Personality’s Genetic Signal Points to the Brain

TL;DR

• All ten strongest associations in the current 265-variant GFP set have AlphaGenome predictions in neural contexts.
• BMAL1 gives the leading region an experimental connection to learning, memory, and synaptic plasticity.
• The remaining hits nominate changes in RNA production or assembly in brain tissue and neural cells.

AlphaGenome gives the ten strongest GFP associations a plausible neural connection.

A personality questionnaire begins with judgments about ordinary life: how sociable you are, how reliably you finish things, how easily you become anxious. A genome-wide association study asks whether differences in those answers track differences in DNA. Follow the strongest associations in our current General Factor of Personality analysis through AlphaGenome, and something concrete emerges: all ten have predicted molecular effects in brain tissue or neural cells.

That is the central result. The statistics identify the variants; the sequence model supplies a plausible biological connection. At the strongest end of the list, that connection reaches a gene experimentally involved in learning, memory, and synaptic plasticity. Elsewhere, it takes the form of predicted changes in RNA production or assembly in particular neural contexts. The abstract personality score begins to acquire a cellular description. The path from words describing character to statistical dimensions now extends into molecular biology.

From personality measurements to DNA #

The starting point is a constructed General Factor of Personality, or GFP. The question behind it follows from the shared structure of personality traits: how much of their variation can be described along a common direction? Our analysis combines results from the five ReGPC 2026 personality GWAS, using fixed literature-derived weights and reversing neuroticism. In ordinary terms, the underlying studies connect alleles to personality measurements; the construction combines those associations into a shared statistical direction. The weights were fixed before this molecular interpretation. Brain biology did not determine which variants occupied the top ten. Construction method.

Here, “top ten” means the ten strongest associations among the project’s 265 retained variants. Their association P values range from approximately 2.6 × 10⁻²⁷ to 5.8 × 10⁻²¹. Taking that list in order matters: it keeps the biological story anchored to the statistical ranking. Association and allele audit.

AlphaGenome adds a different kind of information. It uses DNA sequence to predict molecular measurements, including gene expression and RNA splicing, across tissue and cell contexts. Comparing the two alleles at a variant can therefore nominate changes in how much RNA a cell produces or how it joins an RNA transcript together. These are molecular steps through which a DNA difference could influence cell function. Avsec and colleagues, 2026.

Applied to these ten GFP hits, that process returns a neural connection for every entry. “Connection” has a precise meaning here: a predicted RNA or splicing effect in a neural context. Every one of the leading associations therefore has something specific to investigate in the brain.

What connects each of the top ten hits to the brain? #

The complete list shows what that means. All entries below come from the top-ten molecular observations; effects are model predictions.

BMAL1 connects the leading hits to learning and memory #

BMAL1 gives the argument its strongest biological anchor. Hits #1 and #3 connect to it through predicted RNA or splicing effects. This gene helps run the cellular clock, including clocks within brain circuits. Its connection to cognition has been tested directly: deleting Bmal1 from excitatory forebrain neurons in mice impaired learning and recall while leaving the central circadian pacemaker intact. Local timing within those circuits mattered for memory. Price and colleagues, 2016.

A later study brought BMAL1 directly to the synapse. Researchers found that phosphorylated BMAL1 rhythmically localizes to hippocampal synapses and helps regulate signaling involved in long-term potentiation, a lasting strengthening of synaptic transmission. This provides a concrete link between a molecular clock component and the machinery through which neural connections change. Barone and colleagues, 2023.

The GFP variants predict small molecular changes, and the nominated BMAL1 junction in the follow-up query changes only slightly. The experiments establish why BMAL1 is a compelling brain target; AlphaGenome nominates it as one possible route from these particular associations to neural biology. Alternative targets remain visible in the table, including the noncoding RNA that produces the largest CNS RNA score for hit #1.

Hits #2 and #9 occupy the same broader BMAL1/RASSF10/PTH neighborhood. Their neural predictions extend across frontal cortex, cerebellum, neural progenitors, and neuronal stem cells. Together, the four chromosome 11 entries give the leading part of the ranking a recurring connection to this region. They are four associated variants in a shared neighborhood, so the natural biological unit to investigate is the region and its transcripts.

Neural RNA gives the other hits a molecular interpretation #

The chromosome 8 hits broaden the story from the clock to RNA processing. Hits #4 and #5 both nominate changes in XKR6 transcript assembly, in motor neurons and prefrontal cortex respectively. A splicing prediction concerns how pieces of RNA are joined. It gives a more specific molecular hypothesis than a gene name alone: the allele could change the transcript a neural cell assembles. These variants also produce predictions involving noncoding RNAs in neural stem or progenitor cells and Purkinje cells. The developing cell and the mature neuron both appear in the molecular follow-up.

Hit #6 supplies another such example. The variant lies within BLK, and AlphaGenome predicts both an RNA change in glutamatergic neurons and a splicing change in motor neurons. BLK also has experimentally established immune functions, so the neural prediction adds a cellular context to a gene with functions elsewhere. For this essay’s narrow question, the useful finding is the explicit neuronal output associated with changing the allele. Top-ten molecular observations.

Hits #7 and #8 illustrate why this approach is useful even when it supplies no familiar neuroscience headline. For #7, the strongest CNS RNA target is CLDN23 in Purkinje cells; the nearest coding gene is PPP1R3B. For #8, the leading predictions involve pseudogene transcripts in anterior cingulate cortex and Purkinje cells. In both cases, the output identifies a transcript and a neural context that a nearest-coding-gene label would miss. Their downstream functions are unresolved, but their contribution to the ten-of-ten result is concrete: changing the associated allele changes the model’s predicted neural RNA output.

Hit #10 makes the RNA-assembly story especially tangible. It lies within MFHAS1. A follow-up AlphaGenome query predicted an increase in one nominated RNA junction in motor neurons, cerebellar hemisphere, and two prefrontal-cortex assays. In one prefrontal assay, the predicted signal rose from approximately 0.000850 to 0.000937: about 10% on a very small baseline. These are model units, not a measurement of protein abundance. The valuable feature is the specificity: a particular connection between RNA boundaries moves in the same direction across all four returned neural tracks. Nominated-junction predictions.

The statistical signal acquires a cellular description #

The resulting picture is broader than a list of famous brain genes. It includes a clock protein with experimental links to cognition, predicted transcript changes in neurons, and noncoding RNA outputs in brain regions. Each entry adds a molecular hypothesis to an association that began with personality measurements.

That is why the result matters. The statistical construction starts with how people describe their behavior. Its strongest retained associations can then be followed into predictions about what neural cells produce and how they assemble it. The evidence is most developed around BMAL1, but the molecular follow-up reaches every hit in the top ten.

AlphaGenome gives the GFP signal a plausible neural interpretation all the way down the list. It turns an abstract ranking into named transcripts, cell types, and molecular changes. A signal extracted from personality measurements now has places to look inside the brain.

Method note: This analysis uses the current ReGPC 2026 literature-weighted GFP construction, restricted through the project’s AADR intersection and LD-clumping procedure. The top ten are ranked among 265 retained variants, rather than all possible full-GWAS leads. Neural predictions were inspected within a pre-existing CNS track classification; demonstrating that these effects are unusually common or strong would require a matched comparison set. The clustered variants do not establish ten independent mechanisms. Signed effects in the underlying analysis are oriented to the GFP-increasing allele. The follow-up splicing results are additional model queries, not laboratory replication.

Sources #

1. GFP construction method, ReGPC 2026 personality suite.
2. Top-ten association and allele audit, with all returned observations.
3. Top-ten molecular observations, September 9, 2026, and nominated-junction predictions.
4. Avsec, Ž., Latysheva, N., Cheng, J., et al. “Advancing regulatory variant effect prediction with AlphaGenome.” Nature 649 (2026): 1206–1218.
5. Price, K. H., et al. “Modulation of learning and memory by the targeted deletion of the circadian clock gene Bmal1 in forebrain circuits.” Behavioural Brain Research 308 (2016): 222–235.
6. Barone, I., et al. “Synaptic BMAL1 phosphorylation controls circadian hippocampal plasticity.” Science Advances 9 (2023): eadj1010.
7. “Mimicry of Pre–B Cell Receptor Signaling by Activation of the Tyrosine Kinase Blk.” Journal of Experimental Medicine (2003).