Francisco Torres

Background: Alzheimer’s Disease (AD) is the most common form of dementia, a collection of syndromes characterized by cognitive, behavioral, and functional deficits. Pathological hallmarks of AD include amyloid -plaques (A) and tau protein neurofibrillary tangles (NFTs).[i]^,[ii],[iii] Although there are two main categories of AD, Familial Alzheimer’s Disease (FAD) and Spontaneous Alzheimer’s Disease (SAD), this review focuses on the latter. One gene associated with SAD is TREM2, which codes for both a membrane protein (TREM2) and soluble protein (sTREM2). TREM2 has been studied for its diagnostic and therapeutic value in SAD.^iv Traditionally, methods like genome-wide association studies (GWAS) have been used to identify relationships between TREM2 and other important genes.^iv,v However, because the boom in artificial intelligence (AI) and machine learning (ML) may allow scientists to uncover new insights related to SAD, it’s worth evaluating how valuable these new tools actually are.^v

Objective: The objective of this review was to explore how ML methods compare to traditional methods like GWAS in SAD research.

Search Methods: A PubMed Database search was conducted per the MSGR guidelines. The keywords “Alzheimer’s Disease,” “artificial intelligence,” and “machine learning” were used.

Results: GWAS identified a relationship between sTREM2 and the MS4A gene region. No associations between sTREM2 and other genes, like APOE, were found.^iv MS4A and TREM2 are mechanistically related via DNAX-activating protein of 12 kDa (DAP12).^vi MS4A4A-MS4A6A complex prevents DAP12 and TREM2 from interacting.[iv] ML methods, however, suggest a more complex relationship involving APOE and MS4A.^v Seeding experiments in a modified 5xFAD mouse model demonstrated that TREM2 knockout in mice worsens A plaque formation and may involve lysosome activity in microglia.[v] Another study demonstrated that that TREM2 knock-out in P301S mice expressing APOE4 (TE4-T2KO) had increased levels of tau protein-dependent neurodegeneration compared to the same mouse models without TREM2 knockout (TE4), although TREM2-independent factors are also involved.[vi]

Conclusions: In this review, ML identified a relationship between TREM2 and APOE. The mechanism underlying this relationship is in the process of being uncovered, as illustrated by the studies described above. Ultimately, this review suggests that ML methods compliment traditional methods, like GWAS, when seeking to uncover relationships between genes involved in SAD.

Work Cited

[i] Zhang, J., Zhang, Y., Wang, J. et al. Recent advances in Alzheimer’s disease: mechanisms, clinical trials and new drug development strategies. Sig Transduct Target Ther 9, 211 (2024). https://doi.org/10.1038/s41392-024-01911-3

[ii] Frisoni, G., Hansson, O., Nichols, E., et al. New landscape of the diagnosis of Alzheimer’s disease, The Lancet, Volume 406, Issue 10510, (2025) Pages 1389-1407. https://doi.org/10.1016/S0140-6736(25)01294-2.(https://www.sciencedirect.com/science/article/pii/S0140673625012942)

[iii] Frisoni, Giovanni B. et al. Alzheimer’s disease outlook: controversies and future directions. The Lancet. 2025; 406:10510. 1424-1442. Published 2025 Sep 22. Doi: 10.1016/S0140-6736(25)01389-3.

[iv] Rosner D, Sun J, Cacace R, et al. The Alzheimer’s disease risk genes MS4A4A and MS4A6A cooperate to negatively regulate TREM2 and microglia states. Neuron. 2026;114:1-14. doi:10.1016/j.neuron.2025.11.022

[v] Kaji S, Berghoff SA, Spieth L, et al. Apolipoprotein E aggregation in microglia initiates Alzheimer’s disease pathology by seeding β-amyloidosis. Immunity. 2024;57(11):2651-2668.e12. doi:10.1016/j.immuni.2024.09.014

[vi] Maud Gratuze, Johannes C.M. Schlachetzki, Ricardo D’Oliveira Albanus, Nimansha Jain, Brenna Novotny, Logan Brase, Lea Rodriguez, Clayton Mansel, Michal Kipnis, Sydney O’Brien, Martina P. Pasillas, Choonghee Lee, Melissa Manis, Marco Colonna, Oscar Harari, Christopher K. Glass, Jason D. Ulrich, David M. Holtzman, TREM2-independent microgliosis promotes tau-mediated neurodegeneration in the presence of ApoE4, Neuron, Volume 111, Issue 2, 2023, Pages 202-219.e7, ISSN 0896-6273, https://doi.org/10.1016/j.neuron.2022.10.022.