Sparse QSAR modelling methods for therapeutic and regenerative medicine
- 32 Downloads
The quantitative structure–activity relationships method was popularized by Hansch and Fujita over 50 years ago. The usefulness of the method for drug design and development has been shown in the intervening years. As it was developed initially to elucidate which molecular properties modulated the relative potency of putative agrochemicals, and at a time when computing resources were scarce, there is much scope for applying modern mathematical methods to improve the QSAR method and to extending the general concept to the discovery and optimization of bioactive molecules and materials more broadly. I describe research over the past two decades where we have rebuilt the unit operations of the QSAR method using improved mathematical techniques, and have applied this valuable platform technology to new important areas of research and industry such as nanoscience, omics technologies, advanced materials, and regenerative medicine. This paper was presented as the 2017 ACS Herman Skolnik lecture.
KeywordsQuantitative structure–activity relationships QSAR Machine learning Deep learning Sparse feature selection Regenerative medicine Skolnik award
I would like to acknowledge the very talented members of my group, Frank Burden (my long-term collaborator in neural networks), Vidana, Epa, Anna Tarasova, Julianne Halley, Mitch Polley, Tu Le, and my current collaborators at CSIRO, Imperial College, MIT, and Nottingham. Their contributions are captured in the cited publications and I’m extremely grateful for their dedication and valuable intellectual contributions. I’ve also been very fortunate to have some excellent mentors during my career. I’m especially grateful to Prof. Toshio Fujita, Prof. Peter Andrews, and Prof. Graham Richards for valuable guidance and mentorship. I would also like to thank the ACS for the Herman Skolnik award and travel support, and the speakers in the Skolnik symposium for their great support. Support of UK Engineering and Physical Sciences Research Council (EPSRC) Grant EP/N006615/1 for the Programme Grant in Next Generation Biomaterials Discovery and a Monash University-Nottingham travelling fellowship are also gratefully acknowledged.
- 1.Mitchell M (2009) Complexity: a guided tour. Oxford University Press, OxfordGoogle Scholar
- 29.Burden FR, Winkler DA (2009) Bayesian regularization of neural networks. In: Livingston D (ed) Artificial neural networks: methods and applications, vol 458. Humana Press, TotowaGoogle Scholar
- 34.Winkler DA, Le TC (2017) Mol Inf 36:(1–2)Google Scholar
- 39.Mauri A, Consonni V, Pavan M, Todeschini R (2006) Match Commun Math Comput Sci 56(2):237Google Scholar
- 46.Puentedura RR (2003) The Baldwin effect in the age of computation. In: Weber BH, Depew DJ (eds) Evolution and learning: the Baldwin effect reconsidered. MIT Press, CambridgeGoogle Scholar
- 47.Hinton GE, Nowlan SJ (1987) Complex Syst 1:495Google Scholar