A stated vesicle mixture equation gave 83.6% inferred release at its 46% group weight in cultured mouse synapses

We tested a published calculation that used vesicle counts to infer release in cultured mouse synapses. It assumed that postfusion proximal counts equal ongoing-fusion counts, and that nonrelease counts equal control counts. The stated equation gives a nonrelease fraction of 35.76% within the mixed group. At the stated group weight of 46%, this gives 83.55% inferred release. The reported 17% fraction gives a different mixed mean. Rounding does not explain the error. We also tested a stronger model that applies one mixture fraction to twelve distance bins. It has no exact fit to the displayed means. This is a model check, not a statistical rejection. If postfusion mean vectors can be any nonnegative values, two specified control choices give release ranges of 83.70–100% and 82.88–100% in a selected 54-synapse panel. We construct mean vectors that give 92% release and match all displayed mixed means. The firm result is a correction to the stated calculation. The actual release probability remains unresolved. The package includes source data, provenance, code, proofs, tables, a figure, and repeat instructions.

synaptic release · vesicle counts · cryo-electron tomography · structure and function · mixture model · partial identification · reproducibility