Introduction

Genome-wide association studies (GWAS) have resulted in the discovery of tens of thousands of genetic associations for various traits and phenotypes. Polygenic risk scores1, innovative drug discovery2, and gene-editing3 have all been enhanced, or even based on, GWAS results. Genome-wide association studies investigate the association of individual single nucleotide variants (SNVs) on a phenotype of interest (for example coronary artery diseases)4. Most GWAS identify SNVs with, individually, small effects4. This supports the notion that most diseases are polygenic, rather than monogenic, in nature5.

To observe the small effect of individual SNVs, GWAS have relied on increasingly larger sample sizes4. Recent advances have seen rapidly increasing sample sizes, particularly with the establishment of large biobanks. The most widely used and analyzed biobank in human genetics is the UK Biobank (UKBB)6. Analyses done in the UKBB and other similar biobanks have the opportunity not only to identify new associations but also to replicate previously proposed associations that arose from other GWAS investigations. It is not unexpected that some SNVs that were considered to be associated with a phenotype in an earlier GWAS may not be replicated in a subsequent GWAS. Even if they are replicated, their effect size may change, e.g. because of the winner’s curse phenomenon12,13,,13. With improved phenoty**, it seems plausible that these scores will continue to improve. Nevertheless, in the meantime there may be other ways to enhance current binary GWAS results for polygenic risk scores. First, our results clearly show a superior replication rate with quantitative phenotypes. These quantitative phenotypes are often more in line with physiological processes (e.g. systolic blood pressure) than clinical diseases (e.g. coronary artery disease). As such, future GWAS that directly use metabolomic data as outcomes (such as protein expression) are likely to, similarly, have higher accuracy than clinical disease phenotypes. Future research merging metabolomic outcomes and GWAS may be a useful addition to our scientific knowledge. For instance, some evidence suggests that the use of ‘intermediate’ phenotypes—between the genotype and the disease-based phenotype—may improve disease prediction14. For example, a 2021 study showed that the integration of polygenic risk scores for both disease-associated biomarkers and polygenic risk scores for the disease itself showed enhanced prediction over the polygenic risk score for the disease exclusively14. Second, almost all SNVs for binary traits with an OR > / = 1.2 were replicated, whereas the majority of SNVs with an OR below 1.2 were not replicated and this may reflect lack of power in the replication dataset. Of note, many of the replication UKBB datasets that we considered here did not use the full UKBB data, and power is likely to improve as complete biobank data are used and many biobanks are combined.

Limitations in comparison to previous literature

We were surprised to find only nine phenotypes where two GWAS had been conducted in truly independent participants and where inclusion or not of UKBB data was a distinguishing feature. It is plausible that further independent GWAS on the same traits exist, although this seems unlikely given the thorough and systematic search we performed of the GWAS atlas8. It is, however, likely that more GWAS are available, but they contain overlap** samples between GWAS (i.e. two GWAS of the same phenotype are not truly independent as they contain similar cohorts of participants), aren’t of sufficient quality to be included in the GWAS Atlas, are conducted in a non-European population, or have not made their summary statistics available. An earlier study15 reports building a model for SNV replication using GWAS for over 50 phenotypes, although it is unclear what, if any, measures were taken to determine if these numerous GWAS were truly independent i.e. did not include overlap** participants. Also, this study validated their model in two, small GWAS of one trait. Furthermore, this study didn’t actually quantify a SNV replication rate, nor did they stratify their results by binary and quantitative phenotypes. A further limitation of our study is that we didn’t include other SNV features, ideally we would have liked to include, for instance, LD as predictors in our model. However, this data was sparsely available. Lastly, it should be acknowledged that large disease-specific consortiums generally qualitatively describe the replication of SNVs as their consortium increases. Our study quantifies this formally and, importantly, quantifies replication across more than one phenotype.

Future research

We have identified a number of future research priorities. First, improving the phenoty** of binary phenotypes seems to be a priority for GWAS. Second, to facilitate an assessment of SNV replication, future independent cohorts are likely required. Many efforts to do this are already underway (e.g. AllofUs cohort and Millions Veteran Program).

Conclusions

The replication of SNVs discovered from GWAS was high for quantitative phenotypes. Genome-wide Association Studies appear to be entirely sufficient to detect SNVs associated with quantitative traits. For binary traits, however, the replication rate is modest. We have built a simple prediction model that can accurately ascertain SNV replication in later GWAS. It may be of use for researchers and clinicians that utilize GWAS results.