[addon] Cell Identification Helper

#1

I’m currently working on an addon called “Cell Identification Helper” which should help with inputting correct names of cells during the Cell Identification process. The work is in progress, so there isn’t currently any link available for installing it. However, I’m starting this topic to ask you for your opinions and your help.

Here’s how it currently looks:

cih
cih1819×1009 258 KB

It working is simply - you select a “family” of a cell, then select it’s number and a suffix, if needed. You can also check the “Putative” checkbox to add the word before the whole name.

Here are some questions for you:

  1. What do you think about the UI/UX of it (how does it look, how to use it)?
  2. The current list on the left is copied from @Nseraf’s list in the Q&A gsheet:
    a. Is the list complete?
    b. Should it be in alphabetical or any other order, than it currently is?
    c. Should it display both long and short names or only short (maybe only long)?
  3. Should I add other exisitng long and/or short names for already existing cells?
    a. If so, should the alternative names also be displayed in the list or only in the target text?
  4. Should I add an author and a year for all the cells?
    a. Does it should look like (Fischach et al., 1988) or maybe other format, maybe add other authors too?
    b. Should I add initial(s) for autors’ name(s)?
  5. Should I also attach a DOI identificator for all the known cells? (after all, one author can have multiple papers in a single year)
    a. If so, what would be the preferent format?
  6. How does the final name should look like (currently it is like on the picture)?
  7. If there are more than one long and short names, should I display them like that:
  • long_name1; long_name2; short_name1; short_name2
    or like that:
  • long_name1; short_name1; long_name2; short_name2
  • a different way

As for a list of authors and DOIs, I’ve just started making such list. So I’ll try to do as many as possible and then ask for your help for the cells, I won’t be able to find sources.

4 Likes
#2

looks great!!! Nice work :smiley:

  1. ui/ux looks good to me, simple and functional
  2. I got all ‘my’ names from: Optic Lobe Cell Name Guide - Google Slides plus 1-2 names from papers posted in the forums by players etc for illustrations and so on, but i think im missing some like that Mi(new)1 ive sen floating the last few days. I think admins have asked us we have both full name and abbreviation ie: Centrifugal - C2. I tried to go with alphabetical order but after a while it wasnt easy in the gsheet lol.
  3. id say yes
  4. idk, on the one hand it may be of help to either researcher(s) and/or players but on the other hand idk if we need/want to over-convolute the thing with a lot of information.
  5. not sure what this would do/is.
    6 + 7 i like it as is on the picture, Concise and to the point.
2 Likes
#3

Thanks for the replies :slight_smile:

DOI is a permanent link to a scientific paper. Usually there only exists one DOI for each paper.
For examle DOI to the famous Fischbach’s paper is: http://dx.doi.org/10.1007/BF00218858

2 Likes
#4

I see, didn’t know, thanks. How do I see/read the paper? it’s paywalled.

2 Likes
#5

Yes, unfortunately the DOI links often go to paywalled papers. But such paper shows the title and the abstract of a given paper, so it’s enough to know, which paper we’re talking about. If I’d go with full-text links, they might change with time and such info wouldn’t be useful in the identification database. But DOIs are rather permanent. Also, the full-text links aren’t always legal from the copyright point of view.
If you want to find the full-text, just copy-paste the title into Google. Often there are links to the open versions (e.g. on ResearchGate).

2 Likes
#6

  1. Having a list of names looks great, but not so sure about the number panel, will probably be faster to just write in the number on keyboard.

  2. a think the T cells are called Trans cell, not sure where i got this from, but what i have been using in the gsheet
    2b/c i think it would be easiest to use if in alphabetical order with short name only.

  3. for search purposes, it might be helpful with other names also, but then you would need the authors also, but not sure what the policy on this is. Are you sure a cell is grouped the same from different authours for instance on cell named mi in on paper and dm in another( example taken out of the air)
    On the other hand it could go wrong if someone add in a entry based on another paper than what is stated in the referral. The list should be as short and clear as possible

  5. difficult to add a DOI for a group of cells since some cell are in one paper, with new types beeing discovered in another paper.

  6. my preferred naming would be Mi17, medulla intrinsic or if unsure Mi, resembles Mi17, medulla intrisic

  7. definitively * long_name1; short_name1; long_name2; short_name2

i know they use putative to mark beeing unsure about a cell, but i have no idea what it really mean since it is first time i have heard the expression. that make me think it is not a very user friendly term like a question mark or resembles/ looks like. when looking trough the list of cell names i was 100% sure it was some cell that was named putative and did scroll fast past it.

3 Likes
#7
  1. The number panel is useful for cases like Lccn for example, where the entry could look like:
    Lobula complex columnar neuron [number]; Lccn[number]; LPLC[number]; LPL[number]
    If we’d go with putting all existing names. In such cases you entry the number once and it will be placed automatically in all the right places, with correct spaces, etc. You can still enter the number or anything else manually.
    Actually, I forgot to add - if you enter something manually and after that you select something from the list or from the numerical “keyboard”, then your entry will be replaced with what you’ve selected, so if you want to do any alterations, do it after selecting, what you want to select (or don’t select anything, just write the answer).
  2. Never heard of the term “Trans” for the T cells, but will take a look at it. For me alphabetical also makes the most sense. At first I tought about making them in order form the most distal to the most proximal (e.g. retina → lamina → medulla → etc.), but that wouldn’t make much sense, because the groups interwine with each others.
  3. Yes, I’m thiking about keeping the list short and readable. The authors and DOIs wouldn’t be displayed in the list, just in the answer textarea, when user selects a proper name. Also, I’m thinking about adding the authors and potentially DOIs only for the cells, that have well-established names (definitely the Dischbach’s ones, but looks like the ones without the (new) infix are also rather permanent).
  4. The authors and DOIs would be added only, if the script recognized a conrete cell, not a cell group. For example, if you enter “Tm”, it wouldn’t add anything (the full name would be "Transmedullary Y; TmY), but if you enter “Tmy5”, then there then full name would look like: "Transmedullary Y 5; TmY5 (Fischbach, 1988; http://dx.doi.org/10.1007/BF00218858).
  5. Such things as “resembles…” could be difficult to script. I would have to create another list for cells in the “resembles” place or create a switch, so the script would know, if user currently selects the main type or the “resembles” type. In such cases you can just select the appropriate category (Mi in this case) and fill in the rest manually.
  6. Agree.

As for the “putative” I also didn’t hear that word before. But it’s the word, they gave in the examples visible during submiting Cell Identification and it exists in the database of identified cells. That’s why I’ve added it here. If you want other words, I can add them too (just tell me the word(s) and place, where it should be inserted).

4 Likes
#8

1 was thinking more like a box you would write in the number, instead of the number panel, that would set it in too the right places

3 Likes
#9

Oh, good idea. Actually, there’s already a box for that (the input field under the “clear” button), but for now it just shows the currently selected value. User can write there anything (currently an error, should be readonly), but it doesn’t change anything. However, I can do it the way you’re saying and maybe make the field a little bit more prominent.
That way there would still be a way to select the name using only the mouse, but a user would also be able to select the category from the left menu and type in the number and potentially the suffix from the physical keyboard.

2 Likes
#10

I found something like this:
“The compound eye consists of ~800 ommatidia, each containing eight rhabdomeric photoreceptor cells (R cells)” in here.

And for “retinula”:
“a group of elongate neural receptor cells forming part of an arthropod compound eye: each retinula cell leads to a nerve fiber passing to the optic ganglion.” - source

Did I understand it correctly, that retinula is the collection of eight R cells (R1 to R8), and a single cell is named “rhabdomeric photoreceptor cell”?

1 Like
#11

1. What do you think about the UI/UX of it (how does it look, how to use it)?

Looks good to me, although a dropdown for the celltype may be better (some screens may struggle with the full list? Not sure, just a thought).

2. The current list on the left is copied from @Nseraf’s list in the Q&A gsheet:
a. Is the list complete?

Seems to be.

b. Should it be in alphabetical or any other order, than it currently is?

I would personally group it by neuropil, going from Re, Lam, Med, Lob, LP, Misc

c. Should it display both long and short names or only short (maybe only long)?

Long and short as currently shown.

3. Should I add other exisitng long and/or short names for already existing cells?

Not sure what you mean exactly by this question.

4. Should I add an author and a year for all the cells?

I would say no, simply because we may have multiple references for a single cell type that differ, and that we might find/discover along the way. I would make this an encouraged but optional thing to add from Flyers.

I’ll note that I have intention at some point of creating a catalog per neuron type of all official and unofficial imagery per cell, and with that I can include the proper citation for each image for those who want to reference them during cell ID.

5. Should I also attach a DOI identificator for all the known cells? (after all, one author can have multiple papers in a single year)

See above

6. How does the final name should look like (currently it is like on the picture)?

The way you have it in the picture is how I submit my IDs, so it looks good to me.

7. If there are more than one long and short names, should I display them like that:

► long_name1; short_name1; long_name2; short_name2

Fischbach refers to the R1-6 cells as “retinal fibres” and others refer to them as “photoreceptors” ex:

Within the retina, photoreceptor cells (R-cells) are organized into ∼800 regular subunits, called ommatidia. Each subunit contains eight R-cells (R1–R8): R1–R6 axons innervate the lamina, whereas R7 and R8 axons pass through the lamina and terminate in two of ten distinct neuropil layers in the medulla

2 Likes
#12

To follow up:

In biology, the term is commonly used to describe an entity or a concept that is based on what is generally accepted or inferred even without direct proof of it, meaning it denotes something, like an inference or a supposition, that is accepted because it is deemed to be the case or what has been commonly believed even without solid evidence to back it up.
Putative - Definition and Examples | Biology Online Dictionary

Putative genes can share sequence similarities to already characterized genes and thus can be inferred to share a similar function, yet the exact function of putative genes remains unknown.
Putative gene - Wikipedia

In short, “putative” in science is a fancy way of saying “it looks like a duck and it quacks like a duck, so I think it’s a duck - but I haven’t tested it scientifically to make sure it’s a duck.”

4 Likes
#13

Awesome, KK! The current list in the gsheet is missing Tm3Y and Tm5Y (from Fischbach), btw. As for the rest, I don’t have much to add, as others have already responded brilliantly, in the above.

1 Like
#14

It does now :stuck_out_tongue:

1 Like
#15

@Nseraf Nice! That was quick! :smiley: I forgot to mention Tm1a, Tm4a, and Tm18a (also from Fischbach), but those variants are pretty rare, in my experience so far. Nice job on the list though…it is incredibly thorough, lol. :slight_smile:

#16

ad. 1. The current version of the list fits into 1280x720 screen (tested). That’s why some other things in the popup window has smaller fonts and margins. I’ve made that this way to save us one click. If it were to be more types, I can do it the dropdown style. It’s just a single attribute to remove, so an easy fix, if needed.

ad. 2b. Maybe I’ll do it in groups - by neuropil and then alphabetically inside the groups

ad. 2c. We are probably more familiar with the short names, so I was thinking about keeping long and short names but change the order - first the short name, then the long one.

ad. 3. Some cells can have multiple long and short names, for example:
Lccn - lobula complex columnar neuron (Fischbach)
LPLC - lobula plate-lobula columnar
LPL - lobula plate-lobula neuron
I was wondering, if I should add all the names and abbreviations or just a single one (which one?)

ad. 4. Ha, I was also going to do similar list - [neuron’s type name; title of the paper, where the type was first mentioned; year and autors of the paper; link to the full version of the paper, if exists; DOI]. I was also thinking about adding images.
Maybe we could do it as yet another tab in the gsheet. It would be useful for the players, who would like to add the author’s name to the identification and for the people, who would like to read more about such type. And the @annkri’s list would be like crown jewel, with all cells actually found and confirmed in our dataset. And, ofc, still very useful for confirming the look of cells during identification.

ad. “R cells”. Maybe then we should just stick to the short version of this cell’s type? (R1, R2, etc.)

2 Likes
#17

That makes good sense, I’m for that :slight_smile:

I would go with Fischbach and perhaps also list (also LPLC/LPL) at the end?

That’s what I think would be best, imo.

3 Likes
#18

Wow this is amazing!

2 Likes
#19

Great idea!

General note to all thinking about cell naming. As we’re trying to bring consistency and clarity to our naming conventions please review this document for the most up-to-date naming conventions we currently have, and please label your cells accordingly. I did recently make a few tweaks to after a review.

If anyone has additional tweaks they would like to suggest I’ve also created this google form where you can submit suggested edits. Please try to be conservative with the amount of edits you suggest, and make sure they are necessary and not already satisfied by the current names.

Thanks to everyone for the enthusiasm and collaborative spirit you are bringing to FlyWire!

5 Likes
#20

And what about cells, that aren’t in the document, but have well established names? Like higher numbers of Mi, Tm, TmY, Dm, Y etc. cells. For example, The document shows Y(1-6), but there are at least up to Y12 described in some scientific papers, and some of the higher Y’s were probably already found by us.

Also the table at the beginning of the document shows Dm(1-8), while links with examples show Dm20 (the google form link tells me, that I don’t have permissions, so I’m writing it here).

1 Like