Unit CLASSICAL MOLECULAR DYNAMICS AND MACHINE LEARNING: SIMULATIONS OF BIOSYSTEMS
- Course
- Biology
- Study-unit Code
- A005340
- Curriculum
- Biomolecolare
- Teacher
- Maria Noelia Faginas Lago
- Teachers
-
- Maria Noelia Faginas Lago
- Paola Belanzoni
- Hours
- 47 ore - Maria Noelia Faginas Lago
- 10 ore - Paola Belanzoni
- CFU
- 6
- Course Regulation
- Coorte 2025
- Offered
- 2025/26
- Learning activities
- Affine/integrativa
- Area
- Attività formative affini o integrative
- Academic discipline
- CHIM/03
- Type of study-unit
- Opzionale (Optional)
- Type of learning activities
- Attività formativa monodisciplinare
- Language of instruction
- English
- Contents
- In this course, different computational chemistry methods commonly employed for the simulation of BIO processes will be explained during the theoretical lectures (35 hours), including classical molecular dynamics, Umbrella sampling and hybrid quantum mechanics/molecular mechanics (QM/MM) methods. Furthermore, during the computer exercises (12 hours) applications of different BIO processes with free software: AmberTools, Orca, VMD, DL_POLY. A second part is aimed at providing a basic understanding of the main machine learning (ML) algorithms and concepts, with a focus on the analysis of data from molecular simulations. Subsequently, we will explore how machine learning can enhance the analysis and prediction of data obtained from classical molecular dynamics simulations.
- Reference texts
- 1. M. P. Allen & D. J. Tildesley – Computer Simulation of Liquids¿Oxford University Press (Second Edition, 2017) 2. D. Frenkel & B. Smit – Understanding Molecular Simulation¿Academic Press (Second Edition, 2001) 3. Aurélien Géron – Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow¿O'Reilly Media (3rd Edition, 2022)
- Educational objectives
- The student will be able to combine MD and ML simulations to analyze complex biosystems in an advanced way, identifying hidden patterns, predicting molecular behaviors and potentially accelerating simulations and to apply basic ML techniques to molecular datasets and explore complex structures and hidden dynamics.
- Prerequisites
- * Basic knowledge of physics, computational chemistry and programming. * Fundamental knowledge of linear algebra and statistics
- Teaching methods
- Frontal lessons and computer lab.
- Other information
- For information on support services for students with disabilities and/or DSA visit the page https://www.unipg.it/disabilita-e-dsa
- Learning verification modality
- Oral exam with project presentation.
- Extended program
- Module 1: Introduction to Molecular Dynamics (MD) * Fundamentals of classical mechanics applied to molecular systems * Equations of motion and numerical integration methods * Force fields: description of atomic interactions * Software and tools for molecular dynamics (DL_POLY, AMBER) Module 2: Simulations of Biological Systems * Molecular models for proteins, DNA and membranes * Boundary conditions and thermostats/barostats * Advanced sampling techniques (Replica Exchange, Metadynamics) * Validation and analysis of simulation results Module 3: Introduction to Machine Learning for Molecular Dynamics * Basic notions of machine learning (ML) and deep learning (DL) * Supervised and unsupervised learning algorithms * ML frameworks for molecular simulations (TensorFlow, PyTorch, scikit-learn) * Datasets and feature engineering for molecular systems Module 4: Applications of Machine Learning to Molecular Dynamics * Prediction of molecular properties with ML * Acceleration of simulations with ML (potential of learned force, surrogate models) * Generation of molecular trajectories via neural networks * Dimensionality reduction and clustering of conformational states Module 5: Case Studies and Applications * Simulation of protein-ligand interactions * Study of conformational transitions (e.g. folding) * Dynamics of membranes and ion channels * Applications in drug discovery * Analysis of mutations and pathologies at the molecular level
- Obiettivi Agenda 2030 per lo sviluppo sostenibile
- 4