Atypical energy-related symptoms define biologically distinct subtypes of major depressive disorder

Major depressive disorder (MDD) is one of the most pervasive challenges in global mental health. It remains a notoriously difficult target for precision medicine. Currently, the medical community treats MDD as a relatively unitary diagnosis. This approach groups together individuals who exhibit vastly different physical realities. For instance, one patient may struggle with debilitating insomnia and rapid weight loss. Another may suffer from excessive sleep and significant weight gain. Because current diagnostic criteria treat these opposite symptom directions as essentially equivalent, the underlying biological drivers are often washed out in large-scale clinical studies.

A new study by Harder et al. suggests that these "opposite" symptoms are not just clinical variations. They may actually indicate fundamentally different biological subtypes. By analyzing massive genetic datasets, the researchers demonstrate that depression characterized by hypersomnia and weight gain (AERS+) possesses a distinct genetic architecture. This architecture is tied heavily to metabolic and inflammatory processes. Meanwhile, the opposite presentation (AERS-) follows a different biological logic.

The failure of the unitary diagnosis

The central problem in modern psychiatry is heterogeneity. This refers to the fact that a single label like "MDD" masks a multitude of different physiological states. In the current Diagnostic and Statistical Manual (DSM), neurovegetative symptoms (physical functions such as sleep, appetite, and motor activity) are used as specifiers rather than primary dividers. This creates a research phenotype that aggregates individuals with opposing metabolic profiles.

As noted in the paper, this aggregation is problematic. It ignores long-standing clinical observations. "Atypical" depression (characterized by weight gain and hypersomnia) has historically responded differently to certain antidepressants than "melancholic" depression (characterized by weight loss and insomnia). By treating these as a single pool, researchers risk diluting specific genetic signals. If one subtype is driven by neuroinflammation and another by neurotransmitter depletion, a combined study will find nothing clearly belonging to either. This obscures the mechanisms required for targeted treatment.

Decoding the genetic architecture of energy symptoms

To resolve this, the authors moved away from the unitary model. They implemented a tripartite classification system based on the directionality of energy-related symptoms. They defined three distinct subgroups: 1. AERS+: Individuals experiencing hypersomnia (excessive sleep) paired with increased appetite or weight gain. 2. AERS-: Individuals experiencing insomnia or hyposomnia (reduced sleep) paired with appetite or weight loss. 3. Uncategorized: The largest group, representing individuals whose symptoms do not fit these specific directional patterns.

The researchers conducted three separate genome-wide association meta-analyses (GWAS)—large-scale scans of the genome to find associations between specific genetic variants and traits. They used data from eight different cohorts of European ancestry. This totaled hundreds of thousands of individuals. This allowed them to compare the "genetic load" of each group.

The methodology relied on several sophisticated statistical layers. They used LD score regression (LDSC) to estimate heritability (the proportion of variation in a trait attributable to genetic factors). They also measured pairwise genetic correlations to see how much the subtypes overlapped. Crucially, they employed a case-case GWAS to directly compare AERS+ individuals against AERS- individuals. This helped isolate the genetic differences existing specifically between the two extremes.

Divergent biology: Metabolism vs. Neurotransmission

The results reveal a stark divide in how these subtypes are built. The authors report that the AERS+ subtype exhibits a higher SNP-based heritability of 10.9% compared to 7.9% for AERS- .

This means a larger portion of the variance in this subtype is linked to common genetic variants.

The researchers also observed differences in polygenicity (the proportion of genetic variants that contribute to a trait). The AERS+ group showed a lower polygenicity of 1.7% compared to 2.9% for AERS- . In the AERS+ group, the authors also observed larger effect sizes at its associated loci [Figure 1H].

The most striking finding lies in the metabolic axis. The researchers found that AERS+ is strongly and positively correlated with Body Mass Index (BMI), metabolic syndrome, and various pro-inflammatory biomarkers .

Conversely, AERS- shows weak or even negative correlations with these same metabolic traits. To ensure this was not simply a matter of "depressed people being heavy," the authors used multi-trait-based conditional and joint analysis (mtCOJO) to adjust for BMI. Even after removing the direct genetic influence of BMI, the AERS+ subtype maintained distinct signals. This included associations with the NEGR1 gene, which involves both neuronal growth and feeding behavior .

Furthermore, Mendelian randomization—a method used to infer causality by using genetic variants as proxies for environmental exposures—suggested a bidirectional relationship. Increased genetic liability to metabolic syndrome and higher BMI appear to causally increase the risk of AERS+. The AERS- subtype shows the opposite trend .

Limits of the metabolic model

While the evidence for a metabolic-inflammatory axis in AERS+ is compelling, the study is not a panacea for psychiatric classification. First, the entire analysis was restricted to individuals of European ancestry. This is a significant caveat. The relationship between BMI and depression has been shown to differ in East Asian populations. This might render these subtype definitions less universal.

Second, the study relies on retrospective self-reporting of symptoms. This introduces the possibility of recall bias. A patient's current state might color their memory of their "worst" depressive episode. This is particularly sensitive for sleep symptoms. These are notoriously difficult to quantify accurately through questionnaires. Finally, the study does not account for medication history. Many antidepressants significantly alter weight and sleep patterns. It remains unclear how much of the AERS+ signature is inherent to biology versus a secondary effect of long-term pharmaceutical intervention.

A verdict for precision psychiatry

The evidence presented here is a powerful argument for moving beyond a "one-size-fits-all" approach to MDD. The researchers have demonstrated that the direction of neurovegetative symptoms is not mere clinical noise. It is a window into distinct biological realities. The AERS+ subtype emerges as a biologically coherent entity linked to metabolic dysfunction and systemic inflammation.

Whether these metabolic traits are a cause of depression or a consequence of it remains an open question. However, the implication for treatment is immediate. If AERS+ is indeed driven by an immunometabolic pathway, then therapies targeting metabolic health or inflammation might offer a specialized route for this subgroup. The data is available for replication at https://zenodo.org/records/20610434. The analytical code can be found at https://github.com/Ararder/AERS-GWAS.

Figures from the paper