Repository content
prep: Folder containg files for preparing PDB of target structure to use in MD simulations
Replica_X: Contains "cryst", "equi", "prod" folders to run an independent replica for target strcuture
Replica_X/cryst: Contains input coordinates and prmtop file of target
Replica_X/equi: Minimization and equilibration of target structure
Replica_X/prod: Production run of MD simulation (1 µs)
Structures: Contains initial PDBs of Pin20, Pin21, and Pin23 as well as the cpptraj files for data extraction as well as PDBs of the clusters
All other files are used below to simplify running the MD simulation or used in the MD simulation if not mentioned (e.g. md_nvt_red_01.in) 


Guideline to perform MD simulations of self cleaving DNA

Step 0: Changing structure for MD simulations (optional)
-Copy Template folder and rename it to which structure you want (optional but helpful for multiple changes)
-Open "Replica_replace.sh" and use "sed/XXX/'Target_Structure'/g" to replace it with e.g. Pin20 or Pin23
-After saving, run "Replica_replace.sh" to replace the template of Pin21 with your target
-Finally, copy the PDB file of you target structure to the "prep" folder
Note 1: The temperature can be adjusted similiar to above by uncommenting the respective lines
Note 2: This can be done with any DNA/RNA structure, adaption to proteins requires different forcefields
 
Step 1: Preperation of MD simulations
-Go to prep Folder and execute "Prepare_DNAzyme.sh" via bash
-This should generate a Pin21_solv.pdb and the corresponding inpcrd and prmtop files for the MD simulations and copy them to all replicas
-Note 1: The concetration for Cd2+ is 2 mM and can be adapted by changing "--solute Cd_hexa.pdb --solute_con 0.002" (2 mM) to "--solute Cd_hexa.pdb --solute_con 0.02" (20 mM)
-Note 2: Path to the Amber software package may need to be adapted on different systems 

Step 2: Run of MD simulations
-After running Prepare_DNAzyme.sh the entire MD simulation folder can be upload to a high performance cluster or immediately submited
-For starting the MD simulation go back to main folder (which contains the replicas and prep folders) and execute "Submit.sh"
-Executing "Submit.sh" will submit "cpu-par.srv" to the cluster, performing initial minimization
-The simulation will submit any following scripts until 1 µs of simulation for each replica has been reached
-Note 1: All .srv scripts in "Replica_X/equi" and "Replica_X/prod" are using the Slurm configuration of the Gohlke research group, adaption may be necessary, most likely in regards to the partition
-Note 2: Slurm was used as workload manager but with some adjustments to the header, each script can be submitted via PBS as well 
-Note 3: A total of 40 trajectories (each 25 ns) should be generated in "Replica_X/prod", resulting in a total of 1 µs per replica
-Note 4: If any of the scripts failed, you can restart the simultion by resubmitting a previous scripts (e.g gpu_ser3.srv when an error with gpu_ser4.srv occured)

Step 3: Data extraction of MD simulations via cpptraj
-Submit "Full_Analysis.sh" or run it locally, this will take approximately 4 hours for 10 µs of MD simulations
-Note 1: In case the script did not copy the resulting data from the "Replica_X" folders back into the main folder, you may have to copy them by hand
-Note 2: You can decrease the amount of cores, which may then take longer than 24 hours, keep that in mind for queue time 
-Note 3: "cpptraj_all.in" may need to be adapted to work an Pin20 and Pin23, adapted files have been provided in the "Structures Folder"

Step 4: Clustering and 2D plot (optional)
-Submit "Clustering.srv" after running your MD simulation

Step 5: Data analysis with Jupyter notebooks
-Open MD_analysis.ipynb via Jupyter with at least Python3.9
-Run the notebook, ideally restarting the kernel as well to ensure any variables are reset
-Note 1: Additional modules may need to be installed but these are listed in the respective notebook fields

Step 6: Running aMD (optional)
-After successfully running the low temperature simulation copy the entire MD simulation as one folder into the "aMD" Folder
-Execute "prep_MD_sims_aMD.sh", which should generate all required files
-Upload to HPC cluster similiar to Step 2
-Note 1: "prep_MD_sims_aMD.sh" can also be used to generate input files for the unbiased MD simulation by uncommenting the section