Predicting Transcription Factor Binding Sites Using Protein-DNA Models
Author Information
Author(s): Angarica Vladimir Espinosa, Pérez Abel González, Vasconcelos Ana T, Collado-Vides Julio, Contreras-Moreira Bruno
Primary Institution: Universidad de Zaragoza
Hypothesis
Can atomic-detail structural information be used to predict transcription factor binding sites?
Conclusion
The study demonstrates that atomic-detail structural information can effectively predict transcription factor binding sites.
Supporting Evidence
- The algorithm was tested on eight transcription factors from different structural superfamilies.
- The method showed comparable predictions to existing structural methods.
- Sampling of interface side-chain rotamers improved prediction accuracy.
Takeaway
The researchers created a computer program that helps predict where proteins attach to DNA by looking closely at how they fit together.
Methodology
The study used a structure-based methodology to predict sequence motifs from TF-DNA complex coordinates, combining direct and indirect readout information into a statistical model.
Potential Biases
Potential bias due to reliance on existing protein-DNA complex data from the PDB.
Limitations
The method may not perform well for transcription factors with few or no experimentally characterized binding sites.
Statistical Information
P-Value
1.49E-04
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website