Experimental research roadmap
Five stages, each designed to decide whether the next one is worth doing.
The logic of the order
Each stage exists to decide whether the next one is worth doing. That is the only honest way to plan research you cannot yet afford.
The stages are not a schedule and they are not funded. Stage 1 is running. Stages 2 to 5 describe what would follow, and are written so that a laboratory considering a collaboration can see immediately where the real decision points are and where the money would go.
Most available cell models express the paralogue eEF1A1, either exclusively or alongside eEF1A2 — transformed lines, immortalised lines, iPSC-derived cells and primary cultures alike — which confounds functional work on this gene. Any plan that does not solve the isoform-background problem first will produce results that cannot be interpreted. This is why Stage 3 is not simply “make iPSC neurons”.
Computational investigation running
Molecular dynamics comparing wild-type and D91N eEF1A2 across nucleotide states, with the analysis plan fixed before the data were examined.
Candidate measurements: RMSD, RMSF, hydrogen-bond occupancy, solvent-accessible surface area and water penetration into buried regions, structural fluctuation of the switch regions, and the coupling between nucleotide state and conformation.
What this stage can decide: which mechanism class remains open, and therefore which experiment in Stage 2 is worth someone's money. What it cannot decide: anything about disease mechanism on its own. Full method and limitations.
Gate to Stage 2: the positive-control test must pass — the wild-type model has to distinguish its own nucleotide states. If it does not, the simulation is uninformative regardless of what the mutant appears to do, and that is reported as such.
Protein biochemistry
The stage this project most needs, and the one it cannot do itself. Purified wild-type and D91N protein, measured directly.
Candidate experiments: thermal and conformational stability; degradation and half-life; steady-state abundance; GTP binding affinity; GTPase activity; aminoacyl-tRNA binding and delivery.
Why it comes first among the wet experiments: it is the cheapest way to answer the question that determines everything downstream — whether D91N is a protein that cannot work, or a protein that works wrongly. Those imply opposite interventions, and no amount of cellular phenotyping resolves it as directly.
Cellular models
Feasibility to establish, not a commitment: patient-derived iPSC lines; iPSC-derived neurons differentiated far enough that the isoform switch has actually occurred; isogenic CRISPR-corrected controls from the same line; and an appropriate muscle-cell model.
Why cell-type-specific expression matters so much here. eEF1A2 expression is concentrated in brain, heart and skeletal muscle. In neurons it replaces eEF1A1 during development. A model that still runs on eEF1A1 is a model in which the variant barely matters — which makes a negative result meaningless.
Two published findings make the muscle arm non-optional rather than a nicety: the E122K protein was stable in neurons and unstable in myocytes in the same animal, and heterozygous D252H mice showed neuromuscular phenotypes. Brain-only phenotyping could miss where the variant does its work.
Translation and cellular stress
Candidate measurements: translation efficiency; ribosome occupancy at codon resolution; evidence of ribosome collisions; integrated stress response activation; ribosome-associated quality control markers; proteostasis abnormalities.
Design constraint that follows from our own hypothesis review: measure under load, not at rest. If the phenotype is a loss of reserve rather than a loss of baseline function, resting-state measurements will look normal. Challenge conditions — metabolic stress, increased synthetic demand — are where a reserve deficit becomes visible.
Second constraint: if the disturbance is transcript-selective rather than global, global measures of protein synthesis are the wrong instrument. Length-matched reporters and transcript-resolved readouts should be included from the start rather than added after a negative global result.
Therapeutic target identification
Nothing in this stage begins until Stages 2 to 4 have produced an experimentally supported abnormality with a known direction. The reason is specific to this disorder: the field has not settled whether pathogenic variants act by loss of function or by toxic gain of function. Those imply opposite interventions. An approach built on the wrong one does not fail neutrally.
If and when that point is reached, candidate approaches would be sorted strictly into four categories, and the labels would be used honestly:
| Category | What it means |
|---|---|
| Established treatment | Approved, with evidence in this or a comparable indication. For this disorder, symptomatic management decided by treating clinicians — nothing disease-modifying is established. |
| Experimental compound | In clinical investigation. Not available outside a trial, and not a basis for any individual decision. |
| Preclinical strategy | Supported in cells or animals only. Most precision approaches for ultra-rare variants sit here, including allele-specific approaches. |
| Unsupported speculation | Mechanistically arguable, experimentally untested. Much of what circulates in rare-disease communities, including some of our own earlier reasoning. |
What this site will never do: give personalised treatment recommendations, paediatric dosing, or instructions for administering experimental compounds. That holds regardless of what the research shows, and it holds for substances available without prescription just as much as for drugs.
What we are asking a laboratory for
Not a commitment to the whole roadmap. The useful first conversation is much smaller: whether Stage 2 is tractable with existing reagents, roughly what it would cost, and whether the question is interesting enough to be worth a graduate student's side project. A reply explaining that the premise is wrong would also be valuable, and would save us from the alternative.
Start that conversation The hypotheses behind the stages
References
Rendered from the site bibliography. Each entry states how far the source was actually read — a database record, an abstract, or the full text — because that distinction changes how much weight a citation deserves.