Neural Network for SH3 Domain-Peptide Interaction
Author Information
Author(s): Ferraro Enrico, Ausiello Gabriele, Helmer-Citterich Manuela
Primary Institution: Centre for Molecular Bioinformatics, Department of Biology, University of Tor Vergata, Rome, Italy
Hypothesis
Can a neural network model improve the detection of SH3 domain interactors compared to traditional methods?
Conclusion
A neural network can more effectively identify SH3 domain specificity than standard methods.
Supporting Evidence
- Neural networks showed higher sensitivity and specificity than PSSMs in detecting SH3 domain binders.
- The study demonstrated that adequate training data is crucial for the performance of neural networks.
- Neural networks can identify both general and domain-specific SH3 domain interactors.
Takeaway
Scientists created a computer program that helps find which proteins stick together better, using a special kind of math called a neural network.
Methodology
The study used neural networks trained on peptide sequences from the S. cerevisiae proteome to predict SH3 domain binding specificity.
Potential Biases
Sequence similarity between binders and non-binders may lead to misclassification.
Limitations
The performance of the neural networks can vary based on the number of binding peptides in the training set.
Participant Demographics
The study focused on peptide sequences from baker's yeast (S. cerevisiae).
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website