Binding Site Prediction for Protein-Protein Interactions and Novel Motif Discovery using Re-occurring Polypeptide Sequences

2011

Predicting Protein Binding Sites Using PIPE-Sites

Sample size: 688 publication 10 minutes Evidence: high

Author Information

Author(s): Amos-Binks Adam, Patulea Catalin, Pitre Sylvain, Schoenrock Andrew, Gui Yuan, Green James R, Golshani Ashkan, Dehne Frank

Primary Institution: Carleton University

Can PIPE-Sites accurately predict protein binding sites using only sequences of query proteins and known binary interactions?

PIPE-Sites is an accurate method for predicting protein binding sites and is applicable to the proteome-scale.

PIPE-Sites predictions were closer to confirmed binding sites than existing methods.
The method was validated using a dataset of 265 yeast and 423 human interacting protein pairs.
PIPE-Sites showed improved performance compared to domain-domain based methods.

The study created a tool called PIPE-Sites that helps find where proteins stick together, which is important for understanding how cells work.

The method uses known binary interactions and sequences of query proteins to predict binding sites.

Potential bias due to the reliance on existing databases for training data.

The method may not perform well on landscapes with low overall height and may exclude spurious peaks.

The study involved protein pairs from humans and yeast.

p<10-21 for human, p<10-20 for yeast

p<0.05

Access the complete publication on the publisher's website