Inconsistency in Male CNS dataset: Soma Side vs Instance for ORN_VM4

[floesche]

None of the cells for ORN_VM4 currently have a soma side, although the majority of the instance names suggest they should.

Suggested solution 1: assign soma side according to instance name:

• for 120657549, 207890, 208854, 211355, 214955, 219326, 221662, 222819, 224031, 239695, 243289, 252813, 254983, 258507, 263646, 264011, 274149, 280533, 286398, 299072, 299288, 303869, 305168, 327161, 339206, 379010, 391675, 419497, 428811938, 436108, 452039, 455596, 477029, 519797468, 532302, 532303, 547138, 550697, 551819, 552143, 554095, 569391, 576872, 577129, 686346986, 712732813, 969501369 : somaSide='R'
• for 261548, 276056, 279337, 285620, 300065, 316132, 318688, 338564, 432746, 921680, 921975, 962000917: somaSide='L'
• for "_unknown": ignore 189202940, 191925150, 364416375, 402504772, 410132390, 422096435, 429689, 452797, 467724, 553982, 712920408, 780103867, 829367655, 902041659, 925509, 925649, 938371337, 942499, 943401 or

Suggested solution 2: change instance name according to reflect soma side:

• for 20657549, 207890, 208854, 211355, 214955, 219326, 221662, 222819, 224031, 239695, 243289, 252813, 254983, 258507, 263646, 264011, 274149, 280533, 286398, 299072, 299288, 303869, 305168, 327161, 339206, 379010, 391675, 419497, 428811938, 436108, 452039, 455596, 477029, 519797468, 532302, 532303, 547138, 550697, 551819, 552143, 554095, 569391, 576872, 577129, 686346986, 712732813, 969501369, 261548, 276056, 279337, 285620, 300065, 316132, 318688, 338564, 432746, 921680, 921975, 962000917: instance='ORN_VM4_unknown'

Broader observation

This could be a broader issue. The following query returns >13k results across 188 neuron types.

match (n:Neuron)
where n.somaSide is null
and n.instance=~ '.*[RL]$'
return n.type, n.instance, n.somaSide

This was observed in male-cns:v0.9. Initially reported for cns on August 16 in #16

[stuarteberg]

There are a few different places where the "side" might be recorded:

• instance
• somaSide
• rootSide -- generally used for cells in which the soma does not reside in the volume (and not just because we didn't manage to find it)

In the case of ORN_VM4 cells, we record the rootSide, though I note that for a subset of them, the rootSide remains unknown due to incomplete reconstruction (not for lack of trying, though).

In all cases, the ORN_VM4 rootSide is consistent with the instance, including the unknown cases.

https://neuroglancer-demo.appspot.com/#!gs://flyem-user-links/short/2025-10-02.143148.294118.json

[floesche]

Thanks, I didn't know about rootSide. That will most likely answer #99 as well.