ChemoSoft is equipped with various diversity
sorting systems enabling it to flexibly adjust the systems to any customer's needs. Diversity sorting
procedures comprise structural diversity sorting (matrix maximization and stepwise elimination
algorithms), Diversity of Heterocycles, and Diversity of Plates. Each algorithm allows optimization of both
intra- and interlibrary diversities.

Structural diversity sorting employs structural spherical radii as screens, while the center of each atom in a molecule being chosen as a center of the radii in its turn. A particular feature of ChemoSoft's algorithm is that the nature of cyclic bonds is taken into account. Cosine coefficient is used as a similarity measure. Please refer to the Request\Info section for more information on the topic.

Sorting by structural diversity in ChemoSoft. Intra- and inter-library diversity.

Diversity of Heterocycles is a unique algorithm in ChemoSoft. A distinct feature of Diversity of Heterocycles is that it reveals heterocyclic cores (scaffolds) of a collection compounds and then uses them as screens for further sorting by diversity. Compounds that do not contain heterocycles settle to the end of a sorting list in this algorithm. The methodology enables you to enrich a collection of compounds with rare and unique heterocycles. Please refer to the Request\Info page for more information on the topic.

Sorting by Diversity of Heterocycles in ChemoSoft. Intra- and inter-library diversity.

Diversity of Plates (another unique utility of ChemoSoft) was invented as a tool for sorting sets of compounds (not compounds in a set) by diversity. For instance, compounds placed in 96-vial plates can be sorted in such a manner that the plates are sorted by diversity. A "plate" is considered a supermolecule in this algorithm. Such a procedure is important when "cherry pick" is impossible or time-consuming and expensive (e.g. for large compound collections after a screening).

Sorting by Diversity of Plates in ChemoSoft.

Selection by similarity is represented by a number of tools in ChemoSoft. They make it possible to select substances based on structural similarity, bioisoster similarity, and descriptor (property) similarity space.

Structural similarity includes the following functions: sorting by similarity, minimum dissimilarity selection and clustering by similarity. All the tools use the Tanimoto coefficient as a similarity measure. Minimum dissimilarity selection employs Monte Carlo algorithm.

Sorting by similarity.	Minimum dissimilarity selection.	Clustering by similarity.

Bioisoster similarity of ChemoSoft is our revolution innovation in the field of computer-assisted drug design, which turns intuitive concepts of bioisosterism into a practical realization. This engine enables you to search compounds based on predefined bioisoster rules that are user-customizable and adjustable. In other words, you can select necessary rules, edit them, and enter new ones. You do not have to select compounds manually. All is automated from now on.

Bioisoster similarity. Adjustable rules. An input and a result (a bioisoster analogue of Bayer's aspirin).

Property similarity is realized in the "Explore data" module along with tools for QSAR studies and other statistics. The function called "Map objects" virtually breaks a database down into clusters showing the results in a Sammon's plot. Thus, you can mark and select clusters of your interest for further research or make a conclusion on voids in your collection.