Genetic testing cannot tell teachers anything useful about an individual pupil鈥檚 educational attainment, as some are claiming. That is the conclusion of a study that looked at how well so-called polygenic scores for education predict a person鈥檚 educational achievements, based on a long-term study of thousands of people in the UK.
鈥淪ome people with a very low genetic score are very high performers at age 16. Some are even in the top 3 per cent,鈥� says Tim Morris at the University of Bristol, UK. 鈥淵ou just cannot make an accurate prediction for any one child.鈥�
And while Morris expects the accuracy of polygenic scores for educational attainment to improve, he doesn鈥檛 think they will ever be good enough to predict how well an individual will do.
Advertisement
Even relatively simple traits such as height are influenced by thousands of genetic variants, each of which may only have a tiny effect. Polygenic scores sum up all these small effects to try to work out the overall impact of all the variants in one person鈥檚 genome.
Educational attainment
It is claimed that polygenic scores can be used to make all kinds of useful predictions, such as how likely a person is to develop various diseases. One company is offering embryo screening based on polygenic disease risk scores.
Some researchers 鈥� notably Robert Plomin of King鈥檚 College London 鈥� think that schools should start using polygenic scores for educational attainment. In most cases we don鈥檛 know why particular gene variants are linked to academic achievement, but the scores may reflect traits such as persistence as well as intelligence.
鈥淭here鈥檚 so much we can do with this,鈥� says Plomin. For instance, he says children could be tested before they start school to identify and help those who are likely to struggle academically.
To assess the usefulness of polygenic scores, Morris and his colleagues calculated them for 8000 people in Bristol who are part of . The participants鈥� genomes have been sequenced and their academic results are available to researchers.
Among other things, the team found a correlation of 0.4 between a person鈥檚 polygenic score and their GCSE results at age 16 (where 1 is a perfect correlation and 0 means no correlation). But there would need to be a correlation of at least 0.8 to make useful predictions about individuals, says Morris.
Plomin, however, argues that the results support his stance. 鈥淸A correlation of 0.4] makes it the strongest polygenic predictor in the behavioural sciences,鈥� says Plomin, who says this matches his own results. 鈥淚t鈥檚 so much stronger than a lot of other things we base decisions on. So it鈥檚 a very big finding.鈥�
Morris says schools already have access to other predictors that are more accurate, such as a pupil鈥檚 earlier tests results. Looking at parents鈥� educational attainments and their socioeconomic status is also a better predictor of a pupil鈥檚 academic results than looking at their genome, his results show. Providing teachers with an extra predictor based on genetics would just confuse matters, says Morris, and the cost cannot be justified.
Reference:
