Abdul Raafik Arattu Thodika

Computational Chemistry   |   Biophysics   |   Machine Learning Potentials

---

I am a recent Ph.D. graduate from the University of Texas at Arlington interested in molecular simulations of protein systems. I develop and apply multiscale quantum mechanical/molecular mechanical (QM/MM) methodologies to investigate the thermodynamics and kinetics of chemical catalysis within enzyme environments. Alongside traditional sampling approaches, I’m exploring the integration of machine learning techniques to improve the efficiency of analyzing such complex chemical systems in terms of simulation speed and accuracy.

The developed approaches and workflows are primarly implemented in the CHARMM, CHARMM-GUI, mlp_qmmm and BLaDE programs.

I did my Ph.D. research under Kwangho Nam focussing on QM/MM simulations of enzymatic reactions and machine learning potentials for free energy simulations, studying catalytic mechanisms of dihydrofolate reductase and adenylate kinase enzymes. Before that, I completed my BS and MS in Chemistry at the Indian Institute of Science Education and Research (IISER) Pune, India where my studies primarly focussed on physical chemistry topics.

Education

• Ph.D. in Physical Chemistry, University of Texas at Arlington, 2021 - 2026, Advisor: Kwangho Nam
• BS & MS in Chemistry, Indian Institute of Science Education and Research(IISER Pune, India), 2015 - 2020

Relevant Skills

• Programming Languages: Fortran, C++, CUDA practices, Python, Bash
• Software: CHARMM, Amber, Gaussian, QChem, VMD, Schrödinger PyMOL, CHARMM GUI
• Machine Learning: PyTorch, Scikit-Learn, NumPy, Alphafold2

Ongoing Projects

• Machine Learning Potentials (MLP) for Free Energy Simulations of Enzymatic Reactions
  • Collaboration: Dr. Xiaoliang Pan, [Dr. Yihan Shao], [Dr. Saroj Kumar Panda]
  • Description: In the ongoing work, we undertake a comprehensive re-examination of the ML architecture introduced by Pan et al (for QM/MM free energy simulations) to rigorously evaluate the transferability of a pre-trained MLP and Δ-learning based MLP models.
  • Relevant Publications:
    • Machine learning quantum mechanical/molecular mechanical potentials: evaluating transferability in dihydrofolate reductase-catalyzed reactions[J. Chem. Theory Comput., 2025]
    • Accurate and Time-Efficient Condensed-Phase Free Energy Simulations with Reaction Specific delta-Machine Learning Potentials in CHARMM [in review, 2026]
      
• Dynamic Perspectives on Catalytic Function in Adenylate Kinase
  • Collaboration: Dr. Magnus Wolf-Watz
  • Description: In our project, we aim to understand the catalytic mechanism of adenylate kinase using a combination of classical molecular dynamics and QM/MM simulations. Our particular interest lies in examining the role of protein dynamics in the enzyme’s catalytic function. Additionally, we are exploring the evolutionary trajectory of adenylate kinase’s thermoadaptation across different species with the aid of Alphafold2 predicted structures.
  • Relevant Publications:
    • Elucidating Dynamics of Adenylate Kinase from Enzyme Opening to Ligand Release. [J. Chem. Inf. Model., 2024]
    • Magnesium induced structural reorganization in the active site of adenylate kinase. [Sci. Adv., 2024]
      
• CHARMM-GUI QM/MM Interfacer Module
  • Collaboration:: Dr. Donghyuk Suh
  • Description: QM/MM Interfacer Module (part of CHARMM-GUI) provides a GUI-based platform to set up QM/MM simulations in CHARMM and Amber. In this project, I assist in debugging, generating template scripts, and testing the module for various QM/MM MD simulations applied to enzyme systems.
  • Relevant Publications:
    • CHARMM-GUI QM/MM Interfacer for a Quantum Mechanical and Molecular Mechanical (QM/MM) Simulation Setup: 1. Semiempirical Methods [J. Chem. Theory Comput., 2024]

Presentations

• 2026 ACS Spring Meeting (Atlanta, Georgia, US): “Machine Learning (ML)-assisted free energy simulations of enzyme reactions: Efficiency and Transferability of QM/MM-based ML potentials”
• 2025 ACS Fall Meeting (Washington, D. C., US): “Thermoadaptive mutations tune domain dynamics and catalytic effieciency in adenylate kinase”
• 2025 10th Institute of Computational Molecular Science Education (i-COMSE) Workshop (OSU Stillwater, Oklahoma, US): “QM/MM Molecular Dynamics Simulations (teaching assistant demonstration sessions)”
• 2024 ACS Meeting in Miniature (MiM) (UDallas, Texas, US): “Machine Learning Potential Models for Free Energy Simulations of Enzyme Reactions”
• 2024 Computational Chemistry Gordon Research Conference (GRC) (UMaine Portland, Maine, US): “Exploring Transferability of Machine Learning Potentials for Free Energy Simulations of Enzymatic Reaction: A Case Study on Dihydrofolate Reductase Catalyzed Reaction”
• 2024 Chemistry at Harvard Molecular Mechanics (CHARMM) Developers' Meeting (UM Baltimore, Maryland, US): “Machine Learning Potential (MLP)-assisted QM/MM Molecular Dynamics Simulations in CHARMM Program”
• 2023 Southwest Regional Meeting (SWRM) of The American Chemical Society (Oklahoma City, Oklahoma, US): “Dynamic Perspective on Catalytic Function in Adenylate Kinase: Decoding The Role of Magnesium Ion with Molecular Dynamics Simulations”
• 2022 Discover Symposium (UT Arlington, Texas, US): “Insights on Evolutionary Thermoadaptation of Adenylate Kinase: Ancestral Protein Structure Prediction with Deep Learning”

---

Last updated: May 2026