Glutamate Receptor, Ionotropic, Gene-Related Neurodevelopmental Disorders Curation Project

Overview

This project organizes curation of Mendelian neurodevelopmental disorders caused by ionotropic glutamate receptor subunit genes, starting with:

• GRIN1-related neurodevelopmental disorder
• GRIN2A-related neurodevelopmental disorder
• GRIN2B-related neurodevelopmental disorder
• GRIN2D-related neurodevelopmental disorder
• GRIA2-related neurodevelopmental disorder

These disorders sit at the overlap of neurodevelopmental disorder, epilepsy / developmental and epileptic encephalopathy, language disorder, and movement disorder curation. The shared biological theme is impaired ionotropic glutamate receptor function at excitatory synapses, but the clinically useful pathograph is not identical across genes. The goal in dismech is therefore not to create a single generic "glutamate receptor disorder" entry, but to curate gene-centered entries that preserve receptor-subunit-specific mechanisms and treatment questions.

This document is a project-level organizing brief, not an assertion that all of these entities should collapse into one umbrella disease file.

Practical Modeling Decision For dismech

Recommendation: gene-level files, project-level umbrella

Why not one umbrella disease file by default?

• The affected receptor subunit changes the proximal biology, not just the severity label.
• NMDAR obligatory-subunit disease (GRIN1) is not equivalent to GluN2A, GluN2B, or GluN2D subunit disease.
• GRIA2 belongs in the same project because it is an ionotropic glutamate receptor disorder, but its AMPA-receptor biology is distinct enough that it should not be forced under a GRIN/NMDA disease umbrella.
• The electroclinical pattern differs materially across genes: some are more epilepsy-aphasia weighted, some are broader developmental disorders with variable seizures, and some have a stronger hyperkinetic / infantile DEE profile.
• Variant-function-informed treatment logic may differ by gene and by gain-of-function vs loss-of-function direction, so over-lumping would hide clinically important distinctions.

Why not split every syndrome label into its own file?

• Many named electroclinical labels in this space are downstream expressions of a shared gene/subunit mechanism rather than entirely different proximal pathobiology.
• Within a given gene, syndrome labels often differ more in age, seizure burden, language regression, or severity than in the earliest mechanistic trunk.
• Defaulting to gene-level entries keeps the pathograph coherent while still allowing has_subtypes, phenotype notes, or explicit variant-function notes where warranted.

Working Granularity Rules

Split into separate top-level entries when:

• The causal gene or receptor family changes.
• The proximal receptor biology changes enough that the early pathograph would no longer be shared.
• Treatment logic is directionally different in a way that a shared file would obscure.
• The literature frames the entity as a distinct disease family with its own recurring clinical and mechanistic signal.

Keep within one gene-level entry when:

• The shared proximal mechanism is still altered function of the same receptor subunit.
• Reported phenotypic differences are mostly severity-weighting or downstream branch differences.
• The main curation problem is preserving functional direction (loss-of-function, gain-of-function, mixed effects), not inventing new root disease files.

Use has_subtypes or explicit within-file structure when:

• A gene has recurring functional subclasses with clear clinical value.
• A seizure-dominant branch and a developmental-dominant branch are both common enough to deserve internal organization.
• A gene has a well-recognized epilepsy-aphasia or movement-disorder-heavy branch that still shares the same proximal receptor mechanism.

Existing KB Anchors

• kb/disorders/CTCF-related_Neurodevelopmental_Disorder.yaml
• Useful precedent for a gene-first neurodevelopmental disorder entry rather than a broad syndrome bucket.
• kb/disorders/Dravet_syndrome.yaml
• Useful precedent for a gene-driven developmental and epileptic encephalopathy entry with a clear circuit-level mechanistic chain.
• kb/disorders/Epilepsy.yaml
• Useful umbrella precedent showing that broad epilepsy should not absorb all mechanism-specific subtype content.
• projects/EPILEPSY.md
• Useful precedent for writing down split/lump rules before subtype curation.
• projects/RETINOPATHIES.md
• Useful precedent for using a project document to distinguish project labels, family umbrellas, and true standalone disease entries.

Current state:

• There are no existing GRIN1 / GRIN2A / GRIN2B / GRIN2D / GRIA2 disorder files in kb/disorders/.
• A local MONDO search already resolves GRIN1-related complex neurodevelopmental disorder; exact MONDO labels and IDs for the other working targets should be confirmed during YAML creation rather than guessed here.

Initial Target Set

Shared Mechanistic Backbone To Capture

1. Pathogenic variants alter receptor subunit abundance, assembly, trafficking, agonist response, or channel gating.
2. Ionotropic glutamate receptor signaling becomes quantitatively or qualitatively abnormal at excitatory synapses.
3. Synaptic plasticity, dendritic maturation, and activity-dependent circuit refinement are disrupted during development.
4. Cortical, hippocampal, and thalamocortical network balance is altered, increasing risk of epileptiform activity and seizures.
5. Persistent circuit dysfunction contributes to developmental delay, intellectual disability, language impairment, autism-related features, hypotonia, and/or movement abnormalities.

This shared trunk should guide project-level consistency, but it should not erase subunit-specific differences in receptor physiology.

Gene-Specific Modeling Emphasis

GRIN1

• Treat as the most general NMDAR-core entry in the set because GluN1 is an obligatory receptor subunit.
• Expect broad neurologic involvement: developmental delay / intellectual disability, epilepsy, hypotonia, and movement-disorder signal.
• Do not assume a single therapeutic direction across all variants; functional effect matters.

GRIN2A

• Expect stronger language and epilepsy weighting than in several other GRIN disorders.
• Keep epilepsy-aphasia-style labels inside the gene-centered file unless a truly separate proximal mechanism emerges.
• Capture speech and language regression carefully as a major downstream branch, not as a substitute for pathophysiology.

GRIN2B

• Expect a broader developmental-neurobiology framing, often with severe global delay / intellectual disability and variable epilepsy burden.
• Emphasize synaptic development, cortical circuit maturation, and plasticity disruption.
• Avoid over-centering epilepsy if the literature for a given branch is more developmental than electroclinical.

GRIN2D

• Expect a relatively sharp infantile DEE / movement-disorder profile.
• Capture hyperkinetic or abnormal-movement features as a major downstream consequence of receptor dysfunction when supported.
• This may become the cleanest use case for an internal subtype split if recurrent functional subclasses prove treatment-relevant.

GRIA2

• Keep in the same project because it belongs to the same ionotropic glutamate receptor mechanistic neighborhood.
• Model separately from the GRIN/NMDAR branch because AMPA receptor composition, trafficking, and excitatory-synaptic-strength effects are distinct.
• Treat it as the main adjacent-family test case for whether project-level curation rules generalize beyond NMDA receptor disorders.

Family-Specific Curation Rules

• Do not create one disease file per named seizure syndrome by default.
• Do not generalize treatment response across genes or across functional directions within a gene.
• Where functional studies are central to interpretation, make the evidence type explicit (HUMAN_CLINICAL vs IN_VITRO vs MODEL_ORGANISM).
• Prefer exact receptor-function language in pathophysiology nodes ("altered NMDA receptor gating", "reduced surface expression", "impaired excitatory synaptic transmission") rather than vague "synaptic dysfunction" wording.
• Keep phenotypes downstream from mechanism nodes. Do not collapse "seizures", "language disorder", or "autism" into proximal mechanism descriptions.
• Revisit within-gene subtype structure only when there is enough repeated evidence that it improves curation clarity rather than creating ontology noise.

Initial Execution Plan

1. Confirm exact MONDO label and ID for each target disorder before creating YAML files.
2. Start with GRIN1-related_Neurodevelopmental_Disorder.yaml as the anchor NMDAR-core entry.
3. Follow with GRIN2A-related_Neurodevelopmental_Disorder.yaml to establish the epilepsy-language branch pattern.
4. Curate GRIN2B-related_Neurodevelopmental_Disorder.yaml as the development-heavy counterpart.
5. Add GRIN2D-related_Neurodevelopmental_Disorder.yaml once the GRIN family template is stable.
6. Add GRIA2-related_Neurodevelopmental_Disorder.yaml as the adjacent AMPA receptor comparison case.
7. After the first 2-3 entries, decide whether a reusable project module or checklist is warranted for receptor-function direction, electrophysiology evidence, and precision-treatment notes.

Open Curation Questions

• Should recurring gain-of-function vs loss-of-function branches be represented as has_subtypes, or is that too variant-specific for current dismech disease-level files?
• When a gene-specific disorder is strongly framed by an epilepsy-aphasia syndrome in the literature, should that syndrome live only in phenotypes / notes, or as an explicit subtype?
• Is there enough shared proximal biology across GRIN1 / GRIN2A / GRIN2B / GRIN2D to justify a future reusable pathophysiology module, even if no umbrella disease file is created?
• Should precision pharmacology observations be modeled only in treatments, or also summarized in pathophysiology notes when variant function directly determines response?

STATUS

Scope and Modeling

• [x] Defined the project as a gene-family curation program rather than a single umbrella disease entry
• [x] Identified five initial target disorders spanning NMDA and AMPA receptor biology
• [x] Recorded default modeling rule: gene-level files first, subtype structure only when mechanistically justified
• [x] Recorded the main family-specific curation risks: over-lumping, syndrome proliferation, and treatment overgeneralization

Next Curation Steps

• [ ] Confirm exact MONDO mappings for GRIN2A / GRIN2B / GRIN2D / GRIA2 working targets
• [ ] Create GRIN1-related_Neurodevelopmental_Disorder.yaml
• [ ] Create GRIN2A-related_Neurodevelopmental_Disorder.yaml
• [ ] Create GRIN2B-related_Neurodevelopmental_Disorder.yaml
• [ ] Create GRIN2D-related_Neurodevelopmental_Disorder.yaml
• [ ] Create GRIA2-related_Neurodevelopmental_Disorder.yaml
• [ ] Reassess whether reusable within-family subtype conventions are needed

NOTES

2026-04-09

• Added an ionotropic glutamate receptor neurodevelopmental disorders project brief to guide future dismech curation.
• Chose a project-level umbrella with gene-level disease files as the default pattern.
• Kept GRIA2 in scope because it is mechanistically adjacent and useful for comparison, while explicitly not forcing it into a GRIN/NMDA disease umbrella.
• Recorded the main curation stance that named electroclinical syndromes should not automatically become separate root entries when the proximal receptor mechanism is still gene-centered.