pvalle6 / smlde

This is the repository for the Engineered Protein Directed Evolution Project.

Structure-informed MLDE for GB1

2. Scripts and Dependencies for Sequence Input Feature Generation
   • Contains working scripts to run zeroShot on HPC
   • Contains unworking/untested scripts for glycanDock use on HPC
   • Contains original working scripts for FlexPepDockUse
3. Scripts and Dependencies to Calculate RMSD on HPC / 5. Scripts and Dependencies to Run FPD on HPC
   • Contains scripts for analyzing Wittmann Data and pulling encodings appropriately for Training and Testing Models

smlde's Issues

Pre-Review Prep

Make new python package and put code into with detailed instructions
Make juypter notebook with detailed instructions

Need a script to pull specific index given a 4 letter ID

Please modify the script you use to pull MSA Embeddings using the mutant's 4 letter ID so that it can pull a number from a single column when given a mutant's 4 letter ID. Cool thanks. I can send you the file if you need it to test.

Clean up and Document for paper

Need to clean up repository for reference and possible fork when publishing.

Remove Local Paths from Scripts

RMSD for 1000 Mutants

So for this we need a script that can pull a specific file from each folder in a directory, load that file into PyMOL, run a series of PyMOL commands, and then record the outputs from those commands into a simple 2D matrix. The FlexPepDock scripts folder has a script called PreProcess.sub that essentially performs two tasks. The first task the script performs is pulling that specific file from each folder in a directory. The second task the script performs is running a series of PyMOL commands in the terminal without opening the PyMOL GUI. Not using the GUI significantly reduces the computational demand and run time.

SO, what I think would be easiest is to adapt the PreProcess.sub script to simply run a different series of PyMOL commands and then output the result. I think the majority of the code would be the same. Kenny sent us a PowerPoint from CCT last week that explains some of the code, I've attached it here. Dr. Li wrote the script so it looks like gibberish to lower experience coders. The PreProcess.sub script is on the Github if you can't find it. Try to take a look at this and get back to me before the meeting tomorrow so we can say we have something.

Lane

Please add auto contrast to Tiff to PNG image script.

Image/Adjust/Contrast:Brightness then under that Auto for each image under a directory or sub-directory.