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

Deep Learning Models Capture Umbrella Sampling-Derived Energetic Trends: A Troponin C Case Study.

Akshaya Narayanasamy1, Rocco Moretti2,3, Jens Meiler2,3,4,5,6

  • 1Department of Chemistry and Biochemistry, University of California, Los Angeles, California 90095, United States.

Journal of Chemical Information and Modeling
|July 13, 2026
PubMed
Summary

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

Incorporating Surfaced-Induced Dissociation Mass Spectrometry Data into an AlphaFold-derived deep learning network improves protein structure prediction.

bioRxiv : the preprint server for biology·2026
Same author

Comprehensive Tutorial for Computational Methods of Protein Structure Prediction Incorporating Mass Spectrometry Data.

Mass spectrometry reviews·2026
Same author

Prediction of pre- and postfusion conformations of class I fusion proteins with AlphaFold2.

PloS one·2026
Same author

Structure-guided compound prioritization strategy for virtual screening identifies putative binders for the nuclear receptor LRH-1.

bioRxiv : the preprint server for biology·2026
Same author

Profiling the CFTR Variant Selectivity and Off-Target Interactions of VX-121.

bioRxiv : the preprint server for biology·2026
Same author

LCK-targeting molecular glues overcome resistance to inhibitor-based therapy in T-cell acute lymphoblastic leukemia.

Blood·2026

Deep learning models like ThermoMPNN can now rapidly predict protein energy changes, offering a faster alternative to traditional simulations for understanding mutations in Troponin C (TnC). This advance aids in designing proteins with specific calcium-binding properties.

Area of Science:

  • Computational Biology
  • Biophysics
  • Protein Engineering

Background:

  • Traditional methods like umbrella sampling for calculating protein free energies are computationally intensive.
  • Troponin C (TnC) regulates muscle contraction via calcium-induced conformational changes, crucial for understanding its function and disease-related mutations.
  • Disease-associated mutations and isoform differences in TnC alter the free energy of hydrophobic patch opening, impacting muscle function.

Purpose of the Study:

  • To evaluate deep learning models (ProteinMPNN and ThermoMPNN) as rapid alternatives for energetic readouts in protein dynamics.
  • To investigate the ability of ThermoMPNN to quantify free-energy differences in Troponin C (TnC) variants and compare with traditional methods.
  • To assess ProteinMPNN's capability in distinguishing isoform-specific conformational preferences.

Related Experiment Videos

Main Methods:

  • Utilized ThermoMPNN to calculate free-energy differences for disease-linked variants in cardiac TnC, comparing results with umbrella sampling simulations.
  • Analyzed free energy of hydrophobic patch opening in TnC using both deep learning and umbrella sampling methods.
  • Employed ProteinMPNN to analyze sequence probabilities for distinguishing conformational preferences between TnC isoforms.

Main Results:

  • ThermoMPNN showed a strong correlation (R²=0.82) with umbrella sampling results for hydrophobic patch opening free energy.
  • The model accurately predicted the effects of hypertrophic cardiomyopathy (HCM) and dilated cardiomyopathy (DCM) mutations on TnC's hydrophobic patch opening.
  • ProteinMPNN successfully differentiated isoform-specific conformational preferences, aligning with previous umbrella sampling findings.

Conclusions:

  • Deep learning frameworks, specifically ThermoMPNN and ProteinMPNN, offer a computationally efficient and accurate alternative to traditional free energy calculation methods.
  • These deep learning models can reliably predict the impact of mutations and isoform differences on TnC's conformational states and function.
  • Deep learning-based metrics hold significant potential for guiding the rational design of TnC mutations for therapeutic applications, such as modulating calcium sensitivity.