MATLAB and Simulink for Chemistry

Analyze, visualize, and build predictive models for chemical data

With MATLAB and Simulink, you can acquire and process experimental data, including AFM, Cryo-EM, NMR and EPR. MATLAB lets you generate and visualize simulated big data and create predictive models for molecular structures and properties using machine learning and deep learning.

MATLAB and Simulink help you to:

  • Simulate and fit various spectroscopy data by applying numerical techniques and visualization methods
  • Develop advanced predictive models for molecular properties predictions
  • Develop new theoretical frameworks to model complex chemical systems and provide analytical and numerical solutions
  • Teach chemistry-oriented programming skills across all levels of chemistry courses

Please join MathWorks at these upcoming webinar series focusing on how to use Quantum Computing and Deep Learning in chemistry.

See How Others Use MATLAB for Chemistry Research and Teaching

Psi4-MATLAB Molecular Dynamics Simulation Workflow

You can use Psi4 (an open-source suite of ab initio quantum chemistry program) with MATLAB to build an automated molecular dynamics (MD) simulations workflow for data generation and processing. This Psi4 example starts with a single molecular structure input, rotates it around a C-C bond, and calculates the molecular energy at the desired theory level. The output of the Psi4 computations is then processed in MATLAB to extract data and to build a single .mat file for further analysis.

The folding of a neuropeptide with seven amino acids, APRLRFY, is examined on gate-based quantum processor within coarse-grained model.

Ground-State Protein Folding Using Variational Quantum Eigensolver

With MATLAB, you can use qubits to encode a protein fold on a 3-D tetrahedral lattice. Using this ground-state protein example, the ground-state is found through a simulated variational quantum eigensolver routine. The final circuit from the simulation is run on a real quantum processor unit for comparison.

Functional Groups Classification Using Graph Attention Networks

MATLAB enables you to classify molecules that have multiple functional groups using graph attention networks (GAT). In this multilabel graph classification example, the training is done using the QM7-X data set, a collection of graphs that represent 6950 molecules. This demonstration considers the functional groups CH, CH2, CH3, N, NH, NH2, NOH, and OH.

Schematic illustration of multiple functional groups classification workflow using GAT.
Schematic illustration of atomic classification workflow using GCN.

Atoms in Molecules Classification Using Graph Convolutional Network

You can use MATLAB to predict the types of atoms in a molecule using a graph convolutional network (GCN). Using this node classification example, you can learn how to train a GCN with the QM7 dataset, which is a molecular data set consisting of 7165 molecules composed of up to 23 atoms.


Learn about the products used for chemistry applications.