How can we better predict someone's risk of developing a condition?

Predicting someone’s risk for a condition has been revolutionary for clinical decision-making. New research by Bradley Jermy and colleagues uses data from the 100,000 Genomes Project to look at factors affecting the accuracy of our current prediction tools, and how we could improve them.

Can genetic changes cause disease?

Many conditions can be caused, or partially caused, by changes in a person’s genes.

This means that the genetic changes people have can influence their chance of developing a certain condition, for example heart disease or cancer.

Some genetic changes are associated with a higher risk of developing disease, whereas others are associated with a lower risk.

This means that someone’s overall genetic risk for a condition depends on the combination of genetic changes they have.

You can read more about genetic changes in a previous blog.

What is a polygenic risk score?

A polygenic risk score uses information from someone’s genetic make-up to estimate their overall risk for a specific condition.

It combines the individual risk associated with each genetic change to estimate their overall effect.

This estimate can account for thousands of genetic changes that have a small effect, but also rare genetic changes that have much larger effects.

Why are polygenic risk scores useful?

Polygenic risk scores can help to allow the early diagnosis or even prevention of conditions.

Currently, there are several ways or ‘risk calculators’ that are already used in the clinic to estimate a person’s risk of developing a condition. However, these calculators tend to estimate risk over a short period of time, typically the next 5-10 years.

Polygenic risk scores, however, provide an estimate that is relevant for life.

They can be calculated from birth, meaning that younger individuals not typically targeted by disease calculators can receive a life-long estimation.

This gives polygenic risk scores the potential to revolutionise disease prevention by identifying high-risk individuals from early on in life.

What are the limitations of polygenic risk scores?

There are various factors that could impact how accurate a polygenic risk score is.

Common factors that affect disease-risk, for example age or sex, can often interfere with the accuracy of estimates.

If polygenic risk scores are to be used to make decisions in the clinic, it is vital that we account for these factors to make estimates more accurate.

What did this study do?

Recent research by Bradley Jermy and colleagues has provided a new method of estimating disease risk.

Their method would allow us to account for age, sex and geographical location when predicting someone’s risk of developing a condition, making the prediction more accurate.

How did the researchers do this?

The researchers focused on 18 different conditions, all selected due to their high impact in the Global Burden of Disease 2019 Study.

In total, they examined genomic and clinical data from 1.2 million participants from 7 different projects, spanning 4 different countries.

Amongst the projects used was the 100,000 Genomes Project, a landmark research study led by Genomics England in partnership with the NHS.

The researchers came up with a new approach to predict someone’s risk of developing a condition.

When used in the clinic, this new approach would provide each patient with a polygenic risk score specific to their age, sex, and the country they are from.

Why should we consider age and sex in polygenic risk scores?

In the study, researchers identified 2 key findings showing that age and sex are vital to account for when accurately predicting conditions.

1) Polygenic risk scores are less accurate for older individuals than younger individuals.

As we age, the influence of environmental factors becomes more prominent in causing disease. This, in effect, reduces the influence of genetic factors, making genetic-based predictions less accurate.

2) Some genes are only associated with conditions in one sex.

Researchers identified 5 conditions where faulty genes caused a condition in one sex, but not in the other. This could be explained by biological differences, or by unbalanced participation of men and women in research studies.

Using the new approach developed by this study, we could account for these external factors to produce a more accurate, personalised estimate, allowing for more effective prediction and prevention of conditions.

What does this mean for patients?

Estimating the likelihood that a patient will develop a condition creates better opportunities for us to detect and treat it.

This research offers a method that would provide sex, age and country-specific risk predictions for several different conditions.

This would allow patients to receive a personalised prediction from birth, providing more detailed information about their health, and in turn helping inform their healthcare.

For example, patients at high risk for a certain condition may be advised to undergo regular screening, adjust their lifestyle habits to reduce the risk, or even receive preventative treatment, all according to their personalised prediction.

The future of polygenic risk scores

Polygenic risk scores could be transformative in healthcare if used alongside other data to help inform decisions.

However, further research is important to ensure that polygenic risk scores are accurate and beneficial for everyone.

This study demonstrates the need to account for age and sex, but further research is essential so that scores may be accurate across diverse ancestries.

Understanding how to optimise the accuracy of estimates is a key step towards their widespread use in disease prediction and treatment.