Based on the paper from JJ:

https://pubs.rsc.org/en/content/articlelanding/2020/ra/d0ra01127d#!divAbstract

Create a PPI module based on global + IF exploration

Dani

@danielSoler93 Hello Daniel, I would like to try the software...I am not sure how to get the executable file.. I tried to send email but in vain.

I'm using peleffy from the cli to generate templates and then passing them to the platform using the templates parameter. However for some reason peleffy is only generating the OPLS2005 folder inside DataLocal and not the OpenFF. When trying to copy the template raises a TemplateFileNotFound, but the template exists. Offending code is

It would be interesting to have a flag to set which PELE executable, documents and data folders do you want to use instead of having a hardcoded default.

Additionally, take into account that when using PELE > 1.6.1 you do not need to restrict the amount ligand rotamers because PELE will select a few of them randomly at each step.

Consists of: Box of 10-15A around interfece + induced fit simulation + refinement.

Project pipeline:

1. Build software
2. Build dataset
3. Test
4. Improve
5. Benchmark
6. Publish

Consists of: Box of ligand all around the protein + global exploration + induced fit simulation

Project pipeline:

1. Build software
2. Build dataset
3. Test
4. Improve
5. Benchmark
6. Publish

Link growai package and fragpele

The "equilibrationLength" is always set to 1 even if you specify a larger number of equilibration steps in the yaml file with the equilibration_steps flag.

Rdkit messes atoms when symmetry is found. Maybe we can first atom map and then correct the symmetry bug with that.

minimization step:

We saw at some simulation we find the pocket but the best struct is not well scored, but it is after a minimization step

0. Check if the step is actually useful. Run the minimization for several poses and see if they also go down in energy as much as the good one. (danielSoler93)

1. If yes, build the pipeline. (frumpowy)

   1.1 Cluster 1000 bestStructs from refinement into 125 clusters
   1.2 From the previously 125 obtained structures, DIFFERENT simulations for each cluster (each simulation has 2 cpus) --> 3h
   1.3 Analysis of all together, gather the 125 minimization simulations together and get best 200
   structures

Build RNAi package.

Project pipeline:

1. Get forcefield to work with RNA

2. Get Dataset

3. Test & Improve

4. Benchmark & paper

10135

hola2PELE-138

hola3

FragPele from sdf is still unstable due to rdkit connection. We should massively test that.

PELE-137

We detected in the PDB 3E64 3e64.pdb.txt two PTR residues defined as HETATM in chain A (1007-1008).

When pele_platform preprocesses and prepares this file for pele run, it introduce TER lines before and after this residues
3e64_processed.pdb.txt instead of renaming to ATOM (exist a template in PELE for this residue).

Alert user when passing a wrong key