NEW YORK – Bringing in rare genetic risk variant data has the potential to improve the performance of polygenic risk scores (PGS) for type 2 diabetes (T2D), according to new research by investigators at the Broad Institute, Boston Children's Hospital, Harvard Medical School, and elsewhere.
"[O]ur results suggest that adding rare variants to complex trait PGSs will be valuable, albeit through polygenic effects quite distinct from those that might have been expected from early predictions of rare variants with a high impact on common disease," co-senior and corresponding author Jason Flannick, a researcher affiliated with the Broad, Harvard, Boston Children's Hospital, and the AMP-T2D-GENES Consortium, and his colleagues wrote in Nature Genetics on Monday.
Using exome sequence data for more than 87,700 UK Biobank or AMP-T2D-GENES study participants of European, Hispanic, African American, East Asian, or South Asian ancestry, the researchers searched for individual variants or genes associated with a T2D-related blood glucose trait known as hemoglobin A1C (HbA1C), an oft-used measure for diagnosing T2D.
"HbA1C is influenced by common genetic variants that affect both pathways central to glycemic control and pathways that influence erythrocytic properties such as cell lifespan," the authors explained, noting that "[e]rythrocytic variants do not affect the risk of T2D and can in fact cause diabetes misdiagnoses by altering the expected relationship between measured HbA1C and true blood glucose levels."
The team highlighted significant HbA1C associations with more than three dozen moderate- or high-impact variants, along with rare variants that were overrepresented in the PIEZO1, GCK, and G6PD genes. These and other rare variants in the erythrocyte function-related genes PIEZO1 and G6PD appeared to influence HbA1C levels used to diagnose T2D, as well as the performance of an existing T2D PGS that is based on common genetic variants.
Expanding their view to include suspicious rare variants with more tenuous ties to HbA1C in the association analysis or to T2D in a prior exome sequencing-based analysis, the investigators put together several rare variant-based PGSs, settling on a proposed T2D PGS that included nearly 21,300 variants spanning more than 150 genes. This set was largely comprised of ultra-rare variants that turned up in three or fewer of the individuals they considered.
While the rare variant diabetes PGS showed promise for predicting T2D, the team found that an approach that brought together both rare and common PGSs for T2D had even better accuracy. The group estimated that the combined PGS could reclassify some 4.9 million misdiagnosed T2D cases in the US alone, compared to 3.4 million reclassifications with a common variant PGS and 2.4 million reclassifications with the PGS that is comprised of rare variants.
"These results suggest that rare variants, despite a comparatively modest impact on complex traits, may meaningfully contribute to genetic-based diagnostic strategies for complex disease," the authors wrote, noting that the current findings "provide a method for constructing complex trait PGSs from rare variants and suggest that rare variants will augment common variants in precision medicine approaches for common disease."