SMA Research Platform

Evidence graph for Spinal Muscular Atrophy

Biology-first target discovery
Christian Fischer / Bryzant Labs
14,766Targets
453Trials
84Drugs
7Datasets
6,987Sources
64,683Claims
72,052Evidence
29,649Hypotheses

SMA Top-10 Intervention Board

v2 LOCKED · 2026-05-06
10 curated · ranked 1–10

This board replaces the “29,625 hypotheses” vanity metric per Phase 1 of the breakthrough plan. The platform’s job is killing or proving fast, not accumulating.

Every entry below carries an explicit kill criterion, a required experiment, and a current claim-state on the six-state ladder. None has yet been experimentally validated. The full 29K hypothesis archive remains available at /hypotheses/ (research-grade signal, not validated leads).

— Synthesised from the 3-LLM cross-review (ChatGPT + Gemini + Claude Opus) and the locked breakthrough plan v2.

01

DOK7-MuSK-Agrin-LRP4 NMJ ternary complex stabilization

2 · Computationalsmall moleculeaxis: NMJconf 55%

Stabilization of the agrin-LRP4-MuSK-DOK7 ternary signalling complex restores neuromuscular-junction integrity in SMA. The complex is dynamic, multi-protein, underserved by classical structural biology — exactly where a multi-model AI consensus stack (AF3 + Boltz-2 + Chai-1 + NeuralPLexer-3) gives an edge over Roche / Biogen / Novartis. Whitespace target: SMN-independent, NMJ-stability-axis.

Required experiment

iPSC patient-derived motor neuron + myotube co-culture; readout: AChR clustering area, synapse count, MN survival, calcium activity, neurite length. CRO partnership scheduled Phase 3.

Kill criterion (falsifies the hypothesis)

No measurable AChR cluster size increase AND no MN survival benefit on top of risdiplam standard-of-care after 14-day exposure.

02

AGRN-LRP4 binding axis stabilization (NMJ maturation)

1 · Hypothesisantibodyaxis: NMJconf 50%

AGRN-LRP4 binding triggers MuSK clustering and AChR aggregation at the synapse. Stabilizing this initial step rescues NMJ maturation in patient iPSC motor neurons. Antibody-class candidate (ARGX-119 / adimanebart) under industry development for neuromuscular disease.

Required experiment

iPSC patient-derived MN + myotube co-culture + LRP4-stabilizing antibody panel. Readouts: synapse maturation kinetics, AChR cluster size, miniature endplate potentials.

Kill criterion (falsifies the hypothesis)

No synapse-maturation acceleration vs vehicle in patient iPSC NMJ within 21-day differentiation window.

03

Apitegromab-class myostatin antagonism on top of SMN restoration

1 · Hypothesisantibodyaxis: muscleconf 65%

Latent myostatin antagonism rescues residual muscle weakness in SMA patients already receiving SMN restoration. SAPPHIRE Phase-3 hit endpoints (Lancet Neurology 2025), opening the lane for orally bioavailable small-molecule myostatin antagonists.

Required experiment

SMA Δ7 mouse on nusinersen + apitegromab analog vs nusinersen alone. Readouts: grip strength, motor-function score, muscle mass, lifespan.

Kill criterion (falsifies the hypothesis)

No incremental motor-function gain vs nusinersen-alone arm at 90-day endpoint (paired study, n ≥ 8 per arm).

04

STMN2 axonal-maintenance restoration in SMA motor neurons

1 · Hypothesisasoaxis: axonalconf 40%

STMN2 (Stathmin-2) regulates axonal microtubule dynamics. Restoring STMN2 splicing or expression rescues axonal length and branching in SMA motor neurons. Open whitespace — TDP-43 + STMN2 axis dominant in ALS literature, underexplored in SMA.

Required experiment

Patient iPSC motor neurons + STMN2 splice-rescue ASO panel. Readouts: full-length STMN2 transcript fraction, axon length, axon branching, retrograde transport.

Kill criterion (falsifies the hypothesis)

No axon-length or axon-branching rescue in patient MNs at 14-day exposure to STMN2 splice-rescue ASO.

05

SMN2 splice ternary-complex modulation (RNA-protein-ligand)

2 · Computationalsmall moleculeaxis: splicingconf 60%

RNA-protein-ligand ternary complex modulators of SMN2 splicing (risdiplam-class). Technical capability moat — running AF3 + Boltz-2 in proper RNA-protein-ligand ternary mode is what separates our platform from the field.

Required experiment

SMN2 minigene splice-rescue assay in patient fibroblasts + AF3-ternary-predicted ligand panel. Readouts: full-length SMN2 transcript fold-increase, SMN protein increase, MN survival.

Kill criterion (falsifies the hypothesis)

No SMN2 full-length transcript increase ≥ 1.5× over vehicle for any AF3-nominated ligand at 1 µM (panel of 8 nominations).

06

ρ-ROCK-LIMK-Cofilin actin-dynamics axis (under re-audit)

2 · Computationalsmall moleculeaxis: actinconf 45%

The ρ-ROCK-LIMK-Cofilin pathway controls actin dynamics; actin pathway dysregulation is triple-validated in SMA transcriptomic data (GSE287257, GSE290979, GSE255215). Specific-compound rescue claims (Fasudil, ripasudil) are PENDING re-verification per QMS audit. Axis-level hypothesis remains open; specific-paper evidence requires re-audit before re-citation.

Required experiment

iPSC patient-derived MN + ROCK inhibitor (Fasudil, ripasudil, Y-27632) panel. Readouts: actin polymerization, AChR clustering, MN survival.

Kill criterion (falsifies the hypothesis)

No AChR-cluster rescue with any ROCK inhibitor at 1-10 µM in patient iPSC MNs (panel of 3 compounds, n ≥ 4 per condition).

07

SMN backbone + NMJ-rescue combination (Phase 3 wet-lab target)

1 · Hypothesiscombinationaxis: combinationconf 40%

Risdiplam (or nusinersen, or onasemnogene) PLUS a hit from the NMJ-rescue lane (Top-10 #1 or #2). Hypothesis: additive synaptic-density and motor-function rescue beyond SMN restoration alone.

Required experiment

iPSC patient MN + myotube co-culture; arm A: risdiplam alone, arm B: NMJ-rescue compound alone, arm C: combo. Readouts: AChR cluster density, mEPP frequency, MN survival.

Kill criterion (falsifies the hypothesis)

No additive AChR cluster density (≥ 1.5× over either arm alone) at any tested combo concentration.

08

SMN backbone + myostatin antagonism combination

1 · Hypothesiscombinationaxis: combinationconf 55%

SMN therapy plus apitegromab-class myostatin antagonist. Direct extrapolation from SAPPHIRE Phase-3 design.

Required experiment

SMA Δ7 mouse, n=8 per arm: nusinersen alone, myostatin-antagonist alone, combo. Readouts: grip strength, treadmill, lifespan.

Kill criterion (falsifies the hypothesis)

No incremental grip-strength gain ≥ 15% over nusinersen-alone arm at 90-day endpoint.

09

SMN backbone + cytoskeletal-rescue combination

1 · Hypothesiscombinationaxis: combinationconf 35%

SMN therapy plus a hit from the actin-axis lane (Top-10 #6, #4 STMN2). Hypothesis: synergistic neurite + axonal preservation beyond SMN-directed therapy alone.

Required experiment

iPSC patient MN; arm A: risdiplam, arm B: STMN2-rescue ASO, arm C: combo. Readouts: axon length, branching density, MN survival.

Refuting / contradicting (0)

No refuting papers attached — actively searching.

Kill criterion (falsifies the hypothesis)

No additive axon-length or branching density vs either arm alone.

10

NRF2 / mitochondrial-modifier repurposing for SMA motor neurons

1 · Hypothesisrepurposingaxis: redoxconf 30%

NRF2 antioxidant response activators (e.g. dimethyl fumarate, omaveloxolone) rescue oxidative-stress phenotype in SMA motor neurons. Existing approved drugs in adjacent indications (MS, Friedreich's ataxia) provide a fast translation path.

Required experiment

iPSC patient MN under oxidative stress + dimethyl fumarate / omaveloxolone panel. Readouts: ROS levels, glutathione, MN survival.

Kill criterion (falsifies the hypothesis)

No oxidative-phenotype rescue (ROS ≥ 30% reduction OR survival ≥ 20% improvement) at any tested concentration.

Maintenance:Top-10 ranking is curated by Christian + CORTEX CEO from the locked breakthrough plan. Auto-generated hypothesis stubs do not appear here. Every quarter, one entry on this board becomes the OSF pre-registered prediction with the “published in 30 days regardless of outcome” commitment.
Login → Command Center