Background:Rare variants in the SORL1 gene have been associated with increased risk of Alzheimer′s disease (AD). While protein−truncating variants (PTVs) are observed almost exclusively in AD patients, most variants are rare missense variants that can be benign, risk−increasing, and recent reports have indicated that some variants are causative for disease. However, since SORL1 is currently not considered an autosomal dominant Alzheimer Disease gene (ADAD), segregation analyses are not performed, which complicates the identification of additional clinically important missense variants.Methods:We prioritized highly conserved and functionally relevant SORL1 missense variants by considering the functional effects of homologous variants on proteins that share domains with SORL1 (domain-mapping of disease mutations, DMDM) into. We used this variant prioritization approach to annotate SORL1 variants identified in a previously assembled exome sequencing dataset encompassing 18,959 AD cases and 21,893 non-demented controls, and we tested the effect of high, moderate, low and no priority missense variants and specific variant subtypes on disease risk and age at onset.Results:High priority missense variants (HPV) associated with a 6.4−fold increased risk of AD (95%CI: 4.3 − 9.7, p=2.1×10−24), which concentrated on early onset AD (OREOAD 10.5, 95%CI: 6.8 − 16.3, p=3.0×10−29) vs. late onset AD (ORLOAD=4.5, 95%CI 2.85 − 6.94; p=4.9×10-11). The median age at onset of HPV carriers was >8−years earlier than carriers of wild-type SORL1. Intriguingly, specific subtypes of HPVs, including those affecting residues in the YWTD-motif or the calcium cage, occurred only in AD cases and carriers of these variants had an earlier age at onset compared to carriers of PTVs, indicative of a dominant negative effect. Carriers of other HPVs had an age at onset that overlapped with carriers of PTVs, suggesting they lead to haploinsufficiency. Yet other variants had a slightly later age at onset than PTVs, suggesting that their effect on SORL1 function was milder than losing a copy. Variants annotated as moderate, low and no priority did not have an effect on AD.Conclusions:Next to carriers of SORL1 PTVs, carriers of selected missense variants should be considered for segregation analyses, which will likely provide evidence for autosomal dominant inheritance for additional SORL1 missense variants.