Great, thanks!
1 Like
I feel a little responsible for that proportionally small amount of SACs being done, lol. Especially, that I was quite active with the MFs earlier. Iām just not a huge fan of thin long wiggly branches. Hereās how it looks for me:
- annotating > proofreading,
- straight thick branches (MFs) > twisty thin twigs,
- condensed convoluted cells (e.g. bipolars) > large sparse cells (SACs, parasol-like neurons, etc.).
Thatās why, after initial interest in the SACs, Iāve switched back to annotating in BANC. Especially, that thereās still quite a few unannotated cells there.
3 Likes
Itās all good, not every cell type is for everyone!
Thanks for the feedback on your interests, itās helpful to know!
2 Likes
Iām kinda the opposite, lol. I tried doing annotation cell type work in FW and got bored really fast (more or less when I stopped working in the FW dataset), having to learn each cell type from every other and then sub-types ME 1 vs ME 2 vs ME 2a etc is just not something that peeks my interest. I am more of a proofread kinda guy and go through as many different cell types as possible to see all the different shapes, forms and sizes there might be in any new dataset. SACs have a special place in my heart (along w/ BPs and the long āspiderā like neurons we used to get in EW marathons) b/c they are cell types that have a lot of funny, happy memories āattachedā to them and I like the challenge of finding impossible connections (or impossibly thin?) that end up working out (not mergers).
I think the only thing I like more than SACs and/or datasets of new āanimalsā and/or ālocationsā (cortex, retina, hippocampus etc) is cleaning up mergers, like EW2 has, like Pyr has/d and very much like the BANCās giga-mergers, cleaning them up and claiming all of those cells for āmeā (well as much as this is applicable in pyr/ew2 w/o lightbulb and/or BANC bot to backbone proofread stuff) and knowing that exactly b/c they were in mergers thereās little to no possibility iāve been unknowingly co-working in a same cell w/ anyone else concurrently.
Which is why before ew2 I used to do pyr, then when gsheet entries would run out (and until we were given new ones) id go temp back to unmergifying stuff in banc. With ew2 in the mix itās probs gonna be similar, ew2 > pyr > banc mergers, or ew2 > banc mergers if no pyr avail.
I will say though that idk why but what i really like in ew2 and pyr vs banc is split is SOOOOO much easier, both selecting red/blue points and then the actual splitting is just somehow better/faster. Also in pyr and ew2 merging is so much faster than banc! Cant compare w/ FW b/c I donāt remember how merge/split behaved there, plus in FW just being able to split stuff vs having to leave large gaps like we used to do in EW was just so novel that probs overwrote any potential āissuesā that might have existed that I notice-d in BANC etc.
2 Likes
And again, Iām the polar opposite - I really donāt like mergers and usually release and audible āsighā whenever I see even a small one 
However, I donāt mind merging stuff, especially with the multimerger toll, we have for some time now and the automatic refresh of the meshes after a merging or splitting operation.
1 Like
lol, we release whole different sighs when we see mergers. Contempt and despair vs contentment and joy. 
1 Like
And Iām so happy, that someone is taking care of those mergers, especially in the lamina. When I saw in the first time I was convinced, weāll have to skip annotating all the R, L and C neurons, lol.
1 Like
SBs do not have axons, right? This is -just- a fused merger at the CB and not a direct proof of otherwise? lol
2 Likes
According to this article (search for " Inner Plexiform Layer Neurons: Amacrine, Interplexiform, and Ganglion Cells") they can have processes, that might be axons, however, it has not been determined yet.
1 Like
Thanks for all the feedback! Something for everyone I suppose! 
2 Likes
Hmm very curiousā¦ I would not expect to see an axon on these cells, buuut this connection is looking somewhat convincing. Let me ask for some 2nd opinions and get back to you.
2 Likes
Okey dokey, thanks! And yeah everything I have come to know from eyewire on is thatā¦SACs have no axons but then that connection is very solid in the 2D and 3D and idk lol
I can ofc re-merge it and proofread it if the scientist(s) let us know that they think it does in fact belong.
1 Like
I found a second similar SAC! (not the cell Iām proofreading but merged into a branch that belongs to my cell)
1 Like
Merger was found on the SAC cell in question: neuroglancer
I also got confirmation that SAC cells would not be expected to have an axon like this so I would look for a merger on any similar cells with this issue!
3 Likes
that makes more sense! I should have looked at that thing more. Thanks! And I spent an hr trying to fill that gapā¦lol
2 Likes
if anyone wants to help (me) find something connecting this branch to a cell, Iād be much obliged.
https://spelunker.cave-explorer.org/#!middleauth+https://global.daf-apis.com/nglstate/api/v1/5332014661107712
Iāve -also- left a Need Help status (line 107) in the gsheet. Thanks in advance.
I found something like this:
1 Like
No problem! Cells up to their old tricks as usual.
@Nik I know for some of the SAC cells you have some other cells listed in the comment. For any that are proofread, could you move them onto their own lines in the spreadsheet? I added two cell types ā NEW-SAC for added SAC types and NEW-Other for non-SACs.
But please finish with the proofreading of the remaining cells on the list first, as that is the most pressing objective atm!
1 Like
will do, and yep am doing so lol, weāre down to 7 entries not claimed/done.
1 Like
yeah that fits, thanks!
1 Like