DiffPaSS: A New Method for Pairing Protein Sequences
Author Information
Author(s): Lupo Umberto, Sgarbossa Damiano, Milighetti Martina, Bitbol Anne-Florence
Primary Institution: École Polytechnique Fédérale de Lausanne (EPFL)
Hypothesis
Can DiffPaSS outperform existing algorithms for pairing interacting biological sequences?
Conclusion
DiffPaSS is a fast and effective method for pairing protein sequences that outperforms existing algorithms.
Supporting Evidence
- DiffPaSS outperforms existing algorithms for optimizing protein sequence pairing.
- It can be applied to both aligned and nonaligned sequences.
- DiffPaSS improves the prediction of protein complex structures using AlphaFold-Multimer.
Takeaway
DiffPaSS is like a smart helper that quickly finds which proteins work together by looking at their sequences, even if they aren't lined up perfectly.
Methodology
DiffPaSS uses a differentiable framework for optimizing the pairing of protein sequences based on similarity and coevolution scores.
Limitations
DiffPaSS currently assumes one-to-one pairings and may not perform well with collections of different sizes.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website