Homologous synapse counts vary less than Poisson sampling in 8 C. elegans and more than Poisson in larval and 3 adult Drosophila connectomes; neuron gains remove half of the fly excess
We tested how different the synapse counts of the same connection are in two copies: left and right in one animal, or two animals. The reference is independent Poisson sampling around one common weight. The test conditions on the total count n of each connection pair, so the unknown true weight cancels. The dispersion index phi is the observed variance divided by the Poisson variance. We used 20 comparisons: 8 C. elegans nerve rings (Witvliet et al. 2021), one first-instar larval Drosophila brain, and three adult Drosophila connectomes (FlyWire FAFB v783, male CNS v1.0, BANC v888). In C. elegans, left and right copies of connections with n of 16 or more had pooled phi = 0.50 (95% bootstrap interval 0.41 to 0.60) in 8 animals; the Poisson null was 1.00 (p = 0.0005). Two animal pairs gave phi = 0.73. Neuron-level gains did not improve held-out predictions. In the larva, left-right phi was about 2 for n from 4 to 63 and did not increase with n. In adult flies, phi increased with n: within one animal, phi for n of 16 or more was 4.1 to 4.6 in all three animals; between animals it was 8.3 to 9.9, and 14 to 17 for n of 64 or more. Input and output gains for each neuron removed 50 to 56% of the excess between animals and 27 to 41% within animals. Input gains removed more excess than output gains in all 9 fly comparisons and gave larger held-out gains. Main limitations: each species uses a different synapse definition and reconstruction method, so absolute phi values are not a pure biological comparison between species. Fly analyses use only cell types with one neuron per side. Worm data cover only the nerve ring and use human annotation.
connectome · wiring variability · synapse count · Poisson · overdispersion · bilateral symmetry · C. elegans · Drosophila · FlyWire · BANC · male CNS · larval connectome · reproducibility · neuron gain · input fraction
References
11- usesdoi:10.1038/s41586-021-03778-8
Eight C. elegans nerve-ring connectomes (Witvliet et al. 2021).
- usesdoi:10.1126/science.add9330
Larval Drosophila brain connectome (Winding et al. 2023).
- usesdoi:10.7554/eLife.83739
Left-right neuron pairs and edge list of the larval connectome; prior bilateral symmetry tests.
- usesdoi:10.1038/s41586-024-07558-y
FlyWire FAFB v783 connectome.
- usesdoi:10.1038/s41586-024-07686-5
FlyWire cell types; prior cross-connectome consistency analysis.
- usesdoi:10.1101/2025.07.11.664377
Princeton synapse detection used for the FAFB table without a pair threshold.
- usesdoi:10.1101/2025.10.09.680999
Male CNS v1.0 connectome.
- usesdoi:10.1038/s41586-026-10735-w
BANC v888 connectome.
- usesdoi:10.7910/DVN/7WTH1N
BANC v888 data release on Harvard Dataverse.
- useshttps://github.com/dwitvliet/nature2021/tree/0646af9d25896ae660f97d462eab2d67282f5625
Source of the Witvliet nemanode JSON files.
- useshttps://github.com/neurodata/bilateral-connectome/tree/0ff9bbe0515504b88ac827c6a31481db47a92ad2
Source of the larval edge list and pair metadata.
Cited by
9 publicationsPublic catalog entries on Sciscape.
- reviewsTwo adult worm maps showed no clear gain in fixed-sample input-source diversity for 37 selected brain cells
Prior homologous synapse-count variability study; provides context, with no code or numerical data reused.
- reviewsMean short-route backup of bilateral neuron pairs remained after weak-edge removal in two adult worm connectomes
Campaign context on left-right count precision. No code or tables reused.
- reviewsMirror-edge presence improved prediction of edge presence between two adult worm maps after count and cell controls
Prior campaign test of homologous synapse-count dispersion. This study instead tests added cross-animal edge-presence prediction.
- Membrane contact explains the higher synapse counts of two-way connections in 7 C. elegans nerve rings; in 2 fly connectomes a 1.2 to 1.3 fold excess stays after coarse controls
Related campaign work on synapse-count variability in worm and fly connectomes.
- reviewsMirror edge presence improved edge prediction beyond synapse count in two adult worm maps
Prior count-dispersion result motivates the separate edge-presence prediction test. No code or derived tables reused.
- In the FlyWire fly brain, the transmitter share of neuron input varied less among neurons of one cell type than matched input mixes; eight worm brains gave no clear result
Related campaign result on stereotypy of homologous synapse counts in worm and fly.
- Variable chemical wiring across 8 C. elegans brains did not predict more variable co-activity across animals; gap junctions in more brains gave stronger, and in moving worms more variable, co-activity
Related analysis of synapse count variability across the same eight worm connectomes.
- Hemibrain and FlyWire head-direction wiring share a ring form: Delta7 inhibition peaks opposite the source, PEN outputs shift 48-58 degrees by side, and a cosine mode pair leads
Earlier campaign result that compares synapse counts between fly connectomes.
- reviewsA broader count distribution improved held-out synaptic contact predictions in one selected MICrONS mouse
Prior dispersion test used homologous pairs in other species and multiple animals.