A deep learning method for predicting interactions for intrinsically disordered regions of proteins
2024

Deep Learning for Predicting Protein Interactions

publication Evidence: high

Author Information

Author(s): Majila Kartik, Ullanat Varun, Viswanath Shruthi

Primary Institution: National Center for Biological Sciences, Tata Institute of Fundamental Research, Bangalore, India

Hypothesis

Can a deep learning method accurately predict interactions for intrinsically disordered regions of proteins?

Conclusion

The Disobind method outperforms existing models in predicting protein interactions involving intrinsically disordered regions.

Supporting Evidence

  • Disobind performs better than AlphaFold-multimer and AlphaFold3.
  • Combining Disobind and AlphaFold-multimer predictions further improves performance.
  • Disobind can be used to localize IDRs in integrative structures of large assemblies.

Takeaway

Scientists created a computer program that helps predict how certain proteins interact with each other, especially when they are not structured like most proteins.

Methodology

The study developed a deep learning model called Disobind that predicts inter-protein contact maps and interface residues using sequence embeddings from a protein language model.

Potential Biases

The model's predictions may be biased due to the limited number of experimental structures available for IDRs in complexes.

Limitations

Disobind is limited to binary IDP-partner complexes and requires input sequence fragments to be less than one hundred residues long.

Digital Object Identifier (DOI)

10.1101/2024.12.19.629373

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication