Hi All,
I am trying to build a membrane system using CHARMM-GUI. I’m following the guideline from a youtube video. But it seems that it’s not working. Can anyone please suggest other membrane system building for gromacs simulation. There is no specific error message as well.
Thank you.
Hello,
what do you mean with ‘it is not working’? CHARMM-GUI is by far one of the easiest and best solutions out there to build automatically bilayers, I would be really 100% sure that it is not working for you before moving on with other methods. You can use Packmol for example, but again, CHARMM-GUI is faster and much much easier to use.
Hi thank you for your response. It stops at step 4. The step I choose NaCl concentration 0.15. This step fails. There is no specific error message as well. I will try to see some other guide to make a bilayer system using CHARMM-GUI. I only watched the youtube video. Thank you.
Hi is there a possibility that I please send you my pdb structure and the lipid membrane components and ion concentration to help understand what is wrong?
What pdb file are you referring to? If you want to just prepare a bilayer then you should probably use just built in CHARMM-GUI structures.
Yes, you can maybe post specifically what you are trying to achieve? Like lipid composition etc. so to see if we can get to the root of the problem!
Thank you for your response. Please let me know what you mean by built in CHARMM-GUI structures? I have generated a structure using AlphaFold. I converted the pdb file to log file to be able to upload it here. For this structure, Align the First Principal Axis Along Z (looks fine), I added 190 cholesterol,
40 POPC, 88 POPE, 1 POPS and 56 NSM.
For Length of X and Y, I tried several times with different values between 150-300. Then I change ions to NaCl conc 0.15 and calculate the number of ions. After this step it fails.
41_P.log (783.3 KB)
The structure looks a bit weird, especially one of the chains is showing a lot of clashes. Also, you have quite high B factors in many sections, are you sure of the AlphaFold result? Does it resemble more or less what you are expecting experimentally or there is no data?
Regarding CHARMM-GUI, with built-in structure I meant that you have to use the lipids available within the web server. The pdb file has a very large and complex structure, I do not know if CHARMM-GUI is able to handle easily this kind of structure without further guidance. Do you know more of less at what height the bilayer is in the structure, i.e., which part of the protein is the membrane domain? The quantity of lipids sound strange, why just 1 POPS? The molar ratio is so low that it doesn’t really matter if it is part of the membrane or not.
When you build the membrane, what type of insertion do you do? I think in this case would be better to give CHARMM-GUI a rough idea of the box side lengths and then specify just the ratio of the lipids, rather than its number, e.g., 200 Angstrom of side length for an hexagonal box and then 190CHOL/40POPC/88POPE/56NSM as a ratio, not specifically as the number of components. This may help the webserver in building the membrane. Also what kind of error are you getting from CHARMM-GUI?
Hi
Thank you for your response.
There is no experimental pdb structure but it is close to the structure it should be. When I used Align the First Principal Axis Along Z, it did seem that the membrane was added to the right place.
For the lipid bilayer, I used this paper. I’m not too sure how to read it but this I saw the ration in another paper. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1413831/pdf/pnas-0511136103.pdf
I was actually trying ratio itself instead of number. It seems the real problem is my pdb file. I checked the option 'Check/Correct PDB Format ’ while uploading my PDB file and I can the error that the H coordinates are not there. So I used gmx2pdb to get the H coordinates but then it seems it cannot read the coordinates of all the atoms. Could you please help here, with the warning that there is no coordinates?
The error message is too long to be added here. But when I read it there is no reason for the failed job.