manny405 / mcse

Molecular Crystal Simulation Library (mcse) is an open-source Python package for manipulating and analyzing molecular crystal structures
Apache License 2.0
14 stars 1 forks source link

some questions about cocrystal structure prediction.(from Noa's group) #5

Open Jiayihua2001 opened 9 months ago

Jiayihua2001 commented 9 months ago

Dear Dr. Imanuel Bier,

I hope this message finds you well. My name is Jiayi, a first-year PhD student in Prof. Noa Marom’s group. Recently, I have been working on the co-crystal structure prediction project. Noa wants us to try first to generate dimers, then input them in gnrs to generate xtals. I am greatly assisted by the codes you have already created! Many thanks for that!

I have run into a few issues in this field. Could I ask you a few questions regarding the dimer generation and selection process? I would be so grateful if you could assist me.

  1. How do you go about choosing the most promising dimers?
  2. Have you been able to use dimers to create experimental crystal structures?
  3. How can we decide which of all the generated dimers is the most promising?
  4. Do you believe it makes sense to choose the most promising dimers based on their total energy? Taking into account that the interaction between dimers in xtals might not be ignored.

Exp_extracted_dimer can generate experimental xtal structure using gnrs.

I have tested target XI, and there are 15000 dimers generated, dimers with a similar structure with exp_extracted_dimer (rmsd <0.5) are among the first 300 ranks in terms of energy (this could just be a coincidence, and I should test more cocrystal structure). However, using GNRS, even a similar structure is still unable to produce the experimental xtal structure.