Blurring contact maps of thousands of proteins: what we can learn by reconstructing 3D structure
2011

Reconstructing 3D Protein Structures from Contact Maps

Sample size: 1716 publication 10 minutes Evidence: high

Author Information

Author(s): Vassura Marco, Di Lena Pietro, Margara Luciano, Mirto Maria, Aloisio Giovanni, Fariselli Piero, Casadio Rita

Primary Institution: Department of Computer Science, University of Bologna

Hypothesis

How can we improve the reconstruction of protein structures from contact maps despite the presence of random noise?

Conclusion

The study shows that 3D reconstruction quality remains high even when up to 75% of real contacts are deleted, as long as a sufficient number of correct contacts are retained.

Supporting Evidence

  • The algorithm FT-COMAR was tested on the largest dataset of proteins to date.
  • Reconstruction quality was shown to be independent of protein length when sufficient correct contacts were present.
  • The study found that random errors on contacts were more tolerable than errors on non-contacts.

Takeaway

This study helps scientists understand how to build 3D models of proteins even when some information is missing or noisy.

Methodology

The study used a GRID technology to benchmark the FT-COMAR algorithm on a dataset of 1716 non-redundant proteins, analyzing the effects of random noise on 3D reconstruction.

Limitations

The study primarily focused on the effects of random noise and did not explore other potential factors affecting reconstruction quality.

Participant Demographics

The dataset consisted of 1716 protein chains with sequence similarity less than 25%.

Digital Object Identifier (DOI)

10.1186/1756-0381-4-1

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication