Predicting Co-Complexed Protein Pairs from Heterogeneous Data
Author Information
Author(s): Qiu Jian, Noble William Stafford
Primary Institution: University of Washington
Hypothesis
Can a computational method effectively predict co-complexed protein pairs using heterogeneous data sources?
Conclusion
The study presents a computational method that accurately predicts co-complexed protein pairs in yeast, achieving high coverage and low false discovery rates.
Supporting Evidence
- The method achieved an ROC50 score of 0.937, indicating high predictive accuracy.
- Predicted positives were significantly enriched with known co-complexed protein pairs.
- The classifier demonstrated the ability to make predictions from partial and indirect evidence.
Takeaway
Scientists created a computer program that helps find which proteins work together in cells by looking at different types of data about them.
Methodology
The study used support vector machine classifiers trained on various data types, including protein sequences and interaction networks.
Potential Biases
Potential bias may arise from the reliance on existing databases and the selection of training data.
Limitations
The method may be affected by the quality and completeness of the input data, and the presence of missing data can impact predictions.
Participant Demographics
The study focused on yeast proteins, specifically Saccharomyces cerevisiae.
Statistical Information
P-Value
0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website