EF-G Conformation Folding and Ancestral Reconstruction

This repository contains the computational workflows and analysis code supporting the paper titled:
Evolutionary Insights into Elongation Factor G Using AlphaFold and Ancestral Analysis.
**Comput Biol Med 2025 Vol. 191 Pages 110188; Accession Number: 40222265 DOI: 10.1016/j.compbiomed.2025.110188

Overview

This project integrates structural prediction, phylogenetic analysis, and functional validation to explore EF-G conformational evolution. Specifically, it covers:

• Identification of EF-G conformations (con1 vs. con2) using AlphaFold predictions.
• Clustering of multiple sequence alignments (MSAs) based on AlphaFold-predicted structures.
• Phylogenetic reconstruction and ancestral sequence inference.
• Functional and biophysical characterization of two reconstructed ancestral EF-G proteins.

Contents

• myenv.yml
  → computing environment.

• EF-G_folding.ipynb
  → Main notebook for folding, structural and sequence-based clustering, and visualization.

• SEQ/wt.fasta
  → Wild-type EF-G query sequence.

• SEQ/
  → will contain initial ColabFold outputs and parent .a3m alignment files generated via MMseqs2.

• Cluster/
  → will Contain clustered .a3m files from AFCluster, with subfolders for predicted structures and analysis.

• Cluster/folding/
  → will contain PDB files generated from the clustered .a3m files.

• Cluster/folding/sel_pdb/
  → will contain PDB files that passed the pLDDT confidence threshold.

• Cluster/folding/sel_pdb/1st_1-286_RMSD/
  → will contain structure clustering and PSE-based selection data and further analysis.

• a3m_file → example a3m files of con1 and con2 for WT

• WT_list1/2.txt → example lists of a3m after structure clustering

• combine_All_MSA → folder with .fasta files to demonstrate cell #17.

• combine_ → example lists of a3m after structure clustering

• one .pse → example pse file to identify con1 and con2 for WT

• two .treefile and one .state file → example treefiles and ancestor sequence file.

• one node_JSD.csv → example to repproduce figure 5.

attributions

• ClusterMSA_min3.py
  → Modified sequence clustering script using a DBSCAN-based method.

• utils.py
  → Supporting utility functions

  these two attibutions are adapted from:
  H. Wayment-Steele et al., Nature 625, 832–839 (2024), doi:10.1038/s41586-023-06832-9. https://github.com/HWaymentSteele/AF_Cluster

Key Tools and Dependencies

• myenv.yml
• requirement.txt
• AlphaFold / ColabFold — for structure prediction
• IQ-TREE 2 — for maximum likelihood phylogenetic reconstruction
• PyMOL — for structural visualization and comparison
• Biopython, NumPy, Matplotlib — for parsing, analysis, and plotting

modularity

Most cells function as modular units, capable of carrying out specific tasks independently.

Reproducibility

To reproduce the computing environment:

conda env create -f myenv.yml
conda activate myenv  

To reproduce the full analysis pipeline:

1. Set your working directory (default_path).
2. Start from cell #2 in EF-G_folding.ipynb, adjusting the path to your local ColabFold installation.
3. Run cell #3 to generate the Cluster/ directory.
4. Follow through cells #2–8 to reproduce Figures 1–2.
   Cells #9–16 generate Figure 3, and #18–22 produce Figure 5.
5. Cell #23 (standalone) uses Logomaker for Figure 4.
6. Cell #17 produces Supplementary Figure S4A.

Citation

If you use this code or find it helpful, please cite:

Shawonur Rahaman,Jacob H. Steele,Yi Zeng,Shoujun Xu,Yuhong Wang (2025).
Evolutionary Insights into Elongation Factor G Using AlphaFold and Ancestral Analysis.
Comput Biol Med 2025 Vol. 191 Pages 110188; Accession Number: 40222265 DOI: 10.1016/j.compbiomed.2025.110188

Contact

For questions or collaboration inquiries, please contact:
Yuhong Wang
📧 ywang60@uh.edu