Identifying Protein Complexes Using Markov Random Fields
Author Information
Author(s): Rungsarityotin Wasinee, Krause Roland, Schödl Arno, Schliep Alexander
Primary Institution: Max Planck Institute for Molecular Genetics
Hypothesis
Can we identify protein complexes directly from high-throughput TAP data using a model-based approach?
Conclusion
The study demonstrates that protein complexes can be identified from high-throughput experimental data without prior elimination of proteins or weak interactions.
Supporting Evidence
- The model incorporates false negative and false positive errors, showing robustness to noise.
- The approach does not rely on heuristics, making it suitable for organisms without established reference frames.
- Results compare favorably with prior work on reference data sets, especially for larger unfiltered data sets.
Takeaway
This research shows a new way to find groups of proteins that work together, using a smart model that can handle messy data.
Methodology
The study uses a model based on Markov random fields to analyze TAP data and identify protein complexes without needing an intermediate interaction graph.
Limitations
The model may not capture overlapping complexes and relies on the quality of the input data.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website