For seven fixed mouse ratios, mean astrocyte tracer reach outside the detected source stayed lower with trimmed whiskers for any label overlap

We asked if unknown overlap between HA-positive source cells and streptavidin-positive labeled cells could explain lower astrocyte tracer reach after whisker trimming. We used seven mouse ratios from Cooper et al., Nature (2026). For each mouse, labeled cells outside the detected source per source cell equal the reported ratio minus the fraction of source cells carrying both labels. With the measured ratios fixed, sharp bounds kept the trimmed mean lower by 14.9% to 67.2%. If all source cells were labeled, the decrease was 54.3%. These are bounds for missing overlap, not confidence intervals. Exact enumeration reproduced the reported-ratio one-sided rank P of 1/35, in agreement with the revised source figure. The main text still gives 0.002. We checked all 47,293 weak rank orders, including ties; 1,185 were feasible. Across all overlaps, exact rank P for reach outside the source ranged from 1/35 to 4/35. Thus the sample mean direction was robust, but rank evidence could cross 0.05. If at least 22% of source cells were labeled in each trimmed mouse, exact rank P stayed at 1/35 for every remaining overlap. This is a sufficient assay check, not a measured overlap. The result concerns one small cohort and functional tracer transfer. It does not identify anatomical junction counts or neuronal performance. The package contains code, the proof, source provenance, all results, figures and repeat instructions. Original source files remain external under their original terms.

astrocytes · gap junctions · sensory deprivation · tracer reach · partial identification · exact permutation test · missing label overlap