Scoring consensus

0 if outside this range un is the number of predicted log fCnw values used Cr, is 

1 if the candidate is within the boundaries of the given retention index criterion and 

0 if it is outside this boundary is the number of R1 criteria used Cr is 1 if the conformational energy is below the 90 percentile of a given energy calculation and 0 if this is above this value, while is the number of energy calculations used. 

---

The calculation of the binding affinity with con.sidcration of all tbc.se effects for virtual screening is not possible. In order to circumvent thus difficulty, scoring functions arc used instead, c.g., the Liidi scoring function [80, or consensus scoring functions derived from FlevX score, DOCK score, GOLD score, ChemScore, or PMF score [81 ]. 

Charifson P S, J J Corkery, M A Murcko and W P Walters 1999. Consensus Scoring A Method fc Obtaining Improved Hit Rates from Docking Databases of Three-Dimensional Structures int Proteins. Journal of Medicinal Chemistry 42 5100-5109. 

Charifson PS, Corkery JJ, Murcko MA, Walters,WP. Consensus scoring a method for obtaining improved hit rates from docking databases of three-dimensional structures into proteins. J Med Chem 1999 42 5100-9. 

Wang RX, Wang SM. How does consensus scoring work for virtual library screening An idealized computer experiment. J Chem Inf Comput Sci 2001 41 1422-6. 

A machine-learning method was proposed by Klon et al. [104] as an alternative form of consensus scoring. The method proved unsuccessful for PKB, but showed promise for the phosphatase PTPIB (protein tyrosine phosphatase IB). In this approach, compounds were first docked into the receptor and scored using conventional means. The top scoring compounds were then assumed to be active and used to build a naive Bayes classification model, all compounds were subsequently re-scored and ranked using the model. The method is heavily dependent upon predicting accurate binding... 

Wang RX, Wang SM. (2001) How does Consensus Scoring Work for Virtual Library Screening An Ideahzed Computer Experiment. /. Chem. Inf. Comp. Set. 41 1422-1426. 

Key Words Prioritization compound quality structural diversity consensus scoring regularization molecular complexity structural alerts biological promiscuity. 

The D-score is computed using the maximum dissimilarity algorithm of Lajiness (20). This method utilizes a Tanimoto-like similarity measure defined on a 360-bit fragment descriptor used in conjunction with the Cousin/ChemLink system (21). The important feature of this method is that it starts with the selection of a seed compound with subsequent compounds selected based on the maximum diversity relative to all compounds already selected. Thus, the most obvious seed to use in the current scenario is the compound that has the best profile based on the already computed scores. Thus, one needs to compute a preliminary consensus score based on the Q-score and the B-score using weights as defined previously. To summarize this, one needs to... 

Compute the preliminary consensus score (e.g., Q-score +. 5 B-score)... 

After all the individual scores have been computed and transformed appropriately, one can define the consensus score. This score should weight the various components according to the desires of the project team. For example, the Consensus Score (CS) could be define as ... 

An example of a final consensus list can be seen in Fig. 6. In this figure one can see the Q-score and B-score and the computed preliminary consensus score. On the basis of the preliminary consensus, NP-103930 was chosen as the best compound and selected to be the dissimilarity seed. After the maximum dissimilarity calculation, the diversity score was input and the final consensus score was calculated. As one can see from this figure, the first compound in the preliminary run remains the best. The second compound from the preliminary run does not appear in this list as it was very similar to the NP-103930 and was de-prioritized and moved down the list accordingly. Also the 155th compound in the preliminary ranking moved up the 14th rank because it was considered as a structurally novel compound. This, we feel, illustrates the power of this approach. Compounds with the most desirable properties move up the list and compounds with less desirable properties move down the list. 

Once the final consensus score has been calculated for all compounds, the lists were divided into smaller compound sets for convenience. In one particular example, the total set was split into 6 sets of approx 400 compounds each. This is illustrated in Fig. 7. Selected lists were then sent out for plating and subsequent testing. 

What has been described in this chapter is an approach where one prioritizes compounds for follow-up screening instead of filtering. One starts with a list of compounds ordered by some parameter such as percentage inhibition and then calculates scores that are reflective of various measures of desirability. These scores are combined into a consensus score and then are used to reprioritize the list so that compounds with desirable features are near the top of the list and less desirable compounds move near the bottom of the list. These compounds can be organized into groups to facilitate analysis. It is anticipated that this is a dynamic process and evolves as more experience is gained. 

---