ideasIdea
Build polygenic scores that work in every ancestry before deploying any
Genetic risk scores were built mostly on Europeans and work worse in others. Funding non-European cohorts and setting a portability standard would prevent screening that widens inequality.
Polygenic score accuracy drops substantially in African and South Asian ancestry. Propose a funding programme for diverse cohorts and a programme standard requiring PRS discrimination within 20% across ancestries before use in screening.
Hypothesis
Multi-ancestry training and standards reduce the performance gap for breast, prostate and colorectal PRS to under 20% within five years.
Rationale
The portability gap is a training-data problem; deployment without standards locks in inequity.
What would test it
Benchmark PRS annually against a held-out multi-ancestry cohort.
Maturity
early clinical
Who has to act
research
Cost to try
Large (over $50M)
Years to first evidence
5
Bottlenecks it attacks
- Inherited risk is mostly unidentified · Most people who carry a high-risk cancer gene do not know it until they or a relative gets cancer.
- Trials do not represent the people who get cancer · Older, Black, Hispanic, Asian, rural, poor and multimorbid patients are under-represented, so results may not apply to them.