV), can be most significantly rejected.)

According to our research, we can say that approximately 14 percent of the Central European population (i.e. the distribution of PAV/PAV for Germany according to Sausenthaler 2009) will like the taste of stevia sweetener - and we would have "read this prediction in the DNA".
Certainly, a study on the other 24 genes related to taste would provide further information. Unfortunately, we lack the time and above all the money to purchase further DNA probes for analysis.

Fig. 8 Taste Perception in Relation to Stevia by Genotype

7.) The Taste Perception in Relation to Aronia by Genotype

(130 valid records)

Aronia is very bitter; it consists of many ingredients that taste astringent and bitter. According to our hypothesis, there should be a certain connection to the gene TAS2R38, which we studied. Figure 9 shows the analysis of the questionnaire in relation to aronia by genotype. There is a recognizable trend: the aversion to aronia increases with the PAV/PAV group. This adheres to the fact that people with the genotype PAV/PAV also find the standardized bitter solutions disgusting much sooner.

Fig. 9 Taste Perception in Relation to Aronia by Genotype

Additionally, we purposefully did not optimize the good taste of the aronia juice for the tasting, because we did not want to diminish the bitterness of the aronia berry. Mixing it with orange juice or glucose would have dramatically improved the taste. Then, the healthy fruit would have tasted better to more people.

Figure 10, in which the AVI/AVI genotype is represented, is also interesting. It shows that there is a difference in the taste perception of aronia based on age groups. Younger people perceive aronia to be too bitter. This adheres to evolutionary as well as developmental theory. It is a natural, healthy reflex when children spit out bitter plant parts; it helps them to avoid eating anything poisonous. Only in the course of becoming an adult do people realize that bitterness is not dangerous and can even be tasty. This is also the reason why children seldom like grapefruit or brussel sprouts, but later enjoy eating them as adults.

Fig.10 Genotype AVI/AVI by age group and rating of the taste of aronia

8.) Additional Result: Smoking and Taste

We also asked our test subjects whether they were smokers or non-smokers. It is known that smoke can have an adverse effect on taste receptors and, therefore, the pleasure derived from eating.

Our data shows that smokers tasted the PROP later in the test solutions. It could be said that smokers' taste buds are more numb. Since there were 'too few' smokers in our sample, this result is (for the time being) hardly convincing for the tobacco industry.

Fig. 11 Genotype PAV/PAV, Supertaster in Relation to PROP, by Smoker and Non-Smoker

Most interesting result:

No one with the genotype PAV/PAV perceived stevia as negative. Those with the genotype PAV/AVI found stevia pleasant at a higher rate than those with the genotype AVI/AVI. In short: the more PAV, the better stevia tastes.