Finding. Truncating KAT6A does not make one mitochondrial story. In patient iPSCs, cuts in exons 1–15 destroy the message through nonsense-mediated decay. Cuts in exons 16–17 escape decay and gain function. Those opposite genetics invert neuronal gene regulation, metabolism, and mitochondrial physiology. Faces and DNA-methylation episignatures split with them. A late-truncation cell might want a KAT6A inhibitor. An early-truncation cell wants the gene back. Gene-level diagnosis is the wrong unit.
Why this paper matters
Arboleda-Tham syndrome has been treated as a severity spectrum of one disease. This paper says the spectrum is two mechanisms. That is not a nuance. It is a contraindication: the drug that helps one class should worsen the other.
The mitochondrial clause is the reason it is here. Opposite directions of organelle physiology from the same gene, split only by variant position, is a warning against averaging rare-disease metabolomes. An epigenetic acetyltransferase is writing a mitochondrial setpoint, and the leftover protein (or the missing one) writes it in opposite ink.
What they actually measured
Patient iPSCs, multi-omics, variant-position classes, facial gestalt, episignatures, and neuronal programs that include metabolism and mitochondrial physiology. The abstract is mechanism-first and number-light. Believe the inversion claim; wait for the PDF for the Seahorse or imaging panel.
How to read the score
High seventies. Patient cells, opposite mechanisms, organelle readout, and a therapeutic fork. Confidence is medium because mitochondrial methods are unnamed. Score 78.
What to do with it
If you model KAT6A, chromatin-to-mitochondria coupling, or ARTHS, split your cohorts at exon 16. Pull the episignatures and the mitochondrial directionality plots. Do not design a single KAT6A drug for all truncations. The directional implication is that Mendelian mitochondrial physiology can invert inside one gene.