Variant database
Published EEF1A2 variants, with what has been reported about each and — in every entry — what the evidence does not establish.
This is a reading aid compiled by a family, not a curated clinical resource. It is not ClinVar, it is not a diagnostic tool, and it has no authority. Three rules govern it: a field is filled only where a source we have opened says so; a field we could not confirm is left empty rather than guessed; and every entry carries its own statement of evidence limitations.
For clinical variant interpretation, consult ClinVar, the primary literature, and your clinical genetics service. Several entries here are deliberately incomplete and say so.
Variants along the protein
Each marker is one variant position on the 463-amino-acid protein. Highlighted markers are the variants this project follows most closely — either because they recur in the literature or because they bear on D91N. Domain boundaries are approximate.
Loading…
Reading the functional column honestly
Most of the functional evidence in this table comes from one experimental design: a yeast complementation assay. The endogenous yeast factor is removed and the human variant is asked to do its job instead. If the cell grows, the variant retains enough function; if it does not, the variant is damaging.
This is a good assay for the question "is this variant damaging?" and a poor one for the question "what does this variant do?" It gives a binary readout in a single-celled organism that has no neurons, no developmental isoform switch, and no eEF1A2 at all. It cannot distinguish a protein that folds badly from one that folds well but cannot hydrolyse GTP, and it cannot detect harm that only matters in a cell type yeast does not have.
The same paper's own internal inconsistency is worth knowing: the cohort table records complementation as "not reported" for several patients while the text states that all variants except P333L impaired function. We have not resolved that, which is why some entries carry no functional claim.
Two entries to look at if you want to calibrate your scepticism
- p.M102V — impaired growth in the assay, inherited from an unaffected father. A damaging assay result in a healthy person.
- p.R423C — appears in the structural analysis of a paper but not in that paper's own cohort table. We list it so the discrepancy is visible rather than quietly dropped.
Help us correct it
If you are an author of any of these papers, a clinician with a variant we have mischaracterised, or a reader who spots an error, we would rather be corrected than consistent. Errors in this table are our errors, not the sources'. Please get in touch.