
Researchers developed an overlap-adjusted polygenic risk score (OA PRS) to address bias caused by sample overlap between Alzheimer’s datasets like IGAP and ADNI. OA PRS significantly reduced inflated risk prediction (AUROC dropped from 0.915 to 0.726), aligning with results from non-overlapping data. This method enhances accuracy and prevents overfitting in Alzheimer’s genetic studies.
Like
Save
Share