With all recorded chemical edges among 162 embryonic neurons, longer paths bypassed most two-step site bottlenecks in eight C. elegans brain maps
We tested whether a shared synapse site can block all paths between neurons that have several two-step routes. We used eight C. elegans brain maps and 162 neurons born before hatching. A site loss removes all of that site's target contacts together. We computed exact minimum site cuts for all directed neuron pairs. Among pairs with no direct edge and at least two two-step relays, 280 had one site shared by every first leg. A longer chemical path remained after that site was removed in 263 cases. The map fractions were 79.6 to 100 percent. Across the two adult maps, all 84 local cases had bypasses. Requiring at least two sites on every remaining connection left a bypass for 167 of 280 local cases, including 79 of 84 adult cases. Three-site and five-site requirements retained 91 and 28. Site cuts were smaller than separate-contact cuts for 38.3 to 50.3 percent of reachable pairs. Most of these reduced cuts had an optimal cut restricted to a source or target boundary. Cell-panel and source-node checks used the birth map and both adults. All adult local cases still had bypasses in both checks. The package has all pair results, cut certificates, path witnesses, source hashes, code, figures and repeat steps. The full-path audit was added after the first cut results. This study is exploratory. Paths have no sign, timing or cell-compartment limits. The results do not show functional rescue, biological loss rates or a population age trend.
connectome · Caenorhabditis elegans · polyadic synapse · directed minimum cut · shared failure · path length · network robustness · reproducible benchmark
References
4- usesdoi:10.1038/s41586-021-03778-8
Original chemical synapse site lists, neuron metadata and age metadata at the pinned author repository commit.
- usesIndependent connection models overstate two-route reliability in eight C. elegans brains under independent loss of synapse sites sci:sha256:e4e968f83ac4898fed7ad12dae72e6e8007240a0bd95c1d5ea014ad90cd9fd48
We extend its exactly-two-step question to paths of any length and revisit its selected adult example. We did not reuse its code or derived tables. Its short-route result remains within its stated scope.
- doi:10.48550/arXiv.2001.02817
Prior framework for auxiliary graph representations of hypergraph cut costs.
- doi:10.1137/1.9781611974782.71
Prior connectivity analysis for edges with shared removal groups. General grouped-edge models differ from the one-source groups used here.
Cited by
5 publicationsPublic catalog entries on Sciscape.
- usesWith at least three locations per connection, 1 micrometre depth omissions gave excess loss over source position controls in eight C. elegans maps
We use its CC BY 4.0 list of 162 embryonic brain neuron names. We extend the shared-site loss question to a defined spatial omission model. We do not reuse its cut algorithm.
- challengesA neuron removal bound exceeded exact target cuts for 8,336 pairs in eight C. elegans maps under whole site loss
We reuse its exact global cuts and correct only its claim that the target-only endpoint count is exact. Its global cut values and certificates remain valid.
- Counts of paths with no shared relay gave no clear prediction gain in fixed tests of 13,135 C. elegans calcium responses
This prior study tested synapse-site cuts and longer bypass paths. We test internal neuron cuts and measured response prediction. We do not reuse its data or code.
- In one Platynereis graph with unsigned sensory routes, a five-synapse cutoff raised the count of effectors with one-neuron cuts from 10 to 86 of 138
This prior study tested full-path bypasses and synapse count cutoffs for site cuts in worm brain graphs. Our new analysis tests neuron cuts from a sensory source set to annelid effectors.
- usesAt 50% independent site loss, independent connections overestimated full path reliability by 0.28–7.13 points among 162 embryonic neurons in eight C. elegans maps
We use its verified minimum-cut tables to define fixed pair groups. We test random simultaneous failures instead of minimum deletion counts.