Virtual screening of chemical spaces

A computer-implemented method for searching molecules from among a number of molecules to be screened that are similar to a given set of query molecules, comprising the steps of:

• a) accessing data representing the molecules stored on a computer-readable medium;
• b) determining maximum common substructures of a given set of molecules based on the data representing the given set of molecules;
• c) generating a directed acyclic graph of the maximum common substructures (mDAG);
• d) generating, for each said query molecule, a data set comprising data representing the query molecule, data representing a molecule to be screened locally located in the directed acyclic graph, and data representing their respective maximum common substructure;
• and
• e) storing the data,
• f) wherein all of the foregoing steps are performed using a computer,
• wherein in step c), a directed acyclic graph of the mDAG of a given set of molecules is provided, wherein the vertices are colored graphs, wherein all molecules to be screened and the query molecules are contained as vertices within the mDAG, wherein for each pair of molecules to be screened and the query molecules the mDAG contains paths such that the molecules are connected with their respective maximum common substructure, and wherein molecules are included using different protonation states and representations.

A computer-implemented method for searching molecules from among a number of molecules to be screened that are similar to a given set of query molecules, comprising the steps of:

• a) accessing data representing the molecules stored on a computer-readable medium;
• b) determining maximum common substructures of a given set of molecules based on the data representing the given set of molecules;
• c) generating a directed acyclic graph of the maximum common substructures (mDAG);
• d) generating, for each said query molecule, a data set comprising data representing the query molecule, data representing a molecule to be screened locally located in the directed acyclic graph, and data representing their respective maximum common substructure;
• and
• e) storing the data,
• f) wherein all of the foregoing steps are performed using a computer,
• wherein in step c), a directed acyclic graph of maximum common substructures (mDAG) of a given set of molecules is provided, wherein the vertices are colored graphs, wherein all molecules to be screened and the query molecules are contained as vertices within the mDAG, wherein for each pair of molecules to be screened and the query molecules the mDAG contains paths such that the molecules are connected with their respective maximum common substructure, and wherein a weighting schema is applied to determine the maximum weight common substructures (targeted mDAG).