This web page was produced as an assignment for an undergraduate course at Davidson College.

Article: "A chemical genetic roadmap to improved tomato flavor"

Research suggests commercially grown tomatoes have lost a number of flavor-associated genes

photo from:

1) What was the research project?

Commercially grown tomatoes are available throughout the year and across the world. For this to be possible, tomato farmers have continually selectively breeded for tomatoes demonstrating firmness, high yield, and disease resistance. An implied result of this breeding is the loss of flavorful components over time, leaving us with durable, but rather mundane tasting tomatoes. Researchers utilized tasting panels, chemical analyses, and a genome wide association study of 398 varieties of tomatoes in order to identify genes associated with levels and types of sugars, acids, and volatiles contributing to tomato flavor (volatiles associated with olfactory reception -smell- which we know contributes to taste).

2) Were they testing a hypothesis or doing discovery science?

The study was developed upon the realization that commercially grown tomatoes lacked flavor compared to heirloom varieties. However, researchers had no hypothesis on the relationship between certain chemical compounds and tomato flavor. Thus, this was discovery science as researchers compared the sequenced genomes of tomato varieties with their taste and chemical testing in order to identify any genes that could be associated with rich flavor.

3) What genomic technology was used in the project?

This study used a genome wide association study (GWAS). This means that they sequenced the genomes of 398 varieties of tomatoes and then compared the genome sequences in order to identify key similarities and differences that were related to the existence of certain flavor-associated phenotypes. A key function of a GWAS is to identify single nucleotide polymorphisms (SNPs), or variations in a single base pair of the genome. In this study, researchers aimed to uncover SNPs with a high likelihood of impacting sugar, acid and volatile compound presence. This study targeted SNPs in the exome, or "protein-coding region" of the genome. Thus, SNPs could either alter the amino acid sequence of a protein by exchanging one amino acid with another, truncate a protein by adding a premature stop codon, or shift the reading frame of a protein by inserting or deleting bases in a multiple other than 3 (called an indel). In each case, the amino acid sequence of a protein would be altered, therefore changing a proteins functionality or disabling it entirely. 
For more info on genome wide association studies, check this out:

4) What was the take home message?

The "take home message of the study was that by comparing the exomes of 398 varieties of tomato, researchers were able to identify genes that were likely contributing to the reduction in flavor in commercial tomatoes. In order to make this determination, researchers not only identified differences in the genome, but thoroughly researched the wild-type and mutant proteins involved in order to assess the impact of these variations and prioritize a few genes with a strong likelihood of involvement in tomato flavor reduction. 

fig 1
Figure 1: Genetic interactions between sugar content and weight of fruit (Reproduced from Tieman et al. 2017)

Figure 1A is a Manhattan plot. The x axis represents position in the genome (by chromosome), and each SNP in the genome is represented by a single dot. A manhattan plot enables researchers to visualize association data. This particular plot demonstrates the association between total soluble solids and SNPs across the 398 tomato varieties. Thus on the y axis, in units -log (base 10) of the p value, larger values represent smaller p-values meaning lower probability of type I error (false positive) and therefore a more statistically significant result. We see that SNPs in genes Lin5 and SSC11.1 have some of the strongest significance in association between mutation and soluble solids in tomato fruit. This one good example of making sense of a massive data set acquired by a GWAS. However, finding statistical significance from a bioinformatic approach is only half the battle. The remaining half comes with studies aiming to confirm findings in vivo or in vitro. In figure 1G, researchers present results of an in vivo study. Here, researchers made transgenic plants (meaning they contain addition DNA that is not naturally found in the plant) over-expressing either the wild type (R for reference) or the mutant (A for alternate) at the Lin5 locus. "8059" is a control line over-expressing neither allele. As presented in the figure, the sugar content (y axis) is significantly higher in tomatoes overexpressing the mutant allele.

5) What is your evaluation of the project?

I felt that the project was thorough. The consideration of flavor-associated compounds being not only sugars and acids but also "volitiles" that give certain scents that contribute to taste I felt was very insightful. In addition, the combination of GWAS and its required data mining with targeted, hypothesis-driven experimentation was very effective in demonstrating not only the result in relation to tomatoes but also the power of GWAS and the use of "discovery science" assuming n (sample size) is large enough. One thing I would change is in figure 1G, sample sizes used to generate the data for reference (R) and alternate (A) were 6 and 17, respectively. I am curious as to why researchers did not use equal sample sizes, as tomatoes are easy enough to acquire. One possibility is that using these sample sizes gave researchers the most persuasive data, and that equal, larger sample sizes may not have been as effective in demonstrating their point. Due to the discrepancy, it may be worthwhile to be skeptical of the results of this figure.


Denise Tieman et al. Jan 2017 : 391-394. "A Chemical Genetic Roadmap to Improved Tomato Flavor." A Chemical Genetic Roadmap to Improved Tomato Flavor | Science. Science, 27 Jan. 2017. Web. 29 Jan. 2017.

Return to Nick Balanda's Home Page

Genomics Home

Email Questions or Comments: