New Microprotein Atlas Reveals 1,000+ Overlooked Brain Proteins Linked to Alzheimer's

For decades, scientists hunting for the molecular roots of Alzheimer's disease have trained their attention on genes and proteins—the big, well-catalogued players in the cell. But there exists an entire class of molecules that has slipped through the cracks of standard research: microproteins, tiny chains of amino acids that are produced from small stretches of genetic code and measure 150 amino acids or fewer. They have been nearly invisible in the databases that organize human biology, not because they are rare, but because the tools built to find proteins were never designed to look for something so small.

Researchers at the Salk Institute have now created the first comprehensive map of these overlooked molecules in the human frontal cortex, comparing brain tissue from people with Alzheimer's disease to those without. The work, published in Nature Aging, identified 1,067 microproteins that had never been formally characterized before—molecules absent from the standard protein catalogs that scientists rely on. The atlas was built by analyzing postmortem brain samples from nearly 500 individuals, including tissue from the Religious Orders Study and Memory and Aging Project cohort, using a combination of genetic sequencing, mass spectrometry, and an artificial intelligence tool called ShortStop that the Salk team developed specifically to hunt for microproteins in existing data.

The discovery exposes a fundamental blind spot in how we understand the genome. Every reference proteome—the master list of proteins a cell can make—is constructed from gene models that systematically exclude the small open reading frames from which microproteins are translated. As Alan Saghatelian, the senior researcher leading the work, explained it, the first step was simply to build a search database that could actually see these sequences and point it at the deepest human brain proteomics data available. Once the team reapplied their tools to existing datasets, the hidden landscape of microproteins came into view.

Some of these newly discovered microproteins showed different expression patterns in Alzheimer's tissue compared with healthy brain. Notably, Alzheimer's disease samples tended to display higher overall microprotein expression. The researchers then zeroed in on microglia, the brain's immune cells, which are known to malfunction as the brain ages and degenerates. They found a 63-amino-acid microprotein encoded at the MKKS locus that appeared to be the primary protein product made from that genetic location and was significantly reduced in Alzheimer's disease. When they deleted the gene responsible for making this microprotein in cultured microglia, the cells' ability to generate energy through mitochondrial respiration deteriorated. The finding suggested that this single microprotein plays a role in keeping immune cells energized and functional.

The MKKS discovery illustrates a troubling possibility: the most abundant and most relevant protein product at a given genetic location may be one that has never been annotated in any database. Saghatelian cautioned, however, that not every microprotein identified in the atlas should be assumed to be doing something biologically important. An expressed microprotein might simply be a marker—a sign that something has gone wrong with how a gene is being transcribed or spliced—without the peptide itself having any active role. Distinguishing between these two possibilities will require further investigation.

The atlas points toward a connection between microprotein dysfunction and the immune cell abnormalities that characterize Alzheimer's disease. What makes the work significant is not just the specific findings about one microprotein, but the resource itself: the atlas is now publicly available, offering researchers a new framework for studying these molecules across different brain regions, tissues, and disease contexts. The assumption that we have already discovered all the genes our cells can make, Saghatelian noted, is simply wrong. What we know continues to expand, and this atlas makes it considerably easier to investigate microproteins in aging, neurodegeneration, and potentially in diseases like Parkinson's as well.