Who am I? This is a fundamental question that philosophers and anthropologists have been trying to answer for generations. With the discovery of DNA and the development of technologies to investigate genetic information, people have begun to probe the mystery of identity using biological tools. The ability to read a person’s DNA has helped scientific researchers and medical practitioners to determine an individual’s ancestry, as well as his or her disposition to certain inherited diseases.

In the past decade, the introduction of the do-it-yourself genome kit has made it possible for any person to submit his or her DNA and receive a long list of his or her genetic information. The ease with which people can access their genetic information, including associated risk factors for disease, has led many to wonder if people should have access to information that arguably might require a scientific background to interpret. In December 2013, the Food and Drug Administration issued an order to a company called 23andMe to stop offering health-related genetic reports with their do-it-yourself genome kits []. Such kits were intended to allow any individual to determine if his or her compilation of DNA, referred to as a genome, contains particular markers, indicating a predisposition to certain diseases or conditions, such as breast cancer, Alzheimer’s, or Huntington’s. However, the FDA argues that the kit is not scientifically valid, that it can be erroneous, and that because these results are shared directly with the individual rather than through a doctor or medical professional, it can lead people to make the wrong decisions about their health. Similar kits offered by other companies, such as Navigenics and deCODEme, have been subject to the same skepticism []. What is it that these kits actually do, and is the potential for informational error as large as the FDA suspects?

Reading the genetic message

In order to answer these questions, it is crucial to understand the process of obtaining DNA, reading genetic information, and determining health risks. Many of these kits require you to spit into a tube or swab the inside of your cheek, which deposits a small number of your cells containing your DNA []. From these cells, the companies isolate your DNA and use specialized machines to read out the sequences, composed of chemical bases that are each represented as an A, T, C, or G []. Once the sequence has been obtained, the real work begins. Computers compare your genetic information to a database filled with genetic information from other individuals []. However, they don’t compare every single letter, because humans share the same letters in many regions of their DNA. Instead, they only look at specific locations where the presence of a certain letter has been correlated with a particular characteristic or, as we are focusing on here, a particular disease risk in a population. These locations are referred to as single nucleotide polymorphisms, or SNPs, which means that these locations can have different letters in different individuals []. Depending on which letter you have at that position, your DNA might encode a message for a particular disease, while your friend, who has a different letter at that position, might have a message that does not confer disease.

However, the presence of a different letter at that location does not guarantee that you will develop the disease. This is one of the contributing factors that makes the reports from these genome kits so uncertain. As a result, these kits can only tell you your disease likelihood, or risk. What are some of the reasons these types of kits can only give predictions of risk instead of guaranteeing a particular outcome?

The genetic message and disease

Firstly, your DNA has two sets of messages, rather than one, each one inherited from one of your parents (Figure 1, A and B) []. Sometimes, both sets have the same message, but many times, they do not. Both messages might contribute to the end result: if, at a location in your DNA that specifies your height, one copy has a message for you to be tall and the other copy has a message for you to be short, you might end up being average height as an adult. On the other hand, one message might win out and be the only one to contribute to the final result, a phenomenon referred to as the dominance of the message []. Dominance of a message, including one for a disease, can vary in degrees. If one message specifies that you should develop the disease and the other specifies that you will be healthy, but the healthy message dominates heavily, you might not get the disease.

What else determines how likely you are to develop a health condition? In humans, the answer to this question is complicated, because it does not solely depend on the dominance of a particular message. In past decades, much evidence has surfaced that suggests that your environment, such as the air you breathe or the habits you have developed, might be as influential as your DNA in determining how likely you are to develop a condition (Figure 1, D) []. For example, say you have one copy of a message that specifies high cholesterol, while one copy specifies normal amounts of cholesterol. As a result of the two messages, your body might contain slightly higher levels of cholesterol than your friend’s. However, this does not mean that you will develop heart disease, because the level of cholesterol in your body is also dependent on the amount that comes in from your diet. The degree to which your DNA’s message can be masked by the interactions between your body and the environment is referred to as penetrance []. The lower the penetrance of the message, the less likely the message will contribute to your body’s state. How your genes interact with the environment to produce this variability is still under debate.

To further complicate matters, your genetic information can change over time in a process called genetic mutation []. You might be born with a particular letter in one copy of a message, but that letter can spontaneously change as your cells make mistakes in maintaining your DNA or as you are exposed to harmful chemicals and radiation (Figure 1, C). Sometimes, the change in the letter is harmless, but other times, you might gain the exact change that specifies a new message in your body—possibly one for the development of a disease, particularly cancer. Although rare, because this process of change is random and unpredictable, the sequence of your DNA from three years ago might predict that you won’t get the condition, but you might still get the condition three years from today.

Lastly, most conditions are not pre-determined by a message specified by just one location in your DNA. Dozens of messages, encoded from multiple regions of your DNA, may all influence your disease risk through interactions with each other []. This is especially true of conditions like breast cancer, in which cells in breast tissue grow uncontrollably and cause damage by ultimately overwhelming healthy cells throughout the body []. Messages that control the degree of cell growth, where the growth happens, and what the responses to these cells are, all contribute to whether or not you could develop breast cancer. How many messages there are, and what messages contribute to such conditions, is still unknown.

Figure 1 ~ Your DNA encodes two copies of each message in your body (A and B), depicted in blue and green. Occasionally, a harmful mutation that can cause a disease is made, shown in orange (C). However, there is no guarantee that the message with the error will cause you to develop a disease. Depending on factors like dominance, penetrance, the environment, and interactions between the messages (D, red cross and sun), the likelihood of your developing that disease can be either large or small (E). As a result, the outcome is uncertain, and only predictions, based on statistical calculations, can be made.

Your DNA does not predict your fate

All of this points to the fact that your genetic information is not deterministic, as summarized in Figure 1. That is to say, your DNA almost never completely specifies what the final result will be. Companies like 23andMe, therefore, cannot say with absolute certainty whether or not an individual will get a particular disease. Instead, they provide a risk assessment, determined from statistical calculations of compiled data that correlates DNA sequence to a likelihood of disease outcome [9, 10]. This means that the results of these kits are statistical predictions, not unquestionable facts, and this must be kept in mind when interpreting the data that these kits yield. However, even with this caveat, these kits remain both useful and informative. While the predictions may not always be correct, the degree of information we can obtain from our DNA sequence holds much potential. As scientists and medical professionals uncover more about the millions of messages our body encodes, the hope is that one day, we will further understand genetic predisposition to diseases and ways to cure them.

Emily Low is a Ph.D. candidate in the Biological and Biomedical Sciences Program at Harvard Medical School

References

[] Perrone, M. (2013, November 25). FDA Halts Sales of 23andMe DNA Test Kits. USA Today. Retrieved from http://www.usatoday.com/story/news/nation/2013/11/25/23andme/3699329/

[] Dickinson, B. (2008, August 20). How Much Can You Learn From a Home DNA Test? Discover. Retrieved from http://discovermagazine.com/2008/sep/20-how-much-can-you-learn-from-a-home-dna-test#.UvfZF7SupbF

[] 23andMe. (Owner). 23andMe DNA Processing Lab Video [Video]. (2013, June 5). USA: 23andMe, Inc. Retrieved from https://www.youtube.com/watch?v=6cYGPlQCeLA

[] Lister Hill National Center for Biomedical Communications. What is a Gene Mutation and How Do Mutations Occur? (2014, February 3). Retrieved from

[] Lister Hill National Center for Biomedical Communications. Can a Change in the Number of Genes Affect Health and Development? (2014, February 3). Retrieved from

[] Lister Hill National Center for Biomedical Communications. What are Reduced Penetrance and Variable Expressivity? (2014, February 3). Retrieved from http://ghr.nlm.nih.gov/handbook/inheritance/penetranceexpressivity

[] Lister Hill National Center for Biomedical Communications. What Does It Mean to Have a Genetic Predisposition to a Disease? (2014, February 3). Retrieved from

[] American Cancer Society. What is Breast Cancer? (2014, January 1). Retrieved from http://www.cancer.org/cancer/breastcancer/detailedguide/breast-cancer-what-is-breast-cancer

[] Chang, E., Wu, S., Kleinman, A., and Eriksson, N. (2013, October 16). Estimating Odds Ratios for SNPs in Linkage Disequilibrium with Reported SNPs. Retrieved from https://www.23andme.com/for/scientists/.

[] Naughton, B., and Wu, S. (2011, August 25). Guidelines on Vetting Genetic Associations. Retrieved from https://www.23andme.com/for/scientists/.