Predictive medicine

Since the formulation of precise and well defined substructure queries is not trivial, other approaches to identify unwanted substructures are used as well. If the decisions made by medicinal chemists whether to accept or reject individual screening hits based on purely structural criteria has been captured, this can be used to train a statistical model predicting medicinal chemists judgment on chemical structures. The consensus among medicinal chemists has been demonstrated to be limited [48]. Therefore, this exercise must be based on the decision of a larger group of chemists in order not to bias the model towards the preferences of any individual chemist. In a similar way such models can be trained on experimental toxicity data for an individual experimentally determined toxicological endpoint. 

I am confident that with the approach of the postgenome era, an ever-increasing number of human genes will be discovered and their functions elucidated. Combined with the knowledge of human gene polymorphism, genotyping will allow prediction of the genetic predisposition to certain diseases, such as cancer. The new millennium will usher us in a new era of disease-predictive medicine. 

Reddy G, Dalmasso EA. SELDI ProteinChip(R) array technology protein-based predictive medicine and drug discovery applications. J Biomed Biotechnol 2003 2003 237-241. 

Health care will be transformed over the next decade from our current reactive medicine to a new medical practice that is predictive, personalized, preventive, and participatory (P4). Predictive medicine will have two major components 1) Individual... 

Various approaches, including pharmacogenomics, make up the emerging field of predictive medicine. These advances allow clinicians to predict (somewhat) the risk of disease based on genetic testing and to predict whether a particular therapy will be effective or pose a risk of adverse effects in a particular patient. 

Nova Molecular, Inc. (Montreal, Canada) Predictive Medicine (Cambridge, MA) Rosetta Inpharmatics (Kirkland, WA)... 

Eriksson M A L, J Pitera and P A Kollman 1999. Prediction of the Binding Free Energies of New TIBO-like HIV-1 Reverse Transcriptase Inhibitors Using a Combination of PROFEC, PB/SA, CMC/MD, and Free Energy Calculations. Journal of Medicinal Chemistry 42 868-881. 

Pastor M, G Cruciani and S dementi 1997. Smart Region Definition A New Way to Improve tl Predictive Ability and Interpretability of Three-Dimensional Quantitative Structure-Activi Relationships. Journal of Medicinal Chemistry 40 1455-1464. 

In a different context, a micropipette has been applied to monitor the current through a single-ion channel in a biological membrane. The patch-clamp technique invented by Sackmann and Neher [119] led to their Nobel Prize in medicine. The variations in channel current with voltage, concentration, type of ions, and type of channels have been explored. While the functions of specific channels, in particular their ionic selectivity, have been well known, only a handful of channels have the internal geometry and charge distribution determined. The development of a theory to interpret the mass of channel data and to predict channel action is still lacking. 

Hood L, Heath JR, Phelps ME, Lin B. Systems biology and new technologies enable predictive and preventative medicine. Science 2004 306 640-3. 

Even as the computational prediction error rate is reduced to acceptable levels, many cases will be encountered in which the predictions are indistinguishable to within error. In a scenario in which several different in silico designs are given equivalent but favorable activity predictions, the end user s medicinal experience may help decide which to promote to synthesis. The quality of that decision at this point will be strongly influenced by how easy it is to understand the different contributions to the computational predictions. Interpretability is thus critical for synergistically utilizing the experience of the end user. 

In the context of drug discovery, computational methods do not add value unless they can achieve practical results. Results must be produced quickly enough so that they can influence decision making in chemical synthesis. Most importantly, computational methods must be accurate enough to maintain the trust of the medicinal chemist. Without this trust, computational predictions will rarely be tested in the laboratory, which will then prevent the generation of critical data useful for improving the original predictions. 

---