Thank you for your contributions to the BANC dataset! Proofreading is moving along at an impressive pace, with over half of the cells in the brain completed.
Today we ask for your help this holiday season by identifying specific visual projection cells in the right optic lobe: LC cells, LPCs, LLPCs and LPLCs. These cells enable the brain to recognize objects and sense movement, translating sight into behavior and decision making (think: should I run from this frog or fly toward that tasty banana). They are of great scientific interest in the BANC as we can finally see their connectivity to VNC. But they have not yet been identified in BANC. There are over 40 different types of LC cells.
Krzysztof has a great overview of these cells in this post.
Resources:
Find example cell types on the OL Boundary Cell Types page in Codex. Ctrl+F for LC, LPC, LLPC, LPLC
Thank you for your help in finding and identifying these cells in the BANC! If you have any questions, please feel free to add them in this thread. These cells will be added to the BANC-bot shortly and this post will be updated when they are.
Amy, Marissa, Mala, Sebastian, and the FlyWire team at Princeton
Cool! I’ll start right away.
It would be useful to have annotations defined for them though (including the hierarchy, like we have for the optic lobes’ ones).
For now, we can (probably) label them only as a freeform annotation.
Hooray! Good point about annotation list. We can work on tjos. If you would rather take the first pass, feel free! Marissa is double checking the correct OL
I’ve been working on collecting the LC4s into a single set. Up to 27 so far (FAFB had 50).
Do we have a clear way on identifying these yet? There are actually a few LC4s in the dataset (Neuroglancer) but I cannot figure out how that was added, unless it was before the current banc bot setup.
Edit: I’ve got 54 (two damaged/incomplete) LC4s now - there seems to be one or two missing based on the gap in the canopy but I haven’t been able to locate them. Set complete here: Neuroglancer
LC4s appear to be done (other than one or two I think are missing), outside of getting cell info added. Link’s above.
Since I’ve found a good bundle area to work from I’ll be continuing on LC types- next up is what appears to be LC12s as they are also common in the area I’m working from (might actually be two groups of LCs here). Neuroglancer
Small update so folks know where I’m working. Haven’t been able to do much around the holidays but trying to get back to it.
“Complete”: LC4s - Posted earlier, pending Cell IDs. There may still be some strays, I might go over them again.
“In Progress”: LC12s LC17s - up to 155! Tentatively complete.
I split these two into different groups now that I’ve realized/seen the difference and was able to visualize their terminal clusters. Now working through their terminal clusters to fill out these sets (FAFB had about 140+ each, so lots here to go). Admittedly 80% of my work is just finding something already proofread and adding it to the list for cell ID later, but still important work! Since these two are so clustered together I’m working on them in tandem.
With this, I am tentatively calling LC17s “done” - at 155 neurons grouped into this farm (segmentation layer is turned off in this link for your loading sanity.)
There are likely a few more stragglers in here, but the farm size is getting difficult to manage in terms of memory load. FAFB has 142 labeled, so this feels approximately complete!
I’ll be moving on to finish the LC12 farm next. Currently spending more time on Pyr during the weekdays but weekends will likely be more able to focus on the project to identify these VPCs.
(Speaking of, when do we get the ability to mark these with cell IDs?)
Thanks KK! Yeah, let me touch base with the BANC team and possibly Arie about how they may want to track and/or are able to track progress… I’ll make sure to bring it up at the next meeting.
I really like, that in the catalog of the intristic neurons there are images next to each type. Unfortunately, the catalog of the visual projection neurons lacks it, so I decided to do it on my own. I’ve made it in form of a website.
It’s available for download here: https://drive.google.com/drive/folders/1qIpc5qfJVMtv9upanEnOm5-r6i_KZSoy?usp=sharing
Just download everything and open the index.html file in the browser.
Each entry contains a single type with it’s name and two projections of a single cell of this type. The name is also a link to the original catalog with the given type.
I was thinking about deploying it somewhere online, but all the images have together about 200MB, so it would take sometime to load the page each time, you’ve opened it. It would also probably eat fast any free hosting sites’ transfer budgets, lol.
