Translating polygenic risk scores for clinical use by estimating the confidence bounds of risk prediction

Abstract

A promise of genomics in precision medicine is to provide individualized genetic risk predictions. Polygenic risk scores (PRS), computed by aggregating effects from many genomic variants, have been developed as a useful tool in complex disease research. However, the application of PRS as a tool for predicting an individual’s disease susceptibility in a clinical setting is challenging because PRS typically provide a relative measure of risk evaluated at the level of a group of people but not at individual level. Here, we introduce a machine-learning technique, Mondrian Cross-Conformal Prediction (MCCP), to estimate the confidence bounds of PRS-to-disease-risk prediction. MCCP can report disease status conditional probability value for each individual and give a prediction at a desired error level. Moreover, with a user-defined prediction error rate, MCCP can estimate the proportion of sample (coverage) with a correct prediction.

Rank concordance of polygenic indices: Implications for personalised intervention and gene-environment interplay

“…Our findings are of crucial importance now that PGIs are becoming increasingly accessible to physicians, consumers, and applied researchers 19 . We complement recent work that showed that an individual’s PGI can span several deciles when the uncertainty of GWAS estimates are taken into account during PGI construction 45,46 . While the source of uncertainty emphasised in these papers does not derive from the construction method or GWAS discovery sample per se, we draw a similar conclusion: ranking individuals on basis of their position in a PGI distribution is prone to large uncertainty.…”

Association of a Multiancestry Genome-Wide Blood Pressure Polygenic Risk Score With Adverse Cardiovascular Events

“…However, there is an unmet need to assess the potential benefits (reduction of years lost due to disability from ASCVD events) and harms (anxiety and declined adherence to healthy lifestyle in low genetic risk individuals) of early disclosure of the genetic risk for cardiovascular events, through randomized controlled clinical trials in younger adults prior to the onset of ASCVD events. 26,27…”

Polygenic risk scores for the prediction of common cancers in East Asians: A population-based prospective cohort study

“…In addition, group-wise estimates, which arbitrarily classify the top 10%, 5%, or 1% of samples as the at-risk group, are not optimal for decisions at the individual level [34]. Emerging new methodologies that estimate probability values for hypothetically assigning an individual as at risk or not at risk, thus providing individuals with more clarity, may help to overcome this limitation [35]. At this point, PRS may not have yet reached the standards as a clinical tool by itself.…”

“…All our study participants were residents of government housing flats, which were built to accommodate approximately 86% of the resident population in Singapore during the enrolment period. A total of 63 257 individuals (35,298 women and 27,959 men) provided written informed consent [19]. The study was approved by the Institutional Review Boards of the National University of Singapore, University of Pittsburgh, and the Agency for Science, Technology and Research (A*STAR, reference number 2022-042).…”