support for neuron class annotations

[jefferis]

Introduces a new function class2ids() that allows one to search for broad neuronal classes. It would be great to add more including

• KCs
• LHNs/TOONs etc
• CNs (including of different levels)

We also need to figure out a way to do the inverse (i.e. ids2class and possibly ids2type etc).

[jefferis]

@schlegelp could you suggest how we should identify the list of LHONs/CNs at the moment? Thanks!

[schlegelp]

"Strong" CNs are neurons that meet both of these criteria:

1. 5% of inputs or 50 synapses in total from PNs or is an MBON itself
2. 5% of inputs or 50 synapses in total from MBONs or is a PN itself

Ideally the 5% of inputs would be counted per compartment (axon/dendrite).

For LHNs:
Either 5% of their pre- or post synapses in the LH. @alexanderbates used a connectivity approach to define LHNs based on inputs from uPNs.

In general: discard anything that's statusLabel Orphan, Assign or Unimportant

[jefferis]

Thanks @schlegelp! It would be great if you could provide exact code or queries and probably preferably a dump to a google sheet of the neurons thus identified (broken down by CN0/1 etc). The goal is to put you in charge of the definition and make it v quick and easy for others to retrieve the list of ids. Thanks!

…

Sent from my iPhone

On 17 Feb 2020, at 00:01, Philipp Schlegel ***@***.***> wrote: "Strong" CNs are neurons that meet both of these criteria: 5% of inputs or 50 synapses in total from PNs or is an MBON itself 5% of inputs or 50 synapses in total from MBONs or is a PN itself Ideally the 5% of inputs would be counted per compartment (axon/dendrite). For LHNs: Either 5% of their pre- or post synapses in the LH. @alexanderbates used a connectivity approach to define LHNs based on inputs from uPNs. In general: discard anything that's statusLabel Orphan, Assign or Unimportant — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

[mmc46]

I've tested for PNs and RNs, and it seems to be working fine.