search

flyconnectome / tracerutils

R tools for EM tracers
https://flyconnectome.github.io/tracerutils
GNU General Public License v3.0
0 stars 1 forks source link

fix PN glomerulus identification #47

Closed jefferis closed 5 years ago

jefferis commented 5 years ago

These now seem to be identified by name rather than annotation.

Something like this seems to work:

pns.fafb[,'glom']=stringr::str_match(catmaid_get_neuronnames(pns.fafb[,'skid']),'glomerulus (\\S+)')[,2]
jefferis commented 5 years ago

Should add an option to do this to find_glomeruli()

fmlove commented 5 years ago

Are the annotations no longer being updated? I would have thought they'd be more useful, particularly with mPNs.

schlegelp commented 5 years ago

I use a combination of annotations. E.g. for right canonical DA1 uPNs I would intersect annotations glomerulus DA1 and WTPN2017_excitatory_uPN_right.

fmlove commented 5 years ago

That's what I do for uPNs anyway. What I don't understand is why this broke - shouldn't the annotations still be more robust than names? I'm not averse to adding a name-based option to find_glomeruli(), but I'm not sure on the reasoning.

jefferis commented 5 years ago

@fmlove There are a number of glomerulus annotations that are wrong (e.g. glomerulus VC3m/l).

@schlegelp

I use a combination of annotations. E.g. for right canonical DA1 uPNs I would intersect annotations glomerulus DA1 and WTPN2017_excitatory_uPN_right.

It sounds like this finds RHS PNs that innervate a particular glomerulus but does not tell you which glomerulus a neuron innervates when you are given the skid.

jefferis commented 5 years ago

Hi @SridharJagannathan – I think we did some work on this together on your machine. Could you push?

SridharJagannathan commented 5 years ago

Hi @jefferis I have pushed the latest changes, can you please check

SridharJagannathan commented 5 years ago

Also @jefferis won't it be nicer to add testcases for any additional functionality we implement, I tested the examples offlines but ideally would be nicer to have test cases itself.

jefferis commented 5 years ago

Great. You could def add test cases. These could be live against the server conditioned on whether access credentials exist on the current machine (tracerutils is basically specific to those working with adult FAFB).

Sent from my iPhone

On 1 May 2019, at 16:29, Sridhar Jagannathan notifications@github.com wrote:

Also @jefferis won't it be nicer to add testcases for any additional functionality we implement, I tested the examples offlines but ideally would be nicer to have test cases itself.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

jefferis commented 5 years ago

This is now closed as of 282f8779f412c399239a51d0b6e19d1559b8d82a