Predicting drug side-effect profiles: a chemical fragment-based approach
2011

Predicting Drug Side-Effects Using Chemical Structures

Sample size: 888 publication Evidence: high

Author Information

Author(s): Pauwels Edouard, Stoven Véronique, Yamanishi Yoshihiro

Primary Institution: Mines ParisTech

Hypothesis

Can chemical structures of drug candidates be used to predict their potential side-effect profiles?

Conclusion

The proposed method effectively predicts potential side-effects of drug candidates based on their chemical structures.

Supporting Evidence

  • The method predicted 1385 side-effects from the chemical structures of 888 approved drugs.
  • The study demonstrated high-performance prediction power of the proposed method.
  • The method is applicable on large molecular databanks.

Takeaway

This study shows a way to guess what bad effects a medicine might have by looking at its chemical makeup.

Methodology

The study used sparse canonical correlation analysis (SCCA) to correlate chemical substructures with side-effects.

Limitations

Choosing appropriate sparsity parameters and the number of components can be challenging.

Statistical Information

P-Value

almost zero

Digital Object Identifier (DOI)

10.1186/1471-2105-12-169

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication