Alzheimer's is a complicated disease.
There are early-onset forms and late-onset forms that have different genes involved. The risk to develop Alzheimer's as a result of genetic predisposition varies from person to person and family to family, depending on which form is going on and which gene or genes are involved.
We don't know and understand all the factors that cause Alzheimer's yet, so while genetic testing can be helpful, it oftentimes won't be able to tell the whole story.
If you have raw genomic data you've downloaded from a consumer DNA test (like Ancestry, Family Tree DNA, MyHeritage, or 23andMe), I want you to know some of the limitations that have come to our attention.
There are known miscalls (testing errors) happening for some of the Alzheimer's-associated DNA markers. Ancestry's chip (v2) is miscalling the APOE e4 variant as normal in those who actually are positive, for example. 23andMe (v5) seems to be having the opposite issue with falsely miscalling a SNP as positive in one of the early-onset genes.
Also, the markers included on testing at consumer companies change over time, so if you try to run an analysis on your raw data (by using Promethease, as an example), your report can change from one file to the next and vary over time. You might see things pop on one report that don't appear on another, depending on the testing company or the version of testing that was used.
My advice when people ask about results on Alzheimer's genes and the results they get back from testing at home is this: talk with a genetic counselor (like myself or someone else you find in your area) to make sure you are understanding the results, and always follow up a home-based test with a clinical test before believing them as "real".
Interested in reading more? Here are two other posts you can read about Alzheimer's and whether to find out if you might have a genetic risk: