TIES_PM_RNAP_RIF

This repository contains all input files, output data, and processed results used in the study:

"Rapid, accurate and reproducible de novo prediction of resistance to antituberculars"

The TIES-PM simulations were primarily executed on the Polaris High-Performance Computing (HPC) platform, with additional analysis performed on local machines. The data provided here allows for inspection, reproduction, and further exploration of our molecular simulation-based resistance prediction approach.

---

Repository Structure

The repository is organized into three main directories:

1. Input Files

Contains all files required to set up and run the TIES-PM workflow. It is divided into three parts:

• Preparation: Example inputs for wild-type and hybrid residues, along with an automated pipeline (modeling) to generate simulation-ready structures.
• Calculation: A sample directory structure and configuration files (replica-confs) for both com (complex) and prot (apo protein), along with a job submission script for Polaris.
• Post-processing: Scripts for analyzing simulation outputs and computing free energy changes, including an automated analysis pipeline (auto_post0.sh) and support scripts.

2. Output Files

Provides a complete example of raw simulation outputs for the mutation V170F, demonstrating the directory structure used in TIES-PM:

• Each of com/ and prot/ contains 13 lambda windows.
• Each lambda window includes 5 replicas.
• The sim1.alch files from each replica are the key inputs for ΔΔG calculation.

3. Results

Contains all processed free energy results from the TIES-PM workflow. The main results file includes:

• ΔG and ΔΔG values with associated statistical errors
• Mutation classification (Resistant/Susceptible)
• Comparison to experimental annotations (e.g., WHO phenotype, CRyPTIC data)
• Metadata such as sample counts and resistance proportions

Each column in the results file is explained in detail in the corresponding README.md file within the 3_Results/ directory.

---

Usage

Users interested in replicating or extending this work should:

• Review the input structure in 1_Input_files/ and adapt templates for their own systems;
• Use the provided output example in 2_Output_files/ to understand the lambda/replica setup;
• Refer to 3_Results/ for analysis outputs and validation against reference datasets.

The workflow is compatible with other HPC platforms with minor adjustments. Detailed automation and analysis scripts are included or available upon request.

---

Citation

If you use the data or structure of this repository, please cite our paper:

Click here for preprint paper

---

For further questions or collaborations, please contact:
[email protected]