Skip to content

josevilar-qbioai/energorna-public

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

EnergoRNA — VUS Classification via mRNA Thermodynamic Fingerprinting

EnergoRNA

Complementary classifier for Variants of Uncertain Significance (VUS)
via mRNA thermodynamic fingerprinting

Language: English · Español


This project has evolved into EnergyFingerprint, which includes the full open-source implementation, reproducible notebooks, and validated results across 8 genes and 4 protein families.

EnergyFingerprint

The method originally described under the EnergoRNA name has been substantially expanded and published as EnergyFingerprint: a lightweight 1D-CNN (~228K parameters) that classifies missense variants as pathogenic or benign by combining mRNA thermodynamic stacking profiles with ESM-1v protein language model scores.

DOI DOI

Key results (8 genes validated):

Gene Disease Intra-gene AUC Zero-shot AUC
BRCA1 Breast/ovarian cancer 0.943 --- (train)
TP53 Multi-cancer 0.907 0.854
PTEN Multi-cancer 0.994 0.975
PALB2 Breast cancer 0.950 0.954
CFTR Cystic fibrosis 0.979 0.777
HBB Sickle cell disease 0.721 0.707
SCN1A Epilepsy/Dravet 0.868 0.637
MECP2 Rett syndrome 0.911 0.430

For the full code, notebooks, and documentation, visit: github.com/josevilar-qbioai/energyfingerprint

Patent

This technology is the subject of a Spanish patent application:

  • Title: Método y sistema para el análisis termodinámico de transcritos de mRNA mediante perfiles de energía de apilamiento a resolución de nucleótido y redes neuronales convolucionales unidimensionales
  • Application number: P202630522
  • Filing date: April 11, 2026
  • Applicant: Jose Antonio Vilar Sanchez
  • Office: Oficina Española de Patentes y Marcas (OEPM)
  • Status: Pending examination

Citation

If you reference this work in academic publications, please cite:

@article{vilar2026energyfingerprint,
  title={EnergyFingerprint: mRNA Thermodynamic Profiling Combined with Protein
         Language Models Enables Zero-Shot Cross-Gene Missense Variant Classification},
  author={Vilar Sanchez, Jose Antonio},
  year={2026},
  doi={10.5281/zenodo.19831154}
}

Contact


EnergoRNA — Decoding variants through the energy of RNA

Releases

No releases published

Packages

 
 
 

Contributors