Abstract

The recent explosion of high throughput experimental technologies for characterizing protein interactions has generated large amounts of data describing interactions between thousands of proteins and producing genome scale views of protein assemblies. The systems level views afforded by these data hold great promise of leading to new knowledge but also involve many challenges. Deriving meaningful biological conclusions from these views crucially depends on our understanding of the approximation and biases that enter into deriving and interpreting the data. The challenges and rewards of interaction proteomics are reviewed here using as an example the latest comprehensive high throughput analyses of protein interactions in yeast.

Highlights

  • The recent explosion of high throughput experimental technologies for characterizing protein interactions has generated large amounts of data describing interactions between thousands of proteins and producing genome scale views of protein assemblies

  • Taking as an example the latest high throughput analyses by tandem affinity purification of the yeast interaction proteome [7, 8, 20, 21], we review the complex procedures of generating the raw experimental data and translating these data into meaningful descriptions of the biological reality

  • The estimated average error rate of the high confidence (HC) portion of this network, comprising 1622 proteins and 9074 interactions, is lower than for several recently derived HC networks and similar to that of the data set of binary interactions identified by small scale experiments and annotated by the Munich Information Center for Protein Sequences (MIPS)

Read more

Summary

Challenges and Rewards of Interaction Proteomics*

The recent explosion of high throughput experimental technologies for characterizing protein interactions has generated large amounts of data describing interactions between thousands of proteins and producing genome scale views of protein assemblies. The sets of protein components identified in thousands of purification runs are not the final product of a high throughput TAP-MS study but rather an important raw material that must be processed further to derive information on protein complexes that form in the cell. These assemblies showed little overlap with the Toronto complexes

Identified Partners
EVALUATING THE ERROR RATES OF INTERACTION NETWORKS AND COMPLEXES
Complexes by graph partitioning
DESCRIPTIONS OF MODEL ORGANISM INTERACTOMES ARE BECOMING INCREASINGLY ACCURATE
Within complexes Between complexes Random networks
COVERAGE REMAINS AN ISSUE
Findings
CONCLUDING REMARKS
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call