I want to try KALP-15 in DPPC tutorial, because I will be using membrane and peptide systems for my dissertation, but the beginning of the tutorial it says “As of summer 2023, the required Berger lipid parameters are no longer available online. The authors themselves have raised issues of their validity and they are not available. This tutorial will remain online as a reference until I rewrite it. Please DO NOT proceed and expect the tutorial to work. It will not.”. So how can I try the tutorial? Can I use other lipid parameters with this tutorial or should I wait? Any answers would be appreciated.
I was actually able to finish the tutorial before summer 2023 but never had a chance to analyze the ‘results’, and since then have forgotten most of what I did. As its been a while, I am redoing the tutorial but I am now stuck with the “Berger lipid params” issue as well. Have you gotten around this?
I found a repository online (https://lipidbook.org) that provides various lipid parameters and will try to rerun the tutorial with these ‘new’ params (and compare it to the old “Berger” results).
Lmk if you are still interested in completing these tutorials. I am new to MD simulations and it would be nice to learn/follow along with another novice (my dissertation is also focused on biomembranes but with an emphasis on antimicrobial peptides).
It’s important to stress, given you are new to membrane simulations, that each of the available lipid force field parameter sets must be used with their appropriate mdp settings, as well as things like the correct water model. While this is generally true for all simulations, membranes can be especially sensitive to changes made in such types of settings.
I’ve not looked at the tutorial in ages but setting up membrane simulations with other force fields should be a fair bit easier than messing around adding the Berger parameters to force field files that was previously required. One of the easiest approaches is to use the CHARMM-GUI to generate any membrane structure you want. The output also provides all the files you need to run simulations of this membrane with the CHARMM36 lipid parameters in GROMACS. This same membrane structure can also be used straight away with the Slipids force field as it uses the same atom names and ordering as CHARMM. I have also made versions of GROMOS-CKP lipid topologies that also follow such atom naming and ordering so all that needs done here is to remove any unwanted non-polar hydrogen atoms in the CHARMM-GUI output with grep. Happy to share these with anyone if they’re interested? (I had been planning on making them available with a manuscript on using these and other GROMOS lipid parameters in GROMACS that has been stuck at the ~90% completed stage for far too long due to other commitments).
If you’re interested in differences between how different lipid parameters perform, there are lots of publications out there including some recent articles from the NMRLipids consortium.
As the author of the tutorial, I want to echo everything @TomPiggot said. The protocol was state of the art about 15 years ago. Now, it is nothing more than perhaps an advanced exercise in working with GROMACS and editing force field files to understand them. Don’t use the protocol for real simulations and certainly don’t use the force field, even if you can find the files. Among other updates, I’m planning a complete rewrite of the membrane protein tutorial. It’s long overdue.
Actually using the “Berger” force field is an interesting one.
For PC lipids it actually works pretty well and as the lipids diffuse faster than (most?) other lipid models there could be some potential use cases (e.g. for maximal sampling lateral configurations without coarse-graining).
But definitely don’t use non-PC lipid without making further changes/corrections/fudges to get the membranes to behave okay. These usually involve some repulsive potential for head group hydrogen atoms in the non-PC lipids.
The parameters can (technically) also be combined with GROMOS, AMBER or OPLS-AA protein force field families. Given the unusual mix of parameter sources for the Berger model, arguments can be made that all combinations are valid approaches. However the performance of such combinations hasn’t ever been thoroughly explored in detail beyond proteins seemingly behaving okay and some of the ways to combine lipid and protein parameters having been studied with some approaches resulting in overly strong interactions. Indeed in another world with twice the time such testing of lipid and protein force field parameter combinations is something I would like to have done (and indeed something actually started but never anywhere near finished).
Thank you all for your replies! I am using CHARMM-GUI membrane builder as Tom said. And following the README file with minor changes. When I worked with GROMACS, I highly benefited from the tutorial website. So sometimes I feel I just have to follow the tutorial as well. But now I am continuing with peptide adding step to my equilibrated membrane system. Thanks again!!
I also think CHARMM-GUI is currently what I found to be the most intuitive way to create membrane embedded systems (I do atomistic MD on membrane proteins). I even completely forgot how I would build such system from scratch.
The downside is that I cannot create such system in a high throughput manner in CHARMM-GUI (i.e. prepare lots of such systems quickly using the same protocol). I saw some potential in g_membed in GROMACS but I have never been able to make it work. It seems like an outdated functionality that got carried over to newer versions of GROMACS without updates.
I want to build a bilayer membrane system as well but I wish to use the gromos force field, which isn’t there in the CHARMM-GUI membrane builder forcefield options. I was trying out the KALP-15 in DPPC tutorial and came across the inflategro methodology but james here doesn’t suggest that for real simulations. Could you please help me with getting to know what options I have for building my system? Is there anyway that I can translate my parameters from a different ff like charmm to those of gromos?
I am using CHARMM36 for my peptide-membrane system so I know little about the GROMOS ff. If you do not have to use GROMOS ff, maybe you can consider using the FF’s in the CHARMM-GUI’s suggestion.