Genomic Prediction in the Big Data Era
By Gustavo de los Campos, Daniel Gianola
A simple model from the early 20th century remains our best tool for using DNA to predict disease risk and other complex traits.
This Article From Issue
September-October 2023
Volume 111, Number 5
Page 286
In recent years, more than 30 million genomes have been genotyped by companies such as 23andme or MyMedLab. These companies provide customers with ancestry and health-related information, and link genotype data with measurements (called phenotypes) collected from surveys, wearables, and electronic health records. The resulting datasets are routinely used for genomics research. Likewise, several public initiatives have developed large biomedical datasets, such as the UK Biobank and the All of Us program, comprising DNA and extensive phenotypic attributes from hundreds of thousands of participants.
QUICK TAKE
- In 1918, R. A. Fisher proposed a model for hereditary genetics based on Gregor Mendel’s laws of inheritance. It became the basis for predicting phenotypes as the field of genetics developed.
- Although this model does not accurately reflect causal relationships between genotypes and phenotypes, it can predict phenotypes well and is widely used in agriculture and health care.
- The era of big data in genomics has improved the accuracy of DNA-based model predictions of phenotypes. Still, the potential that these technologies offer has not yet been reached.
American Scientist Comments and Discussion
To discuss our articles or comment on them, please share them and tag American Scientist on social media platforms. Here are links to our profiles on Twitter, Facebook, and LinkedIn.
If we re-share your post, we will moderate comments/discussion following our comments policy.