STAR: predicting recombination sites from amino acid sequence
2006
STAR: A Tool for Predicting Recombination Sites in Proteins
Sample size: 945
publication
Evidence: moderate
Author Information
Author(s): Denis C. Bauer, Mikael Bodén, Ricarda Thier, Elizabeth M. Gillam
Primary Institution: The University of Queensland
Hypothesis
Can machine learning be used to predict useful recombination sites from amino acid sequences?
Conclusion
STAR allows users to identify useful recombination sites in amino acid sequences without needing structural information.
Supporting Evidence
- STAR predicts the maximum number of connections that can be broken by recombination for each position in the parent sequence.
- The correlation coefficient between STAR's predictions and the SCHEMA algorithm is very high at 0.89.
- STAR can be used for proteins with unknown structures, expanding the possibilities for protein engineering.
Takeaway
STAR is like a smart helper that tells scientists where they can mix and match parts of proteins to make new ones without breaking them.
Methodology
Machine learning methods were used to predict recombination sites based on amino acid sequences.
Limitations
STAR's predictions are based on a dataset of proteins with known structures, which may not cover all possible sequences.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website