Virtual benchmarking

Market prices are extremely volatile and follow cycles of advance and decline related to fluctuations ia global iaventories and world economic forces. For example, for northern bleached softwood kraft, which is considered a global benchmark pulp grade, market price went from 840/t ia 1990 down to 440/t ia 1993 as a result of oversupply and worldwide recession (3). By 1995 virtually all of the price losses had been recovered. 

Keywords VS, Virtual Screening, Lead discovery, lead, HTS, Pharmacophore-Based, Structure-Based, Fragment-based, Ligand-based, Docking, Scoring, hybrid workflows, VS strategy, Benchmarking VS... 

In traditional toxicological methods of determining virtually safe doses of hazardous chemicals, nominal thresholds for deterministic responses in humans are estimated based on a NOAEL obtained in human or animal studies. In most high-quality studies, NOAEL is approximately the same as the lower confidence limit of the benchmark dose that corresponds to a 10 percent increase in the number of responses. Thus, as an alternative to the benchmark dose method, the nominal threshold in humans could be set at a factor of 10 or 100 lower than NOAEL obtained in a high-quality human or animal study. However, the benchmark dose method preferred by NCRP... 

Estimated thresholds for deterministic responses in virtually all humans based on lower confidence limits of benchmark doses or NOAELs, as described above, would be used as points of departure in establishing allowable doses of chemicals that induce deterministic responses for purposes of waste classification. NCRP s recommendations on the magnitude of safety and uncertainty factors that should be applied to benchmark doses or NOAELs in classifying waste are described in Section 6.3.1.1. 

Optimizing production plants through knowledge transfer if both companies have comparable production plants and/or production processes, benchmarking on the basis of a virtual plant is suitable. This comes from combining the cost leaders in the individual process steps ... 

The study of the interactions among closed shell systems (van der Waals forces) represents a benchmark for theories of electron correlation. We report here the results of our variational MO-VB study of the interaction between two helium atoms [70]. Up to n=10 optimal virtuals are calculated and employed to generate MO-VB final wavefunctions of higher and higher accuracy, in the usual MO-VB form ... 

From the chemists viewpoint the technique has been pioneered by Kowalski, who has described it very clearly in benchmark papers30 31) reviewed it32), demonstrated that it is equally applicable to the classification of the clays used in ancient pottery and to the identification of oils from spillage incidents 33), and has made his set of computer programs ARTHUR available to the scientific community. Many others, have been active within the area and their work has been comprehensively reviewed by Kryger341 and Varmuza35) who has also described virtually all of the techniques which have been applied. All of these have important features in common ... 

According to detailed test calculations and using the standard 6-31G basis set results as the benchmark, the MEDLA electron densities are more accurate than the standard 3-21G ab initio results and are virtually indistinguishable from the results of standard 6-31G ab initio electron density calculations. " ... 

Table 10 CPU and total times per iteration to evaluate several benchmark full Cl wavefunctions using the IBM RS/6000 model 3CT. Fzc/Fzv denote the number of frozen core/frozen virtual orbitals.

Once the BMD is identified, the one-sided lower 95th confidence interval on the BMD called the BMDL (benchmark dose lower bound) is used as the POD. Similar to noncancer risk assessment, the POD is then divided by uncertainty factors to account for potential interspecies differences, intraindividual variability, and so on. The current risk assessment paradigm has accepted that, by accounting for uncertainty through use of the BMDL and other uncertainty factors, the resulting dose is either below a toxic threshold or so low as to constitute a virtually safe dose (Bogdanffy et al. 2001). 

This often used benchmark model is represented by the symmetrically stretched H2O molecule at the DZ level, for which the exact FCI results are available (35). In addition to the standard L-CCSD and CCSD correlation energies, we present the results obtained with RMR CCSD and four of its AL versions (Table 5). Similarly as for the S4 model, we employ two different reference spaces the minimal (2,2) space, involving HOMO and LUMO, and the (4,4) space involving four electrons in four orbitals (2 occupied and 2 virtual ones). For an easier comparison, we also present the corresponding differences from the FCI result in parentheses. 

Irwin, J.J. (2008) Community benchmarks for virtual screening. Journal of Computer-Aided Molecular Design, 22, 193-199. 

---