Arvid Harder, PhD candidate of Karolinska Institutet, Stockholm, Sweden.

Arvid Harder

A genome-wide association meta-analysis with over 460,000 participants has identified 2 genetically distinct subtypes of 

: atypical energy-related symptoms positive (AERS+; hypersomnia with increased appetite and weight gain) and AERS negative (AERS−; insomnia with appetite and weight loss). These subtypes diverge in heritability, polygenicity, and metabolic genetic correlations.¹

Current MDD diagnostic criteria treat opposite-direction neurovegetative symptoms, weight gain vs weight loss, and hypersomnia vs insomnia as clinically interchangeable.¹ The investigators, working within the Psychiatric Genomics Consortium's MDD working group, stratified individuals meeting MDD criteria into AERS+, AERS−, and an uncategorized group before running subtype-specific genome-wide association analyses.¹ The approach builds on the research team's earlier coverage of the AERS framework's genetic architecture.²

"We know if you're at a higher BMI [body mass index], your subsequent risk of depression increases, and vice versa," said Arvid Harder, lead study author. "We are unable to make any claims about the [causal architecture]. We just show they co-occur together, but we don't know exactly [how]."

Metabolic Traits Differentiate AERS+, AERS− MDD Subtypes, With Arvid Harder

AERS+ and AERS− Subtypes Show Distinct GWAS Architecture

The meta-analysis pooled data across cohorts to reach effective sample sizes of 47,858 for AERS+, 156,624 for AERS−, and 215,828 for the uncategorized group. Across the 3 groups, the investigators identified 27 genome-wide significant loci, including 4 for AERS+, 10 for AERS−, and 13 for the uncategorized group.¹

AERS+ showed a greater SNP-based heritability (10.9%) than AERS− (7.9%) or the uncategorized group (8.6%), despite its smaller sample size. AERS+ also showed lower polygenicity, with associated variants covering an estimated 1.7% of SNPs compared with 2.9% for AERS− and 5.3% for the uncategorized group, and larger effect sizes at its associated loci. AERS+ and AERS− were moderately genetically correlated (rg = 0.64; SE = 0.04), indicating overlapping but distinguishable genetic architecture between the 2 symptom-direction subtypes.¹

Metabolic Genetic Correlations Distinguish MDD Subtypes

Metabolic traits emerged as the primary axis separating the subtypes. AERS+ correlated positively with BMI, metabolic syndrome, and related metabolic traits, while AERS− correlated weakly in the opposite direction.¹

What is the difference between AERS+ and AERS− depression subtypes?


AERS+ describes MDD with hypersomnia and weight gain, while AERS− describes MDD with insomnia and weight loss; the 2 subtypes differ in heritability, polygenicity, and metabolic genetic correlations.

How many genetic loci were identified across the MDD subtypes?


The meta-analysis identified 27 genome-wide significant loci in total, 4 for AERS+, 10 for AERS−, and 13 for the uncategorized group.

What clinical implication does the research suggest for depression assessment?


The investigators suggest retaining the direction of neurovegetative symptoms, such as weight gain versus weight loss, in diagnostic instruments rather than treating opposite-direction symptoms as equivalent.

Harder said the pattern may reflect how genetic studies of clinical traits also capture biomarkers shaped by a person's current body weight rather than depression biology alone. AERS−, marked by weight loss during a depressive episode, showed a genetic link to anorexia nervosa alongside more favorable metabolic markers, a pattern Harder attributed partly to weight-related masking of the underlying depression signal.

The findings identify metabolic biology as a central axis differentiating MDD subtypes beyond mood symptoms alone. Harder said continuous assessment of symptom directionality, rather than a single unified depression score, could eventually guide treatment selection if subgroup-specific treatment response is confirmed in future trials.

“It's very nice to have this dialogue between clinicians and researchers on how we should continuously develop these instruments and which questions we ask,” Harder said.

Harder A, Wang R, Bergstedt J, et al. Atypical energy-related symptoms define biologically distinct subtypes of major depressive disorder. medRxiv. Preprint posted online July 2, 2026. 

Harder A. AERS subtypes show distinct genetic architecture in MDD. 

. August 12, 2026. Accessed August 12, 2026. 

https://www.hcplive.com/view/aers-subtypes-major-depression-genetics

Articles

A large genome-wide meta-analysis (>460,000 participants) supports 2 genetically distinguishable MDD subtypes defined by atypical energy-related symptom directionality: AERS+ (hypersomnia, hyperphagia, weight gain) and AERS− (insomnia, reduced appetite, weight loss). Despite smaller effective sample size, AERS+ showed higher SNP-heritability, lower polygenicity, and larger locus effects. The subtypes were moderately genetically correlated, indicating partial overlap with separable architecture. Metabolic genetics most clearly differentiated them, with AERS+ aligning with adverse metabolic traits and AERS− showing opposing correlations.

New genome-wide data show AERS+ and AERS− subtypes of major depressive disorder diverge in heritability and metabolic genetic correlations.

A genome-wide association study of > 460,000 individuals of European ancestry has identified 27 genetic loci separating 2 biologically distinct subtypes of 

 defined by opposing atypical energy-related symptoms (AERS).¹

Current diagnostic criteria count hypersomnia and insomnia, or weight gain and weight loss, as equally valid paths to an MDD diagnosis, despite opposite underlying biology. Investigators from the Major Depressive Disorder Working Group of the Psychiatric Genomics Consortium, based at Karolinska Institutet, built on earlier heritability work identifying this symptom grouping as a genetically distinct target.

 The new analysis expands the approach to 3 subtypes across nearly half a million individuals.

AERS Subtypes Show Distinct Genetic Architecture in MDD

"[Our] lab, led by Dr. Yi Lu, works on depression subtypes, trying to look specifically at how we can group patients with MDD into subgroupings,” said Arvid Harder, of Karolinska Institutet, Stockholm, Sweden. “[This] might give us more of an idea [on] whether they respond to certain medications or whether they're caused by different things.”

AERS Subtype Definitions and GWAS Architecture

The investigators defined 3 subgroups among individuals meeting criteria for MDD:

AERS+, characterized by hypersomnia with increased appetite or weight gain (n = 47,858);

AERS−, characterized by insomnia with decreased appetite or weight loss (n = 156,624); and

Uncategorized group with mixed presentations (n = 215,828).

Across the 3 subtypes, the analysis identified 27 genome-wide significant loci: 4 for AERS+, 10 for AERS−, and 13 for Uncategorized.

No prior MDD genetics study had reported several of these loci.

AERS+ showed greater SNP-based heritability, at 10.9%, compared with 7.9% for AERS− and 8.6% for Uncategorized. Polygenicity followed the opposite pattern: AERS+ variants accounted for 1.7% of SNPs studied, versus 2.9% for AERS− and 5.3% for Uncategorized. The study found larger effect sizes with AERS+-associated loci.

Genetic Correlations and Metabolic Risk in MDD Subtypes

What are the AERS+ and AERS− subtypes of major depressive disorder?


AERS+ is defined by hypersomnia with weight gain; AERS− is defined by insomnia with weight loss. The Uncategorized group has mixed neurovegetative symptoms.

How genetically distinct are the AERS subtypes?


AERS+ and AERS− share a moderate genetic correlation. with AERS+ showing higher heritability and lower polygenicity than AERS−, pointing to overlapping but partially distinct biology.

Could these findings change how depression is treated?


The investigators propose testing subtype-specific approaches, including GLP-1 receptor agonists in AERS+ patients, though no treatment-matching trials have been conducted.

AERS+ and AERS− shared a moderate genetic correlation (rg = 0.64; SE = 0.04), indicating overlapping but partially distinct genetic architecture rather than fully separate disorders.

Metabolic traits emerged as the primary axis of differentiation: AERS+ correlated positively with body mass index (BMI), waist circumference, metabolic syndrome, type 2 diabetes, and blood glucose, while AERS− correlated weakly in the opposite direction. Investigators note the immunometabolic profile in AERS+ resembles preclinical metabolic syndrome, while AERS− appears biologically more protective on standard metabolic markers.

AERS+ also showed a negative correlation with HDL cholesterol and a positive correlation with C-reactive protein, alongside links to ADHD, hypertension, and coronary heart disease.

 AERS− carried a genetic correlation with anorexia nervosa and had a possible shared mechanism with schizophrenia risk.

"Right now, it's hard for clinicians to make definitive changes, because the [goal] of subtyping is tailoring treatment to different subtypes," Harder said. “We're not quite there yet.”

Harder A, Wang R, Bergstedt J, et al. Atypical energy-related symptoms define biologically distinct subtypes of major depressive disorder. medRxiv. Preprint posted July 2, 2026. Accessed August 10, 2026. 

Mitchell BL, Monistrol-Mula A, Thomas JT, Byrne EM. The Role of Genetic Data in Dissecting Depression Heterogeneity. 

A genome-wide association study of >460,000 Europeans stratified major depressive disorder into atypical energy-related symptom subtypes defined by opposing sleep and appetite/weight patterns. Three groups were analyzed: AERS+ (hypersomnia with hyperphagia/weight gain), AERS− (insomnia with hypophagia/weight loss), and mixed presentations. Twenty-seven genome-wide loci differentiated these groups, with AERS+ showing higher SNP-based heritability, lower polygenicity, and larger locus effect sizes. Genetic correlations were moderate between AERS+ and AERS−, while metabolic and inflammatory trait associations most strongly separated the subtypes.

A genome-wide study of > 460,000 individuals identify 27 loci separating AERS+ and AERS− subtypes of major depressive disorder by metabolic profile.