Hands-on session for protein structure prediction by AlphaFold3 and ESMFold
Course Instructors:
Dr. Sanjeet Kumar Mahtha, MK Bhan Postdoctoral Fellow
Sureshkumar V, Research Scholar
Dr. Debasisa Mohanty, Director
AlphaFold3
URL: https://alphafoldserver.com/
| Monomer | Example 1 | Example 2 |
| Dimer | Example 1 | Example 2 |
| Protein + Ligand | Protein Sequence | Ligand: FAD |
| Protein + RNA | Sequences | |
| Protein + DNA | Sequences | |
| Protein + DNA + Ion | Sequences | Ion: Sodium |
| Protein + PTMs | You can play with example 1 |
ESMFold
URL: https://esmatlas.com/resources?action=fold
| Monomer | Example 1 | Example 2 |
| Dimer– API and Web server do not support Dimer. But you may use ESMFold ColabFold Portal | https://colab.research.google.com/github/sokrypton/ColabFold/blob/main/ESMFold.ipynb | http://pdbi.nii.ac.in/AF_ESM/7awv_Dimer.fa [OR] Run on local server |
ESM – API
curl -X POST --data "GAMADIGSMDVLEYFERLKNRELAFVLDDLQLSDMVTRRGFSVIPFDDFDLAREDHPPAFVLVTRLDYHGKLMQAWETAKGISSHLSLAKFDTSPKSVEYSLDQLLSMDFAETLKRRGDYYDSVASTNRMEVVTPGAVLTCDFGNEIEIANNDVEMQKGWLYSVAEFFETSVINLEADRSSYTLNGDLCFTGLIYLCNRPDLKERASATMDELMRMSTRGRNVVSFVDNQIVRMELGGVDMTATLRELIVGKEREGSSTEFAMGCVEYPLAQDWTINSVMNEGSHGIHVGVGMGKEIPHMDFIAKGAELRIAESSDA" https://api.esmatlas.com/foldSequence/v1/pdb/ > 6zsv_Out.pdb
How to predict multiple structures using the script?
Demo fasta file – Multi_batch.fa
Download Script – ESM_script.py
Command: python ESM_script.py
# If biopython is not installed, you install it with “pip install biopython”.
import requests
from Bio import SeqIO
import os,sys
# Define the ESMFold API endpoint
ESMFOLD_API_URL = "https://api.esmfold.org/v1/predict"
input = "Multi_batch.fa"
for i, record in enumerate(SeqIO.parse(input, "fasta"), start=1):
sequence_id = record.id
sequence = str(record.seq)
print(f"Processing sequence {i}: ID = {sequence_id}")
cmd = "curl -X POST --data ""+sequence+"" https://api.esmatlas.com/foldSequence/v1/pdb/ > "+sequence_id.replace("|","_")+".pdb"
#print(cmd)
os.system(cmd)
How to run ESMFold on the local machine?
Option 1:
wget http://pdbi.nii.ac.in/AF_ESM/Script_ESM.sh
wget http://pdbi.nii.ac.in/AF_ESM/ESM_dim_ex.fa
sbatch Script_ESM.sh
Option 2:
module load esmfold/1.0.3
wget http://pdbi.nii.ac.in/AF_ESM/ESM_dim_ex.fa
## Command for Dimer Structure prediction
esm-fold -i ESM_dim_ex.fa -o output_folder -m /opt/apps/esmfiles/
How to run AlphaFold3 with ligand in the local machine?
Option 1:
wget http://pdbi.nii.ac.in/AF_ESM/Script_AF3_Lig.sh
sbatch Script_AF3_Lig.sh
Option 2
module load AlphaFold/AF3
mkdir AF3_Lig
cd AF3_Lig
cp /apps/run_alphafold.py .
wget http://pdbi.nii.ac.in/AF_ESM/AF3_input_test_wMSA.json
python run_alphafold.py --json_path AF3_input_test_wMSA.json --model_dir /opt/apps/af3_model --db_dir /opt/apps/AF3_DB --output_dir Output_AF3 --run_data_pipeline False
Additional Links:
PAE Viewer – https://pae-viewer.uni-goettingen.de/
PyMOL Download Link – https://pymol.org/#download
PyMOL License – https://pymol.org/ep/
PyMOL Tutorial – https://pymol.sourceforge.net/newman/userman.pdf
Chimera – https://www.cgl.ucsf.edu/chimera/download.html
Practice Files – Download (Needed for tomorrow’s session)
Further Reading – Comparison of Structure Prediction Methods and Comparison of AlphaFold2, AlphaFold3 and ESM Fold: Click here
More Details:-
| AlphaFold2 | Article – https://doi.org/10.1038/s41586-021-03819-2 GitHub Source Code – https://github.com/google-deepmind/alphafold |
| AlphaFold3 | Article – https://doi.org/10.1038/s41586-024-07487-w GitHub Source Code – https://github.com/google-deepmind/alphafold3 |
| ESMFold | Article – https://doi.org/10.1126/science.ade2574 GitHub Source Code – https://github.com/facebookresearch/esm |
| RoseTTAFold | Article – https://doi.org/10.1126/science.abj8754 GitHub Source Code – https://github.com/RosettaCommons/RoseTTAFold |
Download Programs for SSH/SCP connection (Only for Windows Users)
PuTTY – Download
WinSCP – Download