When it comes to innovating genetic testing, 23andme is always at the forefront of science. Not only were they one of the first companies on the genetic testing scene, but they were one of the first to get their FDA-approved breast cancer screening test. The company is partially owned by Google Ventures, the start-up investment wing of Google. This gives 23andme a technological advantage over other companies, and they are often the first in the market with a new offer. Recently, 23andMe began offering a polygenic risk score for diabetes, offering another thing to differentiate it in the 23andMe vs ancestor battle and mark it as a leader in both ancestor tests and health tests.
What are Polygenic Risk Scores?
The standard model for determining the risk of genetic disease is to analyze the genome for individual mutations that are known to contribute to the disease. This method is quite accurate, but it has a significant weakness. Most traits are not determined by a single gene. Rather, they are known as polygenic traits, because any number of the 20,000 + genes that humans possess can modify and modify them.
To measure these polygenic traits, scientists need much more data. This is because they not only have to compare a gene with a trait, but they also have to see the effects of that gene with many other combinations within the population. Companies like 23andMe have now assimilated huge amounts of genetic data, and have the ability to study the relationships between many different genes.
23andMe claims that their polygenic risk score for diabetes measures over 1,244 distinct positions within the genome, each of which has some influence on diabetes. Although it sounds great, this is actually just a tiny fraction of the DNA available. This is typical of SNP tests, which only test individual locations within DNA for clues about the genes they carry.
With the 23andMe report, you will be able to see how your genes can affect your risk of developing diabetes. The report simply separates you into one of several risk categories, in part because the science behind polygenic risk scores is incredibly complex. The report, however, is limited by FDA regulations, and can only provide “general welfare ” suggestions.
Although this may be useful information, you should not make bets on conclusions. In fact, the bathroom scale is just as predictive, because your weight is directly related to the risk of diabetes. In fact, simply eating a heavily processed “Western diet” can significantly increase your chances of getting diabetes because nearly two-thirds of Americans are overweight. Knowing this fact is often more useful than worrying about your genetic disposition.
Are Polygenic Risk Scores Accurate?
According to scientists studying genetics, with a fairly large sample complex size relationships between genes could be accurately predicted. While the authors have suggested more than 2,000 subjects for accuracy, 23andMe uses more than 70,000 samples to make predictions, as well as a control group of millions. This allows you to see in a very complete way how the different genes interact.
The problem is that even genes that are polygenic are not completely determined by your genetics. Your lifestyle, eating habits and other environmental factors can drastically affect your health. For example, one researcher speculated that only 17% of breast cancer could be detected early with polygenic risk scores. This means that the other 83% of cases of breast cancer would not be detectable with technology, probably because of environmental conditions.
So, for diseases such as breast cancer and diabetes, it seems that lifestyle is much more predictive than genes. If you have the money and you’re curious, go ahead. It can’t hurt knowing your genetic risk factors. But also know that polygenic risk scores are a relatively new science. When scientists get more data they can find that they can be more predictive. Or, they might find that genetics is much more complex than they imagined. Until then, paying for a polygenic risk score is just a shot in the dark.