Machine virtual screening

Jorissen RN, Gilson MK. Virtual screening of molecular databases using a Support Vector Machine. J Chem Inf Model 2005 45 549-61. 

Machine Learning as a general technique is quite broad topic and its application in virtual screening could easily fill a chapter on its own [139]. Therefore, similar to the topic on numerical optimization, only the tip of the iceberg can be covered here. 

Hert J, WiUett P, Wilton DJ, Addin P, Azzaoui K, Jacoby E, Schuffenhauer A. (2005) New Methods for Ligand-Based Virtual Screening Use of Data-Fusion and Machine-Learning Techniques to Enhance the Effectiveness of Similarity Searching. /. Chem. Inf. Model, (in the press). 

Melville, J. L., Burke, E. K, Hirst, J. D. (2009) Machine learning in virtual screening. Comb Chem High Throughput Screening 12, 332-343. 

With increasing numbers of X-ray structures being solved, the 3D information can be exploited for ligand design and optimization. There is a need for fast methods for structure-based virtual screening. The recent use of machine clusters and porting of software codes to Linux and Windows platforms has contributed to a signif-... 

Todeschini R, Consonni V (2000) Handbook of molecular descriptors. In Mannhold R, Kubinyi H, Timmerman H (eds), Methods and principles in medicinal chemistry 11. WILEY-VCH, Weinheim. Walters WP, Stahl MT, Murcko MA (1998) Virtual screening — and overview. Drug Discov Today 3 160—178. Ward JH (1963) Hierarchical grouping to optimize an objective function. J Am Stat Assoc 58 236—244. Warmuth MK, Liao J, Ratsch G et al. (2003) Active learning with support vector machines in the drug discovery process. J Chem Inf Comput Sci 43 667—673. 

Li LW, Khanna M, Jo IH et al (2011) Target-specific support vector machine scoring in structure-based virtual screening computational validation, in vitro testing in kinases, and effects on lung cancer cell proliferation. J Chem Inf Model 51(4) 755-759... 

Ma, X.H. et al. 2008. Evaluation of virtual screening performance of support vector machines trained by sparsely distributed active compounds. J. Chem. Inf. Model. 48, 1227-1237. 

Simplistic and heuristic similarity-based approaches can hardly produce as good predictive models as modern statistical and machine learning methods that are able to assess quantitatively biological or physicochemical properties. QSAR-based virtual screening consists of direct assessment of activity values (numerical or binary) of all compounds in the database followed by selection of hits possessing desirable activity. Mathematical methods used for models preparation can be subdivided into classification and regression approaches. The former decide whether a given compound is active, whereas the latter numerically evaluate the activity values. Classification approaches that assess probability of decisions are called probabilistic. 

Many different methods can be applied to virtual screening, and such methods are described in other chapters of this book and/or in the Handbooks of Che-minformatics Here we discuss the methods based on a probabilistic approach. Unfortunately, there are many publications in which the probabilistic or statistical approach items are farfetched. The Binary Kernel Discrimination and the Bayesian Machine Learning Models are actually special... 

Hert, J., Willett, P., Wilton, D. J., Acklin, P., Azzaoui, K., Jacoby, E., and Schuffenhauer, A. (2006)Newmethodsfor ligand-based virtual screening use of data fusion and machine... 

To sum up, independent of the method applied, model building is a very delicate task and a lot of experience is required to reliably constmct stable and predictive models. This is some drawback in comparison to similarity searches whose application also requires experience but is conceptionally simpler. However, in cases where a sufficient number of active and inactive compounds are available, machine learning techniques are valuable tools for virtual screening (see Section 3.4.3.4). 

In the following, we will discuss very few important techniques for machine learning. There exists a wealth of methods [69] and we try to focus here on the ones primarily applied to virtual screening. The list, however, is not exhaustive and the interested reader is redirected to excellent machine learning literature in order to get the full picture where details of the applicability domain of models and error estimations are also discussed [69-72]. 

The outcome of virtual screening is very much method, target, and data set dependent. It would be very helpful for practitioners to be able to predict the chance for success of a prospective virtual screening. Building models with machine... 

---