In recent years, the polygenic scores (PGS) have been touted as a revolutionary tool that will transform disease prevention. By combining the effect of multiple genetic variants, they promise to identify people at higher risk of cancer, cardiovascular disease, or obesity. However, the most recent scientific evidence indicates that their predictive capacity is modest and, in many cases, generates more questions than answers.

A rigorous study debunks the myth

Researchers at University College London (UCL) analyzed 926 PGS for 310 diseases using the public database Polygenic Score Catalog. Their findings, published in October 2023 in BMJ Medicine, were overwhelming: on average, the PGS identified only the 11 % of individuals who eventually developed the disease and simultaneously generated 5 % false positives【372282977355559†L19-L52】. For specific conditions such as breast cancer or coronary artery disease, the detection rates were even lower (10 % and 12 %, respectively)【372282977355559†L67-L70】.

This means that for every correctly identified person, many more would receive a reassuring (false negative) or alarming (false positive) result without being accurate. According to the authors, applying PGS in population-based screening programs would involve performing thousands of genetic tests to prevent a single event (e.g., a heart attack) if used to guide statin prescriptions. Furthermore, the scores provide little additional information than conventional factors such as age, sex, blood pressure, or cholesterol.

Why polygenic scores fail

PGS is based on genome-wide association studies (GWAS). These studies identify variants that are more prevalent in people with a disease than in the general population. Variants typically have very small effects (increasing risk by 1 %–2 %), so millions of data points are needed to obtain robust results. Despite this, current PGS has several problems:

• Low predictive capacityAs we've seen, even for relatively heritable diseases, PGS only capture a fraction of the risk. Environmental, lifestyle, and epigenetic factors play a greater role.
• Ancestry biasMost GWAS are derived from European populations. Thus, the resulting PGS are less precise in other ethnicities, which can amplify inequalities. The new obesity PGS, for example, explains 17% of BMI variation in Europeans but only 2% in an African cohort [368994030973972†L942-L948].
• Gene-environment interactionsGenetic variants act in specific environments. A PGS may predict something in a country with a Western diet and not in one with a Mediterranean diet. Furthermore, the response to diet, exercise, or medications is modulated by many other factors.
• Statistical complexityDifferent algorithms and parameters generate different PGS. Companies can choose the score that offers the most striking results, but that doesn't mean it's the most accurate or reliable.

The same UCL study concludes that PGS should not be used for mass screening. Even when combined with traditional factors, the added benefit is small and less than offsets focusing on universal recommendations (eating healthy, quitting smoking, exercising). Professor Nicholas Wald, co-author of the paper, states that genomics is important for drug discovery, but not for predicting who will get sick. 372282977355559†L86-L107

Does this mean that genetics don't matter?

Genetics plays a key role in certain monogenic diseases (for example, mutations in BRCA1 or BRCA2). In these cases, a genetic test detects a very high risk, and clear preventive decisions can be made, such as mastectomies or drug treatments. PGS, on the other hand, offer a relative risk that barely translates into an appreciable absolute difference. For example, going from a 3 % risk of % to a 4 % risk of a disease does not change lifestyle recommendations, but it can generate unnecessary anxiety.

Even so, genomics research remains valuable. Understanding which variants influence health helps identify biological pathways and potential therapeutic targets. Furthermore, PGS can be used as adjustment variables in Mendelian epidemiology studies to infer causality. The problem arises when they are sold to the public without proper context.

Tips for interpreting the results

If you decide to take a genetic test with PGS, keep the following in mind:

• Consult with professionalsA geneticist or medical expert can help you understand the true magnitude of your risk and how to integrate it with other factors.
• Look at the absolute riskAsk what your baseline risk of the disease is and how much it varies with your score. Sometimes the differences are minimal.
• Do not modify your treatment without adviceDon't stop a medication or start a new one just because of your PGS. Medical decisions should be based on robust clinical evidence.
• Consider privacy: Read the terms and conditions regarding how the company will store and use your genetic data.
• Remember the weight of lifestyleEven if your PGS is high, a balanced diet, regular physical activity, and adequate rest can reduce your risk. If your PGS is low, it's no excuse to neglect your health.

Precision nutrition beyond PGS

PGS are only one part of personalized medicine. In Mefood Omics We use data on microbiota, metabolites, dietary habits, and activity patterns to offer comprehensive recommendations. Even without a PGS, measuring your microbiome or metabolomic profile can provide practical information about how you respond to different foods.

For example, certain combinations of gut bacteria are associated with increased energy absorption or a better response to low-carbohydrate diets. Oorenji integrates this data with coaching tools and healthy recipes (recipes.oorenji.com) to guide you in creating menus that respect your tastes and needs. And if you're interested in knowing your genetic predisposition to obesity, you can read the post 1.

Conclusion: Looking beyond genetics

Polygenic scores have provided valuable information for research, but their usefulness as a clinical or population prediction tool is limited. They do not replace healthy habits or known risk factors. Faced with the temptation to base our health on a genetic number, let's remember that we are more than our genes: our environment, upbringing, and daily decisions weigh much more heavily. PGS can complement medicine, but they should not become the sole criterion for decision-making.

References

1. Genetic risk scores not useful in predicting disease – ScienceDaily
2. Performance of polygenic risk scores in screening, prediction, and risk stratification – BMJ Medicine
3. Polygenic prediction of body mass index – Nature Medicine
4. The future of polygenic risk scores in direct-to-consumer genomics – Nature Genetics