moffragmentor
moffragmentor copied to clipboard
kjappelbaum
write fragments into pormake input format
would also be nice to then add some functionality to pormake to update the database
actually, for placing the dummy atoms it would be useful to extend the "dummy molecules" in the molecules in PR #152 even more (because we would add the atom the carboxy-C binds to as dummy and to avoid geometry analysis we could just take the right distance and orientation from the structure)
Hi, Thank you for making all the code on MOFs available!! I am wondering if this issue describes the only open step to connect to PORMAKE. I am looking for a tool that can do this and, if it's "just" rewriting your building blocks, it seems easiest to still use moffragmentor (i.e., even with this issue being open). Or do you maybe suggest another solution? I am asking because I do not have any experience with the other tools either... Thank you so much already!
Hi @vthost, thanks for your interest!
This is not yet implemented in the general case (I have some development code but got sidetracked with another project). In principle, moffragmentor tracks in private attributes the places one would need to place the dummy atoms.
One difficulty is that the dummy atom definition is different in the MOF-assembly code from what I use here internally.
I know that some people have their own internal tools that they do not want to share, but they are often limited to carboxy-linkers.
If you need this urgently, I'd be happy to help you implement this with moffragmentor. Unfortunately, I can't give you a guarantee when I'll come back to it myself (I guess in the next month, but not sure).
Thank you for the explanation! I think I can use it for my purpose, for now, and will look into the PORMAKE integration later.
I experimented a bit, but I think it would be hard to use it within PORMAKE right now. To me, it seems that the fragmentation is done differently. In the appendix of their paper (see eg Fig S5), they show that they split at single bonds. But then, we'd get different building blocks anyway. Or am I missing anything?
If not, I am wondering if it would be feasible (for me) to create a moffragmentor version which fragments similarly or if there are too many pitfalls (or if it would be better to start somewhere else, not from moffragmentor).
Yes, the fragmentation is done differently, and you would need to use some of the private attributes saved in the building blocks to reconstruct the building blocks in the poremake format.
One could use this to write the building blocks in the pormake format (that was the reason this information is stored with the building blocks). However, this requires coding and testing for edge cases (e.g., non-carboxy building blocks).
I got distracted by other projects and do not see myself working on this until April. I can provide some pointers on where to look, but I won't have the time to do much coding and debugging myself.
If this doesn't work for you, you might want to look into other MOF fragmentation tools, such as mBUD.