KaMLs for Predicting Protein p K a Values and Ionization States: Are Trees All You Need?
2024
Predicting Protein Ionization States with KaMLs
publication
Evidence: high
Author Information
Author(s): Shen Mingzhe, Kortzak Daniel, Ambrozak Simon, Bhatnagar Shubham, Buchanan Ian, Liu Ruibin, Shen Jana
Primary Institution: Cold Spring Harbor Laboratory
Hypothesis
Can machine learning models accurately predict protein pKa values and ionization states?
Conclusion
The KaML models significantly outperform existing methods in predicting pKa values and ionization states for various amino acids.
Supporting Evidence
- KaML models significantly outperform the current state of the art in predicting pKa values.
- The models achieved accurate predictions for deprotonated cysteines and lysines, which were blind spots in previous models.
- Innovations included separate treatment of acid and base and data augmentation using AlphaFold structures.
Takeaway
This study created a new tool to help scientists predict how proteins behave in different conditions, which is really important for understanding biology and developing drugs.
Methodology
The study developed machine learning models using decision trees and graph attention networks, leveraging a new experimental pKa database.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website