Bayesian Active Learning for Optimization and Uncertainty Quantification with Applications in Protein Docking https://www.biorxiv.org/content/10.1101/537035v1.full

• C++ 4.8.5 or higher
• cNMA: Download and install the cNMA from "https://github.com/Shen-Lab/cNMA".
• Energy model: Please download the random forest energy model from https://drive.google.com/open?id=17ByuNoYy0t1R8EjuTK_cMyul5K004MHa
• CHARMM: Download the executable 'CHARMM36a1.exe', GBSW: 'radius_gbsw.str', CHARMM27 topology and parameter files and put them into 'dependencies/'.
• Linux Environment

In src/configuration.h, please change the macros as follow:

• src_dir: change to your current "src" path.
• cnma_path: change to your cNMA path.
• scoring_path: change to your 'random_forest.sav' path.
• output_path: change to the directory where you want to output.

In src/configuration.h, please change the macro 'protein_name' to your 4-letter/digit protein code. Also please change the macro 'protein_path" to the path where the unbound protein is.

• Go to 'BAL/src/'.
• Type './complie' to compile.
• Type './for_train' to run BAL.

• a.'end'(dir): containing the refined structures: receptor.pdb ligand.pdb

• b.'scores': The final energy (unit in Kcal/mol) of refined structures.

• c.'PMI_log': The log of area under the posterior.

• d.'cond_prob.dat': The conditional probability of refined structure: P(RMSD(x^,x*)<4 | x* \in Mi)

• e. 'Rmsd_dis': The RMSD(x^,x*) distribution of the posterior

• f. 'UQ': The [lb,ub] values

@article{Cao537035,
        author = {Cao, Yue and Shen, Yang},
        title = {Bayesian Active Learning for Optimization and Uncertainty Quantification in Protein Docking},
        elocation-id = {537035},
        year = {2019},
        doi = {10.1101/537035},
        publisher = {Cold Spring Harbor Laboratory},
        URL = {https://www.biorxiv.org/content/early/2019/01/31/537035},
        eprint = {https://www.biorxiv.org/content/early/2019/01/31/537035.full.pdf},
        journal = {bioRxiv}
	}

Yang Shen: [email protected]