Researchers at Columbia University Irving Medical Center report that the cell-type molecular signatures associated with Alzheimer's disease are largely shared across African American, Latin and white brain donors. The study, published in Nature on 15 July 2026 by Tain Luquez, Jonathan Algoo and colleagues, with Vilas Menon as senior author, used single-nucleus RNA sequencing and ATAC sequencing (a measure of which stretches of DNA are open and available to be read) on 399 post-mortem tissue samples from 56 self-identified African American participants, 31 self-identified Latin participants (of any race) and 80 self-identified white participants. Three regions were profiled where tissue allowed: the dorsolateral prefrontal cortex, the superior temporal gyrus and the anterior caudate. The authors report microglial subgroups marked by GPNMB and CD74, astrocyte subgroups marked by SERPINH1, CD44 and WIF1, and neuronal signatures including SST-expressing inhibitory neurons and superficial-layer excitatory neurons, all associated with Alzheimer's clinical and pathological measures in a region-specific way across all three groups. All donors came from three longitudinal studies with prospective brain autopsy based at the Rush University Alzheimer's Disease Center: the Religious Orders Study and Memory and Aging Project, the Minority Aging Research Study, and Latino Core.

The finding speaks to a long-standing criticism of the field: that most single-cell brain data, biomarker thresholds and trial cohorts derive from overwhelmingly white samples. In the CLARITY AD trial of lecanemab, reported in the New England Journal of Medicine, about 20 per cent of the 1,795 participants globally were non-white, and among those randomly assigned in the United States, 4.5 per cent were Black and 22.5 per cent Hispanic. Shared cell-type signatures are an argument that mechanistic findings from one cohort may carry over to others. They are not evidence that risk, diagnosis or treatment response is the same. The Alzheimer's Association's 2026 Facts and Figures report states that older Black Americans are about twice as likely, and older Hispanic Americans about one and a half times as likely, to have Alzheimer's or another dementia as older white Americans, differences generally attributed to social, economic and cardiovascular factors rather than to different disease biology.

The authors set their own boundaries. In the earlier bioRxiv preprint of the work, posted on 15 April 2025 under the title "Cell type-specific associations with Alzheimer's Disease conserved across racial and ethnic groups", they note that while the number of samples processed is comparable to other large single-nucleus studies, fewer than 100 donors sit in each population group, so the study is powered to ask what is conserved rather than to characterise group-specific differences. The preprint also states that the analysis deliberately uses self-reported race and ethnicity rather than genetic ancestry, in order to capture effects of life experience, and reports iterative clustering of roughly 1.9 million nuclei. The published abstract differs from the preprint in detail, adding a CD44-positive astrocyte subgroup and neuronal signatures and no longer foregrounding a CR1-positive microglial subgroup, a reminder that peer review changed the emphasis.

One correction to how the work has been framed: the study covers three population groups, and East Asian donors were not among them. Nor does the design speak to sequence or cause. Post-mortem tissue captures a single moment, usually years after symptoms began, so it cannot establish whether the microglial and astrocytic states seen are drivers of decline, consequences of it, or both. Brain-bank donors are also a self-selected group, recruited through cohorts that have worked for years to build trust in communities historically under-represented in dementia research, and they are not a random sample of anyone's population.