updated protein 1 tutorial #40
Conversation
|
Heyo! We've talked about reviewing a few tutorials and this was definitely one of them. Especially because a lot of it is now deprecated (the entire gomodel routine). If we are fixing part of it I would suggest we also fix this bit? If you're interested @csbrasnett you could have a go at it, or I can also have a look in the next few days. Basically making the tutorial in line with the GoMartini paper, using martinize instead of the creategovirt.py script. I would say we could then have Paulo @paulocts , for example, give it a read through before we merge this. What do you think? |
|
Ah nice! Yep happy to have a look at the rest of the tutorial as well so we get this done in one PR. I’ll have a look in the next few days and push it so you guys can have a look |
|
@Lp0lp let me know what you think about that. It occurs to me that it might be useful to have a section here about OLIVES, but I'm not sure I'd want to write that ahead of @KasperBuskPedersen ? |
|
@csbrasnett please go ahead and write a small section of how to set up OLIVES if you find that useful. You can basically just copy paste from the OLIVES github. Alternatively just write that OLIVES is an alternative and link the github? Unfortunately I don't have time to write or test that atm. |
|
I think it'd be nice to have everything in one place as much as possible, I'll try to cook something up based on what you've done and also link through for more examples with the varying quaternary structure interaction strength so we don't overcomplicate too. This is defo out of scope atm, but is there any desire to implement olives into martinize2 at some point? I realise it'd be a lot of work, but it might be nice in terms of workflow optimisation? I'd definitely appreciate it for one (and would be up for trying to do it when I have some time)! |
|
@csbrasnett Actually, I think there is a risk of doing new users a disservice by automating it too much. Since olives is more or less additive to the standard martini model, it makes sense to keep them separate. With the new protein model of Luis', there are several reasons to keep martinize2 and olives as two separate calls, e.g. multistate olives would require you to martinize two proteins in 2 separate calls followed by one olives call etc.. I wont be able to maintain martinize2 going forward, but can support olives alone. From a software maintanece standpoint it is also not good to combine everything into a mega program. Though, we could add olives to vermouth, so it would become a command line call like martinize2, such that you won't have to carry the python script around, but again one could just add olives to their .bashrc. Best Kasper |
|
@Lp0lp @paulocts @KasperBuskPedersen hopefully this satisfies everyone now! |
…protein-1-tutorial-update
|
I will have a look and get back to you! |
|
@Lp0lp did you get a chance to have a look yet? |
|
@csbrasnett not yet sorry. Had to tackle a couple of grants and currently on the road. Worst case I'll have it checked out by Monday. Sorry! |
Small changes to wording/formatting. Added old version of the tutorial as a legacy tutorial.
Added Martins collab tutorials. Updated contributor list to reflect Chris and Martin.
|
Heyo! @csbrasnett I am happy with the tutorial. I tweaked some wording and some formatting but nothing major. I've sent it to @paulocts so that he could have a look as well. I added the old iteration of the tutorial to the legacy tutorials. I also added to the tutorial list the two new collab tutorials @mnmelo and @martincalvelo cooked up. They end up working great in Quarto, since quarto automatically renders I updated the tutorial author list to include @csbrasnett and @martincalvelo . |
|
great, thanks! All sounds good to me. To note, I'm currently looking at changing the behaviour of |
|
Hey, this looks great. I'd add a caveat here, though: Of course, non-Colab users should just provide their own dependencies and skip those installation cells. What's your take? |
|
On all the other tutorials we expect the user to already have all the
required compilations ready to go, so I don't see skipping these cells as
anything too confusing/limiting in comparison. Maybe I am wrong though...
In any case, these two tutorials can be a sort of test to see if people run
into issues in case we want to expand on this kind of tutorial.
Big warning at the top reinforcing these were prepared for colab and to
skip the installation/step 0 cells If running on own compilations, I guess?
Alternatively, maybe adjust the notebook such that it knows whether it is
running in a collab or not?
If in_collab:compile; else: skip
Something like this:
https://stackoverflow.com/questions/53581278/test-if-notebook-is-running-on-google-colab
Manuel Nuno Melo ***@***.***> escreveu (segunda, 21/10/2024
à(s) 18:13):
… Hey, this looks great. I'd add a caveat here, though:
Our tutorials were prepared for colab and if run blindly will download and
install gromacs and dssp binaries as well as some pip dependencies. The
most breaking stuff will probably fail unless run as root because it will
attempt to unpack binaries onto system directories, but I imagine it can
still be confusing...
Of course, non-Colab users should just provide their own dependencies and
skip those installation cells. What's your take?
—
Reply to this email directly, view it on GitHub
<#40 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AKHTB55OXD6IDTV2PFV4V4TZ4UR3DAVCNFSM6AAAAABOVPE742VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDIMRXGEZDIMBTGA>
.
You are receiving this because you were mentioned.Message ID:
<Martini-Force-Field-Initiative/Martini-Force-Field-Initiative.github.
***@***.***>
|
|
I like that kind of automation. It's safer. @martincalvelo, do you think you could try to incorporate that? BTW, the code doesn't actually compile. It downloads a zip of already compiled GROMACS that I prepared and make available. |
|
Hi! I updated the two tutorials to detect whether they are running on Google Colab and, based on that, to install or not gromacs, dssp, and some python libraries. I also made some other small modifications to avoid incompatibilities when running locally. Tutorial 4.1 Ceramide — VDAC1 binding in a membrane Tutorial 4.2 Lipidomics study: Mag2 in a POPC and POPC-POPG 3:1 membrane |
…-scfix in line with new cli for martinize
One user over in the discussion forum has had issues with working out the input string for secondary structure specification. From my memory, the previous version of the tutorial had -dssp as the default argument, but I guess when it was revised that was changed for people who didn't have it installed?
However, I'm not sure if how it's been revised is particularly clear for people who aren't already familiar with martinize2 functionality, so I've tried having a go at a small rewrite to hopefully reintroduce some clarity. I realise this is a matter of personal opinion, so please let me know what you think!