How complete are current yeast and human protein-interaction networks?

Area of Science:

• Biochemistry
• Systems Biology
• Bioinformatics

Background:

• Protein-protein interactions (PPIs) are fundamental to cellular processes.
• Current protein-protein interaction network maps are incomplete due to experimental limitations and high false-positive rates.
• Understanding the scale and accuracy of PPI networks is crucial for biological discovery.

Purpose of the Study:

• To estimate the total number of protein-protein interactions in yeast and human interactomes.
• To assess the completeness of current protein-protein interaction network maps.
• To explore strategies for improving the accuracy of protein-protein interaction data.

Main Methods:

• Computational analysis and network inference to estimate the size of yeast and human PPI networks.
• Evaluation of current experimental data completeness based on estimated network sizes.
• Analysis of the impact of false-positive rates on network accuracy.

Main Results:

• The full yeast PPI network is estimated to contain 37,800–75,500 interactions.
• The full human PPI network is estimated to contain 154,000–369,000 interactions.
• Current yeast PPI maps are approximately 50% complete, while human maps are only 10% complete due to high false-positive rates.

Conclusions:

• Despite significant progress, current protein-protein interaction network maps are far from complete.
• High false-positive rates in experimental assays are a major bottleneck for accurate interactome mapping.
• Releasing raw, unfiltered assay data may paradoxically improve the identification of true protein-protein interactions.