Multiple studies have associated rare variants in SORL1 with an increased risk of Alzheimer’s disease. To achieve enough statistical power to detect an association—individual variants being quite rare, with many found in only a single carrier thus far—these studies analyzed sets of variants grouped according to the likelihood that they disrupt gene or protein function, often combined with the frequency of the variant in the study sample or population databases.

The figure below illustrates key findings from these studies. Results shown met each study’s criterion for statistical significance. See the cited studies for methods and additional findings.

• high-priority
• loss-of-function
• strictly-damaging
• CADD > 30
• Revel >0.75
• Revel >0.50
• Revel >0.25

High priority: Variants located at a position designated by Andersen et al. as high priority, plus variants in the VPS10P and 10CC domains with REVEL scores > 0.5. Loss-of-function: nonsense, frameshift, and splicing; referred to as “disruptive” by Nicolas et al., 2016 and as “protein-truncating” by Bellenguez et al., 2017; Campion et al., 2019; and Holstege et al., 2023. Strictly damaging: predicted deleterious by SIFT, PolyPhen-2, and MutationTaster; referred to as “Mis3” by Campion et al., 2019. Risk allele frequencies from ExAC (Holstege et al., 2017), ExAC or ADES-FR (Alzheimer’s Disease Exome Sequencing Project-France) (Bellenguez et al., 2017), ExAC and study sample (Raghaven et al., 2018), gnomAD (Campion et al., 2019; Holstege et al., 2022), gnomAD or study sample (Holstege et al., 2023), and study sample (Nicolas et al., 2016).