SCORE selection criteria

In the case of iterative medium-throughput screening, at any given point in the process, the set of molecules that have been screened thus far is the previously selected set for the next round of screening. In choosing molecules for the next iteration, one may have a selection criterion such as predictive model scores but a diversity criterion may also be applied it is not desirable to screen something identical, or nearly identical, to that which was screened in previous rounds. 

In the prioritization step, all remaining structures that have not been rejected in the selection step are scored using the function described in section 8.3. It is possible to specify a lower limit for the score. LUDI will then accept only those structures with a score better than the user-defined threshold value. In addition, LUDI also tries to estimate the possible maximum score for each fragment (assuming a fully buried surface and the formation of hydrogen bonds with all polar groups of the fragment). The ratio actual score/possible maximum score can also be used as a selection criterion. 

Figure 2A shows a poor selection according to UCC in ID, the second x, bin and the first and last x2 bins are empty. Similarly, in 2D, the lower left cell is empty. Correspondingly, there are overrepresented cells with more than one molecule selected. Mathematically, the UCC criterion averages penalty contributions of the form (n - c)2 across all cells in all subspaces, where for a particular cell n is the number of molecules selected and c is the ideal for the cell (1 if the cell has at least one molecule in the database falling in that cell, 0 otherwise). The design in Fig. 2A generates penalties for the empty and over-populated ID or 2D cells mentioned above. In contrast, the selection in Fig. 2B has a perfect UCC score of 0 as all cells in all subspaces are occupied once. 

The criterions have considered the distribution of the samples in PCA score plot, and 10 samples among outliers and those that represent most part of the analysed samples were selected. 

Figure 7.7 Rank-by-feature framework interface for scatterplots (2D) Scatterplot Ordering. All 2D scatterplots are ordered according to the current ordering criterion (a) in the ordered list (c). Users can select multiple scatterplots at the same time and generate separate scatterplot windows to compare them in a screen. The score overview (f>) shows an overview of scores of all scatterplots. A mouseover event activates a cell in the score overview, highlights the corresponding item in the ordered list (c), and shows the corresponding scatterplot in the scatterplot browser d) simultaneously. A click on a cell at the score overview selects the cell and the selection is fixed until another click event occurs in the score overview or another selection event occurs in other views. A selected scatterplot is shown in the scatterplot browser id), where it is also easy to traverse scatterplot space by changing the x- or y-axis using item sliders on the horizontal or vertical axis. (The dataset shown is demographic- and health-related statistics for 3138 U.S. counties with 17 attributes.) (See color insert.)...

Yih and Jones (1992) proposed using multilayer perceptrons in selecting some candidate rules for further evaluation of their performance. In their proposed approach, a multilayer perceptron wiU take the attributes describing the system configuration and the performance measures and wfll output a proper matching score for each dispatching rule. They also used this approach for multiple-criterion objectives. 

With the field narrowed to a reasonable number of candidates, it is now possible to devote a little effort to refining each of them, get to understand them better, and then construct a new matrix. This time a weighting factor should be agreed to to each of the criteria. A range of 5 for very important and 1 for minimally important will suffice. Now score each of the concepts on a basis of 1 to 10 on each criterion and compute the sums of the weighted scores. This will allow the ideas to be ranked and a selection made of the ideas to be seriously develr d. 

---