Method for selecting an optimally diverse library of small molecules based on validated molecular structural descriptors

A computer implemented method of determining the neighborhood distance characteristic of a validated molecular structural descriptor using multiple literature data sets containing a variety of chemical structures and associated activities, comprising the following steps:

• a. applying the molecular structural descriptor to all compounds represented in each data set to derive descriptor values;
• b. constructing a Patterson plot for each molecular structural descriptor for each data set using the descriptor values for the compounds in each data set and their associated activities;
• c. determining the appropriate Patterson plot line for each data set;
• d. using for each set a point on the axis of the corresponding Patterson plot the end point of an activity difference for which a neighborhood distance is descried, determining the X axis values of the molecular structural descriptor corresponding to the projection from the Patterson plot line of the end points of the activity difference;
• e. determining the average range of values for the neighborhood distance from the plots for each of the data sets.

A computer implemented method of applying a molecular structural descriptor to a set of reactants comprising the following steps:

• a. topomerically aligning the reactants;
• b. determining the CoMFA steric fields for each topomerically aligned reactant;
• and
• c. calculating the field differences between all pair of reactants.

A computer implemented method for selecting, for all possible product molecules which could be created in a combinatorial synthesis from specified reactant molecules and common core molecule, a subset of product molecules, comprising the following steps:

• a. characterizing all the reactant molecules with a molecular structural descriptor, validated as possessing a neighborhood property, appropriate to reactant molecules;
• b. heirarchically clustering the characterized reactant molecules until the intercluster distance substantially corresponds to the neighborhood distance of the validated molecular structural descriptor;
• c. selecting a reactant molecule from each cluster;
• d. combinatorially assembling the selected reactant molecules and core molecule into products which would be created in the chemical synthesis;
• e. selecting a product molecule for inclusion in the subset;
• f. using a molecular structural descriptor, validated as possessing a neighborhood property, appropriate to whole molecules, calculating the descriptor distance between all selected product molecules and all other product molecules;
• g. determining the shortest descriptor distance between each product molecule and all product molecules previously selected;
• h. selecting for inclusion in the subset the product molecule whose shortest descriptor distance from the previously selected molecules is the largest and is greater than the neighborhood distance of the descriptor;
• i. repeating steps f through h until the largest shortest difference between molecules is less than the neighborhood distance of the descriptor;
• and
• j. outputing a list of the selected product molecules and/or the reactant molecules from which the selected product molecules can be formed.

A computer implemented method for selecting, for all possible product molecules which could be created in a combinatorial synthesis from specified reactant molecules, a subset of product molecules, comprising the following steps:

• a. characterizing all the reactant molecules with a molecular structural descriptor, validated as possessing a neighborhood property, appropriate to reactant molecules;
• b. heirarchically clustering the characterized reactant molecules until the intercluster distance substantially corresponds to the neighborhood distance of the validated molecular structural descriptor;
• c. selecting a reactant molecule from each cluster;
• d. combinatorially assembling the selected reactant molecules into products which would be created in the chemical synthesis;
• e. selecting a product molecule for inclusion in the subset;
• f. using a molecular structural descriptor, validated as possessing a neighborhood property, appropriate to whole molecules, calculating the descriptor distance between all selected product molecules and all other product molecules;
• g. determining the shortest descriptor distance between each product molecule and all product molecules previously selected;
• h. selecting for inclusion in the subset the product molecule whose shortest descriptor distance from the previously selected molecules is the largest and is greater than the neighborhood distance of the descriptor;
• i. repeating steps f through h until the largest shortest difference between molecules is less than the neighborhood distance of the descriptor;
• and
• j. outputing a list of the selected product molecules and/or the reactant molecules from which the selected product molecules can be formed.

A computer implemented system for selecting, for all possible product molecules which can be created in a combinatorial synthesis from all specified reactant molecules and common core molecule, a subset of product molecules whose members collectively represent most of the molecular structural diversity in the possible combinatorially synthesized product molecules, comprising:

• a. means for characterizing all the reactant molecules with a molecular structural descriptor, validated as possessing a neighborhood property, appropriate to reactant molecules;
• b. means for hierarchically clustering the characterized reactant molecules until the intercluster distance substantially corresponds to the neighborhood distance of the validated molecular structural descriptor;
• c. means for selecting one reactant molecule from each cluster;
• d. means for combinatorially assembling the selected reactant molecules and core molecule into products which would be created in the chemical synthesis;
• e. means for selecting at least one product molecule for inclusion in the subset;
• means for using a molecular structural descriptor, validated as possessing a neighborhood property, applicable to whole molecules for calculating the descriptor distance between all selected product molecules and all other product molecules;
• g. means for determining the shortest descriptor distance between each product molecule and all product molecules previously selected;
• h. means for selecting for inclusion in the subset the product molecule whose shortest descriptor distance from the previously selected molecules is the largest and is greater than the neighborhood distance of the descriptor;
• i. means for invoking means f through h until the largest shortest difference between molecules is less than the neighborhood distance of the descriptor;
• j. means for outputing a list of the selected product molecules and/or the reactant molecules from which the selected product molecules can be formed.

A computer implemented system for selecting, for all possible product molecules which can be created in a combinatorial synthesis from all specified reactant molecules, a subset of product molecules whose members collectively represent most of the molecular structural diversity in the possible combinatorially synthesized product molecules, comprising:

• a. means for characterizing all the reactant molecules with a molecular structural descriptor, validated as possessing a neighborhood property, appropriate to reactant molecules;
• b. means for heirarchically clustering the characterized reactant molecules until the intercluster distance substantially corresponds to the neighborhood distance of the validated molecular structural descriptor;
• c. means for selecting one reactant molecule from each cluster;
• d. means for combinatorially assembling the selected reactant molecules into products which would be created in the chemical synthesis;
• e. means for selecting at least one product molecule for inclusion in the subset;
• f. means for using a molecular structure descriptor, validated as possessing a neighborhood property, applicable to whole molecules for calculating the descriptor distance between all selected product molecules and all other product molecules;
• g. means for determining the shortest descriptor distance between each product molecule and all product molecules previously selected;
• h. means for selecting for inclusion in the subset the product molecule whose shortest descriptor distance from the previously selected molecules is the largest and is greater than the neighborhood distance of the descriptor;
• i. means for invoking means f through h until the largest shortest difference between molecules is less than the neighborhood distance of the descriptor;
• and
• j. means for outputing a list of the selected product molecules and/or the reactant molecules from which the selected product molecules can be formed.

A method of screening assays against a subset of molecules which could be made in a combinatorial synthesis of specified reactants and cores in which the subset to be synthesized and tested is selected by the following computer implemented method:

• a. characterizing all the reactant molecules with a molecular structural descriptor, validated as possessing a neighborhood property, appropriate to reactant molecules;
• b. heirarchically clustering the characterized reactant molecules until the intercluster distance substantially corresponds to the neighborhood distance of the validated molecular structural descriptor;
• c. selecting a reactant molecule from each cluster;
• d. combinatorially assembling the selected reactant molecules and core molecule into products which would be created in the chemical synthesis;
• e. selecting a product molecule for inclusion in the subset;
• f. using a molecular structural descriptor, validated as possessing a neighborhood property, appropriate to whole molecules, calculating the descriptor distance between all selected product molecules and all other product molecules;
• g. determining the shortest descriptor distance between each product molecule and all product molecules previously selected;
• h. selecting for inclusion in the subset the product molecule whose shortest descriptor distance from the previously selected molecules is the largest and is greater than the neighborhood distance of the descriptor;
• i. repeating steps f through h until the largest shortest difference between molecules is less than the neighborhood distance of the descriptor;
• and
• j. outputing a list of the selected product molecules and/or the reactant molecules from which the selected product molecules can be formed.
• wherein at least some of the molecules in the subset are synthesized and at least some of the synthesized molecules are tested in at least one screening assay.

In a digital computer in which representations of specified reactant molecules and a core molecule have been stored, a computer implemented method for selecting, for all possible product molecules which could be created in a combinatorial synthesis from the reactant molecules and common core molecule, a subset of product molecules, comprising the following steps:

• a. characterizing all the reactant molecules with a molecular structural descriptor, validated as possessing a neighborhood property, appropriate to reactant molecules;
• b. heirarchically clustering the characterized reactant molecules until the intercluster distance substantially corresponds to the neighborhood distance of the validated molecular structural descriptor;
• c. selecting a reactant molecule from each cluster;
• d. combinatorially assembling the selected reactant molecules and core molecule into products which would be created in the chemical synthesis;
• e. selecting a product molecule for inclusion in the subset;
• f. using a molecular structural descriptor, validated as possessing a neighborhood property, appropriate to whole molecules, calculating the descriptor distance between all selected product molecules and all other product molecules;
• g. determining the shortest descriptor distance between each product molecule and all product molecules previously selected;
• h. selecting for inclusion in the subset the product molecule whose shortest descriptor distance from the previously selected molecules is the largest and is greater than the neighborhood distance of the descriptor;
• i. repeating steps f through h until the largest shortest difference between molecules is less than the neighborhood distance of the descriptor;
• and
• outputing a list of the selected product molecules and/or the reactant molecules from which the selected product molecules can be formed.

A computer implemented method for selecting, for all possible product molecules which can be created in a combinatorial synthesis from all specified reactant molecules and common core molecule, a subset of product molecules whose members collectively represent most of the molecular structural diversity in the possible combinatorially synthesized product molecules, providing:

• a. means for characterizing all the reactant molecules with a molecular structural descriptor, validated as possessing a neighborhood property, appropriate to reactant molecules;
• b. means for heirarchically clustering the characterized reactant molecules until the intercluster distance substantially corresponds to the neighborhood distance of the validated molecular structural descriptor;
• c. means for selecting one reactant molecule from each cluster;
• d. means for combinatorially assembling the selected reactant molecules and core molecule into products which would be created in the chemical synthesis;
• e. means for selecting at least one product molecule for inclusion in the subset;
• f. means for using a molecular structural descriptor, validated as possessing a neighborhood property, applicable to whole molecules for calculating the descriptor distance between all selected product molecules and all other product molecules;
• g. means for determining the shortest descriptor distance between each product molecule and all product molecules previously selected;
• h. means for selecting for inclusion in the subset the product molecule whose shortest descriptor distance from the previously selected molecules is the largest and is greater than the neighborhood distance of the descriptor;
• i. means for invoking means f through h until the largest shortest difference between molecules is less than the neighborhood distance of the descriptor;
• j. means for outputing a lest of the selected product molecules and/or the reactant molecules from which the selected product molecules can be formed;
• and comprising the steps of:
• a′. characterizing all the reactant molecules with a molecular structural descriptor, validated as possessing a neighborhood property, appropriate to reactant molecules;
• b′. heirarchically clustering the characterized reactant molecules until the intercluster distance substantially corresponds to the neighborhood distance of the validated molecular structural descriptor;
• c′. selecting a reactant molecule from each cluster;
• d′. combinatorially assembling the selected reactant molecules into products which wold be created in the chemical synthesis;
• e′. selecting a produce molecule for inclusion in the subset;
• f′. using a molecular structural descriptor, validated as possessing a neighborhood property, appropriate to whole molecules, calculating the descriptor distance between all selected product molecules and all other product molecules;
• g′. determining the shortest descriptor distance between each product molecule and all product molecules previously selected;
• h′. selecting for inclusion in the subset the product molecule whose shortest descriptor distance from the previously selected molecules is the largest and is greater than the neighborhood distance of the descriptor;
• i′. repeating steps f′ through h′ until the largest shortest difference between molecules is less than the neighborhood distance of the descriptor;
• and
• j′. outputing a list of the selected product molecules and/or the reactant molecules from which the selected product molecules can be formed.

A computer implemented method for selecting, for all possible produce molecules which can be create din a combinatorial synthesis from all specified reactant molecules, a subset of product molecules whose members collectively represent most of the molecular structural diversity in the possible combinatorially synthesized product molecules, providing:

• a. means for characterizing all the reactant molecules with a molecular structural descriptor, validated as possessing a neighborhood property, appropriate to reactant molecules;
• b. means for heirarchically clustering the characterized reactant molecules until the intercluster distance substantially corresponds to the neighborhood distance of the validated molecular structural descriptor;
• c. means for selecting one reactant molecules from each cluster;
• d. means for combinatorially assembling the selected reactant molecules and core molecule into products which would be created in the chemical synthesis;
• e. means for selecting at least one product molecule for inclusion in the subset;
• f. means for using a molecular structural descriptor, validated as possessing a neighborhood property, applicable to whole molecules for calculating the descriptor distance between all selected product molecules and all other product molecules;
• g. means for determining the shortest descriptor distance between each product molecule and all product molecules previously selected;
• h. means for selecting for inclusion in the subset the product molecule whose shortest descriptor distance from the previously selected molecules is the largest and is greater than the neighborhood distance of the descriptor;
• i. means for invoking means f through h until the largest shortest difference between molecules is less than the neighborhood distance of the descriptor;
• j. means for outputing a list of the selected product molecules and/or the reactant molecules from which the selected product molecules can be formed;
• and comprising the steps of:
• a′. characterizing all the reactant molecules with a molecular structural descriptor, validated as possessing a neighborhood property, appropriate to reactant molecules;
• b′. heirarchically clustering the characterized reactant molecules until the intercluster distance substantially corresponds to the neighborhood distance of the validated molecular structural descriptor;
• c′. selecting a reactant molecule from each cluster;
• d′. combinatorially assembling the selected reactant molecules into products which would be created in the chemical synthesis;
• e′. selecting a product molecule for inclusion in the subset;
• f′. using a molecular structural descriptor, validated as possessing a neighborhood property, appropriate to whole molecules, calculating the descriptor distance between all selected product molecules and all other product molecules;
• g′. determining the shortest descriptor distance between each product molecule and all product molecules previously selected;
• h′. selecting for inclusion in the subset the product molecule whose shortest descriptor distance from the previously selected molecules is the largest and is greater than the neighborhood distance of the descriptor;
• i′. repeating steps f′ through h′ until the largest shortest difference between molecules is less than the neighborhood distance of the descriptor;
• and
• j′. outputing a list of the selected product molecules and/or the reactant molecules from which the selected product molecules can be formed.

A system for selecting, for all possible product molecules which can be created in a combinatorial synthesis from all specified reactant molecules and common core molecule, a subset of product molecules whose members collectively represent most of the molecular structural diversity in the possible combinatorially synthesized product molecules, comprising:

• a. a general purpose digital computer;
• b. means for characterizing all the reactant molecules with a molecular structural descriptor, validated as possessing a neighborhood property, appropriate to reactant molecules;
• c. means for heirarchically clustering the characterized reactant molecules until the intercluster distance substantially corresponds to the neighborhood distance of the validated molecular structural descriptor;
• d. means for selecting one reactant molecule from each cluster;
• e. means for combinatorially assembling the selected reactant molecules and core molecule into products which would be created in the chemical synthesis;
• f. means for selecting at least one product molecule for inclusion in the subset;
• g. means for using a molecular structural descriptor, validated as possessing a neighborhood property, applicable to whole molecules for calculating the descriptor distance between all selected product molecules and all other product molecules;
• h. means for determining the shortest descriptor distance between each product molecule and all product molecules previously selected;
• i. means for selecting for inclusion in the subset the product molecule whose shortest descriptor distance from the previously selected molecules is the largest and is greater than the neighborhood distance of the descriptor;
• j. means for invoking means f through h until the largest shortest difference between molecules is less than the neighborhood distance of the descriptor;
• and
• k. means for outputing a list of the selected product molecules and/or the reactant molecules from which the selected product molecules can be formed.

A computer implemented system for selecting, for all possible product molecules which can be created in a combinatorial synthesis from all specified reactant molecules, a subset of product molecules whose members collectively represent most of the molecular structural diversity in the possible combinatorially synthesized product molecules, comprising:

• a. a general purpose digital computer;
• b. means for characterizing all the reactant molecules with a molecular structural descriptor, validated as possessing a neighborhood property, appropriate to reactant molecules;
• c. means for heirarchically clustering the characterized reactant molecules until the intercluster distance substantially corresponds to the neighborhood distance of the validated molecular structural descriptor;
• d. means for selecting one reactant molecule from each cluster;
• e. means for combinatorially assembling the selected reactant molecules into products which would be created in the chemical synthesis;
• f. means for selecting at least one product molecule for inclusion in the subset;
• g. means for using a molecular structural descriptor, validated as possessing a neighborhood property, applicable to whole molecules for calculating the descriptor distance between all selected product molecules and all other product molecules;
• h. means for determining the shortest descriptor distance between each product molecule and all product molecules previously selected;
• i. means for selecting for inclusion in the subset the product molecule whose shortest descriptor distance from the previously selected molecules is the largest and is greater than the neighborhood distance of the descriptor;
• j. means for invoking means f through h until the largest shortest difference between molecules is less than the neighborhood distance of the descriptor;
• and
• k. means for outputing a list of the selected product molecules and/or the reactant molecules from which the selected product molecules can be formed.