search

krassowski / multi-omics-state-of-the-field

Analyses for "State of the field in multi-omics research: from computational needs to data mining and sharing"
https://doi.org/10.3389/fgene.2020.610798
MIT License
24 stars 13 forks source link

Captions and methods description #18

Closed krassowski closed 4 years ago

krassowski commented 4 years ago

Extracting @biswapriyamisra's comment from #12:

[1] I would need a "overview short caption" and a "semi-descriptive caption" for the entire Figure. If you want to have 2/3 Figures (Each figure with multiple - typically 4 panels) and fine too.

[2] Please also send me the final 4-5 sentence "methods" description for your analysis and link to Git page that will be eventually shared with publication online.

krassowski commented 4 years ago

Latest version:

image

Self-reflection

WIP captions

vd4mmind commented 4 years ago

Hi @krassowski , Here are my comments on the pic.

  1. Figure A would definitely need a frame to make it more distinct. See you can add this.
  2. Figure B would need 2005 label on the x-axis as our coordinates(x,y) are (2005,0)
  3. Figure C: all, 4 categories are fine but let's reduce the text a bit so as to not cram the CI in the text. Color is fine with me.
  4. Figure D: Remove the color gradient as this is just ranked numbers. Make the text small by 0.2-0.5 if possible and see the figure. I feel removing the color gradient, the cramming issue of text will solve and spaces can be more.
  5. Figure E: This is a bi-axis plot with numbers on both sides of x and x' I believe? I am fine with this.
  6. I think in WIP Caption description for C) is missing.

Can you just remind wherein MS will you fit Figures A, B, C, and D? I think I missed a bit on this when I re-read the text.

This figure is really great and reveals a lot of points in our benchmarking section, where we argued the usage of multi-omics terms mostly >2 and over-representation of cancer. So we will need to put the figures indexes in those sections accordingly.

Let me know if anything is unclear.

Kind regards, Vivek

biswapriyamisra commented 4 years ago

Hi Vivek,

Can you please work at your end on the "WIP captions" and edit to make it more "finalized/ readable/ for clarity" etc. (and also add description for "C"):

WIP captions

On Mon, Aug 3, 2020 at 9:35 AM ivivek87 notifications@github.com wrote:

Hi @krassowski https://github.com/krassowski , Here are my comments on the pic.

  1. Figure A would definitely need a frame to make it more distinct. See you can add this.
  2. Figure B would need 2005 label on the x-axis as our coordinates(x,y) are (2005,0)
  3. Figure C: all, 4 categories are fine but let's reduce the text a bit so as to not cram the CI in the text. Color is fine with me.
  4. Figure D: Remove the color gradient as this is just ranked numbers. Make the text small by 0.2-0.5 if possible and see the figure. I feel removing the color gradient the cramming of text will solve and spaces can be more.
  5. Figure E: This is a bi-axis plot with numbers on both sides of x and x' I believe? I am fine with this.
  6. I think in WIP Caption description for C) is missing.

Can you just remind wherein MS will you fit Figure A,B.C, and D. I think I missed a bit on this when I re-read the text.

Let me know if anything is unclear.

Kind regards, Vivek

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/krassowski/multi-omics-state-of-the-art/issues/18#issuecomment-667791338, or unsubscribe https://github.com/notifications/unsubscribe-auth/AGUCRIFUABMWWC3O6BI66HLR6YZR3ANCNFSM4PS3Z4UA .

vd4mmind commented 4 years ago

I will. Let’s finalize the image first tomorrow with Mike and then I will fix this. It doesn’t need much change but just a bit more description to make it more explicit. But that also means I would need to see the flow diagram as it helps to prepare the narrative for the legend. Does that help @biswapriyamisra ?

biswapriyamisra commented 4 years ago

Exactly my points! Thanks!

On Mon, Aug 3, 2020 at 11:03 AM ivivek87 notifications@github.com wrote:

I will. Let’s finalize the image first tomorrow with Mike and then I will fix this. It doesn’t need much change but just a bit more description to make it more explicit. But that also means I would need to see the flow diagram as it helps to prepare the narrative for the legend. Does that help @biswapriyamisra https://github.com/biswapriyamisra ?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/krassowski/multi-omics-state-of-the-art/issues/18#issuecomment-667812163, or unsubscribe https://github.com/notifications/unsubscribe-auth/AGUCRIAXWLEHN2YUOFT5KHDR6ZD3XANCNFSM4PS3Z4UA .

krassowski commented 4 years ago

Thank you for the comments. Quick answers:

Figure B would need 2005 label on the x-axis as our coordinates(x,y) are (2005,0)

This is no longer the case, because I am now plotting the "complete years". If I retrieved all articles dated 2005 from PubMed, then I got all articles published up until the end of 2005, this is 31 December 2005 which basically is 2006. Also, for 2020 all the articles published up until 25 July 2020 which is why the plot ends after 2020.

Figure E: This is a bi-axis plot with numbers on both sides of x and x' I believe? I am fine with this.

Yes, the lower one is to highlight that this is log10, which is not easy to infer from the numbers alone.

I still have to do some DPhil stuff so will get back in a few hours.

vd4mmind commented 4 years ago

@krassowski , thanks for the responses. Let me know when you would like to go over the latest figures after implementing the suggestions and also generating the flow. :)

DPhil work first and this is all curiosity-driven. :)

krassowski commented 4 years ago

Hopefully, this is moving in the right direction:

image

not sure about the borders now though... a few alternatives:

vd4mmind commented 4 years ago

Hi @krassowski ,

Wonderful. I like it. Few comments and queries.

  1. Is it possible to move the red gradient bar above Figure E? Like just below it’s representative Figure D in a similar horizontal manner? As currently a reader might misconstrue it with red circles in Figure E. Or use a different color gradient scale like blue-green or shades of grey. Any thoughts?

  2. I like the borders. It keeps the images distinct from one another. Let the reviewer come back at us on that part if we need to change them.

Let me know if it’s clear.

Kind regards,

Vivek

krassowski commented 4 years ago

Ok, first serious attempt at captioning:

Figure X. Characterisation of multi-omics literature based on a systematic screen of PubMed indexed articles (up to July 2020). The details of the methods with reproducible code are available at github.com/krassowski/multi-omics-state-of-the-field. The comprehensive search terms (see the online repository for details) were collapsed into four categories; integrated omics () includes integromics and integrative omics, multi-view (**) includes multi-view|block|source|modal omics, other terms (**\) include pan-, trans-, poly-, cross-omics.

The subpanels present: A) Combinations of omics (grouped by the characterised entities) commonly discussed occurring together in multi-omics articles (intersections with degree > 2 and at least 50 papers). The proteins group (1) also includes peptides, the metabolites group (2) includes other endogenous molecules, the epigenetic group (3) encompasses all epigenetic modifications. B) The number of multi-omics articles indexed in PubMed is rapidly increasing (also after adjusting for the number of articles published in matched journals - data not shown); the dip in 2020 can be attributed to indexing delay which was not accounted for. C) Articles of various types mention different numbers of omics; while it is understandable that multi-omics reviews discuss many omics, the computational method articles appear to lag behind all other article types. The detected number of omics may underestimate the actual numbers (due to the automated search strategy) but should put useful lower bound on the number of omics discussed. D) The number of articles mentioning the most popular clinical findings, disease terms (screening based on ClinVar diseases list) and species (based upon NCBI Taxonomy database). Both databases were manually filtered down to remove ambiguous terms and merge plural/singular forms. Only the abstracts were screened. E) The detected references to code and data versioning and distribution platforms and systems (links to repositories with deposited code/data); both the abstracts and full-texts (open-access subset, 77% of all articles) were screened. No manual curation to classify intend of the link inclusion (i.e. to share authors code/data vs to report the use of a datset/tool) was undertaken.

Edit: added E. Edit 2: fixed some grammar. Edit 3: moved to a PR: https://github.com/krassowski/multi-omics-state-of-the-field/pull/19 - if we are using GitHub, we may as well do it properly.

krassowski commented 4 years ago

Is it possible to move the red gradient bar above Figure E? Like just below it’s representative Figure D in a similar horizontal manner? As currently a reader might misconstrue it with red circles in Figure E. Or use a different color gradient scale like blue-green or shades of grey. Any thoughts?

Good point that we have three different things labelled with red. Green might do, but also could change the colors in E.

vd4mmind commented 4 years ago

I like the caption. Detailed and clarified. Provides the rationale. Add the text for E) and then we are done. :)

Change the color gradient in Figure D to something else that’s not repetitive of other colors.

I missed the part of red color in E. Should not overlap with color from A) and B). Helps it to keep things clear, distinct and simpler.

Does that help @krassowski ?

Best,

VD

krassowski commented 4 years ago

I did not find a satisfying colour combination other than what was already there... but I tried to fix the problem by adding letter tag references and adding shapes to D so it is also more colourblind-friendly.

image

I will have a go on the flow diagram tomorrow.

vd4mmind commented 4 years ago

Hi @krassowski

That works. Let’s keep it like this and see what reviewers might have to say.

Now let’s wrap up the flow diagram tomorrow and then you can finish the WIP caption for E) and for the flow altogether. Does that sound good? Impressive work. Keep it up.

Best,

VD

krassowski commented 4 years ago

I added the E caption by editing my previous comment above. I will certainly do one iteration more before saying it's ready and sending SVG fille to @biswapriyamisra, but it will be only tiny details, nothing major should change now.

Notes to self:

vd4mmind commented 4 years ago

Probably also make a TIFF format along with SVG. As often such file formats vary as per journals. I am unsure which tool @biswapriyamisra will he using at his end to compile the comprehensive MS with images + boxes + images.

Best,

VD

biswapriyamisra commented 4 years ago

Mike: Thanks for the cool work- please remove the 'panel titles' on the top of each one. Most journals would not like them. Just A, B, C, D, E are fine, and the descriptions will be in the caption.

Vivek: Please edit the captions in terms of readability and clarity and finalize to send to me to help me add to the MS.

Mike: Any update on the "flow diagram" and GitHub link to share and 3-4 lines methods description ?

On Tue, Aug 4, 2020 at 5:56 AM ivivek87 notifications@github.com wrote:

I like the caption. Detailed and clarified. Provides the rationale. Add the text for E) and then we are done. :)

Change the color gradient in Figure D to something else that’s not repetitive of other colors.

I missed the part of red color in E. Should not overlap with color from A) and B). Helps it to keep things clear, distinct and simpler.

Does that help @krassowski https://github.com/krassowski ?

Best,

VD

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/krassowski/multi-omics-state-of-the-art/issues/18#issuecomment-668304818, or unsubscribe https://github.com/notifications/unsubscribe-auth/AGUCRIDXOH2SWMOLT22F6ILR65IUZANCNFSM4PS3Z4UA .

vd4mmind commented 4 years ago

Hi @biswapriyamisra

We will address the flow tomorrow the first thing when Mike is available. Once that is done we will make the legend and a short text. As for the current image, I am fine with the narrative. If you need more information, let us know and we can expand it.

Best,

VD

krassowski commented 4 years ago

please remove the 'panel titles' on the top of each one. Most journals would not like them. Just A, B, C, D, E are fine, and the descriptions will be in the caption.

I will push back on this one - If a journal tells us to do so, I will remove them, but for now, I believe this is the best way to present this figure:

biswapriyamisra commented 4 years ago

Hi Mike, it's not personal preference of mine or yours, and even Frontiers may not push against/ for it, but I see it as a non-standard. Reason being, when they want to change the aspect ratio of figures and zoom/ stretch for print/ on line, the "fonts start to look weird", not their problem, but ends up being our problem and end result! let us keep it and see how it goes, not pushing for this minor issue, given your effort.

Just examples in most figures for panels: https://www.google.com/search?q=figure+panel+phytcohemistry+assay&tbm=isch&ved=2ahUKEwjCttSFzoHrAhWKFisKHSdSCYMQ2-cCegQIABAA&oq=figure+panel+phytcohemistry+assay&gs_lcp=CgNpbWcQAzoECCMQJzoCCAA6BggAEAUQHjoGCAAQCBAeOgQIABAYOgQIABAeULNVWOqiAWCopQFoAHAAeACAAdoBiAHuF5IBBjE2LjkuMpgBAKABAaoBC2d3cy13aXotaW1nwAEB&sclient=img&ei=kVwpX4KqBoqtrAGnpKWYCA&bih=694&biw=1536&client=firefox-b-d

I will rather wait for

Thanks, Biswa

On Tue, Aug 4, 2020 at 6:21 PM Michał Krassowski notifications@github.com wrote:

please remove the 'panel titles' on the top of each one. Most journals would not like them. Just A, B, C, D, E are fine, and the descriptions will be in the caption.

I will push back on this one - If a journal tells us to do so, I will remove them, but for now, I believe this is the best way to present this figure:

  • if the reader has to go to the caption to understand the figure, it detracts from its utility; I believe that one should understand which panel presents what without having to go to the caption
  • if I move the subtitles back to x-axes, it will increase the overall height thus the contents of the panels will need to be reduced and will become less legible; this is because some panels do not have x-axes which benefit from labelling and we would just get irregular white wasted space here and there (e.g. upset which is already taller than others).
  • reading papers I often find the panels referenced by letter only to be mixed up between the caption and the figure; this is just anecdotal evidence, but one should easily see why using letters only - where an entire sentence can be placed - takes away an opportunity to cross-check if everything is ok
  • the fact that many researchers are doing so is not convincing alone; many are abusing dynamite plots and this is not an argument to use those (speaking of which, we could add density or rugs to C; the problem here is that it is discrete and the "outlier" of reviews which enumerate all omics terms would make this plot illegible as a panel plot, but probably could have a larger one in the repo)
  • I quickly checked submission guidelines of a few journals, did not find such a restriction
    • some specify whether panels should be labelled with upper or lowercase letters, but have not found one which forbids extra text next to the labels.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/krassowski/multi-omics-state-of-the-field/issues/18#issuecomment-668577441, or unsubscribe https://github.com/notifications/unsubscribe-auth/AGUCRICHSMU4R2OUMA3AT4DR6776TANCNFSM4PS3Z4UA .

krassowski commented 4 years ago

If aspect ratio will need to be changed I can update the plot in less than 30 seconds - the benefit of having this automated rather than post-processing manually :)

I will work on the flow diagram and the methods paragraph tonight. In the meantime, feel free to comment on the captions, see #19 (I figured having everyone just comment here might not be the best way to get this done, hence the PR).

biswapriyamisra commented 4 years ago

OK, sure!

On Tue, Aug 4, 2020 at 6:42 PM Michał Krassowski notifications@github.com wrote:

If aspect ratio will need to be changed I can update the plot in less than 30 seconds - the benefit of having this automated rather than post-processing manually :)

I will work on the flow diagram and the methods paragraph tonight. In the meantime, feel free to comment on the captions, see #19 https://github.com/krassowski/multi-omics-state-of-the-field/pull/19 (I figured having everyone just comment here might not be the best way to get this done, hence the PR).

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/krassowski/multi-omics-state-of-the-field/issues/18#issuecomment-668587678, or unsubscribe https://github.com/notifications/unsubscribe-auth/AGUCRIBC3KFIB4KHRMVSFD3R7ACL5ANCNFSM4PS3Z4UA .

vd4mmind commented 4 years ago

Hi @krassowski ,

Dynamic reporting and PR is all good but I would request you send Biswa, the images (multi-panel + flow diagram) once we have finalized along with the Captions, 3-4 lines of text for incorporating within MS, proper GitHub link, that’s in the master and not branch with captions, images, etc so that he can make entries into the MS.

He is currently unable to send the MS for language editing. We shouldn’t delay this. That would be a request. He is a bit lost here. So best is to finalize and send all the above in an email to him. Can we aim to finish this by tonight?

Best,

VD

krassowski commented 4 years ago

That's the plan

krassowski commented 4 years ago

Closed in #19.

krassowski commented 4 years ago

Well, actually the description of the methods is still TBD...

vd4mmind commented 4 years ago

@krassowski your legend is pretty descriptive. The method does not need to be explicitly written for submission. Even the flow diagram should be enough if boxes are well defined.

krassowski commented 4 years ago

I have some things that I want to include and some things that need to be included. Will be short though (3-4 sentences).