Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Molecular Dynamics Fingerprints (MDFP): Machine Learning from MD Data To Predict Free-Energy Differences.

Sereina Riniker1

  • 1Laboratory of Physical Chemistry, ETH Zürich , Vladimir-Prelog-Weg 2, 8093 Zürich, Switzerland.

Journal of Chemical Information and Modeling
|April 4, 2017
PubMed
Summary

We introduce a new molecular fingerprint (MDFP+) derived from molecular dynamics (MD) simulations. This method efficiently predicts solvation free energies and partition coefficients, matching the accuracy of complex computational techniques.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Smoother Alchemical Transformations via Enveloping Distribution Sampling for Free-Energy Estimation.

Journal of chemical theory and computation·2026
Same author

Multiscale Neural Network Potential with Anisotropic Message Passing for the Fast and Accurate Simulation of Protein Dynamics and Enzymatic Reactions.

Journal of the American Chemical Society·2026
Same author

Balancing Data Quantity and Quality: Evaluating Curation Strategies for Bioactivity Prediction in Lead Optimization.

Journal of chemical information and modeling·2026
Same author

How well do classical and multiscale QM/MM molecular dynamics simulations capture stereoelectronic effects? A comparative study on atropisomerism.

The Journal of chemical physics·2026
Same author

Structures of ALG3/9/12 reveal the assembly logic of the N-glycan oligomannose core.

Nature chemical biology·2026
Same author

Unraveling Torsional Preferences: Comparative Analysis of Torsion Motif Torsional-Angle Distributions across Different Environments.

Journal of chemical information and modeling·2025

Area of Science:

  • Computational chemistry
  • Cheminformatics
  • Machine learning

Background:

  • Machine learning (ML) is widely used in cheminformatics for property prediction.
  • Training ML models with molecular dynamics (MD) simulation data is an underexplored area.

Purpose of the Study:

  • To develop a novel fingerprint (MDFP+) using MD simulation data for ML model training.
  • To predict solvation free energies and partition coefficients accurately and efficiently.

Main Methods:

  • Constructed MDFP fingerprint from statistical moments (mean, median, spread) of MD-derived properties.
  • Combined MDFP with 2D structural counts to create MDFP+.
  • Trained ML models using MDFP+ for solvation free energy and partition coefficient prediction.

Related Experiment Videos

Main Results:

  • MDFP+ efficiently encodes physicochemical and entropic information from short MD simulations.
  • ML models trained with MDFP+ achieved performance comparable to established methods like FEP, LIE, COSMO-RS, and SMx.
  • Accurate predictions for solvation free energies across five solvents and partition coefficients for three solvent pairs were obtained.

Conclusions:

  • The MDFP+ approach offers a computationally inexpensive and easy-to-implement method for predicting solvation properties.
  • This technique demonstrates the potential of leveraging MD simulation data for enhanced ML-based cheminformatics predictions.
  • MDFP+ provides a competitive alternative to more computationally demanding free-energy calculation methods.