Analyzing Protein-Protein Interaction Data Using Directed Graphs
Author Information
Author(s): Chiang Tony, Scholtens Denise, Sarkar Deepayan, Gentleman Robert, Huber Wolfgang
Primary Institution: EMBL, European Bioinformatics Institute
Hypothesis
Can directed graph and multinomial error models effectively assess and characterize error statistics in protein-protein interaction datasets?
Conclusion
The study demonstrates that protein interaction datasets can be characterized by their coverage, biases, and stochastic variability, which can inform future experimental designs.
Supporting Evidence
- Directed graph models help assess the quality of protein interaction data.
- Multinomial error models characterize the error rates affecting protein interaction datasets.
- Systematic biases can be identified through statistical testing of interaction data.
Takeaway
This study looks at how scientists can better understand protein interactions by using graphs to analyze data, helping them figure out which proteins work together and which ones don't.
Methodology
Directed graph and multinomial error models were used to analyze large-scale datasets of protein interactions in yeast.
Potential Biases
Potential biases arise from the experimental methods used, particularly in distinguishing between tested and untested interactions.
Limitations
The study relies on the assumption that untested interactions are not reported, which may introduce bias.
Participant Demographics
The study focuses on datasets from Saccharomyces cerevisiae, a species of yeast.
Statistical Information
P-Value
0.01
Statistical Significance
p<0.01
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website