New Drug Targets from Protein-Chemical Complexes
Author Information
Author(s): Kalinina Olga V., Wichmann Oliver, Apic Gordana, Russell Robert B.
Primary Institution: Cell Networks, University of Heidelberg
Hypothesis
Can existing protein-chemical interactions predict new drug targets?
Conclusion
The study successfully predicts 19,578 novel protein-chemical interactions, including 7,793 involving known drugs.
Supporting Evidence
- The method reproduces 84% of known complexes in a benchmark.
- Predictions include structures for known protein-drug interactions lacking a structure.
- The growth rate of confident predictions is twice that of experimental complexes.
Takeaway
Scientists used known protein structures to guess new ways drugs might work with proteins, finding many new possibilities.
Methodology
The study used protein and chemical superimpositions combined with structural compatibility screens to predict interactions.
Potential Biases
The method may be biased against proteins that are difficult to solve, particularly membrane proteins.
Limitations
The predictions may include false positives due to the lack of known negative interactions.
Statistical Information
P-Value
p≤0.05
Statistical Significance
p≤0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website