Architectures, queries, data stores, and interfaces for proteins and drug molecules

• US 20060052943A1
• Filed: 07/28/2005
• Published: 03/09/2006
• Est. Priority Date: 07/28/2004
• Status: Abandoned Application

• Current Assignee: Purdue Research Foundation (Purdue University)
• Original Assignee: Purdue Research Foundation (Purdue University)
• Inventors: Ramani, Karthik, Agrawal, Manish, Turuvekere, Srinivasan S.
• Application Number:
  US11/191,401
• Publication Number:
  US 20060052943A1
• Time in Patent Office: Days
• US Class Current: 702/19
• CPC Class Codes: G16B 15/00 ICT specially adapted for analysing two-dimensional or three-dimensional molecular structures, e.g. structural or functional relations or structure alignment
• CPC Class Codes: G16B 15/30 Drug targeting using structural data; Docking or binding prediction
• CPC Class Codes: G16B 40/00 ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding
• CPC Class Codes: G16B 45/00 ICT specially adapted for bioinformatics-related data visualisation, e.g. displaying of maps or networks
• CPC Class Codes: G16B 50/00 ICT programming tools or database systems specially adapted for bioinformatics
• CPC Class Codes: G16B 50/30 Data warehousing; Computing architectures

Abstract

Architectures, queries, data stores, and interfaces for proteins and drug molecules are provided. Active and binding sites or proteins and drug molecules are modeled and indexed in a data store. The data store may be browsed, searched, and mined to find specific proteins or to find potential drug molecules that interact with the proteins in response to the active and binding sites.

Claims

1. A method, comprising:

acquiring site coordinates and neighboring information for a receptor site of a given protein or a ligand;

determining principal curvatures and directions for each coordinate point of the receptor site or the ligand;

determining maximum span for the receptor site or the ligand;

iterating for each remaining site coordinate of the receptor site or the ligand; and

indexing the given protein, the site coordinates, the neighboring information, the points, the principal curvatures and directions, the maximum span, and the receptor site or the ligand in a data store.

2. The method of claim 1, wherein acquiring further includes acquiring each of the site coordinates as a unique three-dimensional coordinate represented within the given protein.

3. The method of claim 1 further comprising, representing the principal curvatures, the directions, the maximum span, and the neighboring information as a vector for the receptor site or the ligand and storing the vector in the data store with the given protein.

4. The method of claim 3 further comprising, forming a composite vector for the given protein from the vector associated with the receptor site or the ligand and from other vectors associated with other receptor sites of the given protein or other ligands.

5. The method of claim 1 further comprising:

receiving a candidate portion of a different protein or drug molecule;

re-iterating the method with the candidate portion or the drug molecule to acquire candidate principal curvatures, candidate directions, and candidate maximum spans; and

searching the data store with a vector representing the candidate principal curvatures, the candidate directions, and the candidate maximum spans in an attempt to find a matching receptor site or matching ligand in the data store for the candidate portion or the drug molecule.

6. A method, comprising:

acquiring a graphical representation of a given protein;

projecting the graphical representation onto a grid;

identifying cavities and protrusions within the graphical representation using the grid; and

generating a protein from the identified cavities and protrusions.

7. The method of claim 6 further comprising, associating shape features with each of the identified cavities and protrusions, wherein the shape features are retained within the generated protein model.

8. The method of claim 7 further comprising, indexing and storing the shape features, the grid, the cavities, the protrusions, and the protein model in a data store.

9. The method of claim 6, wherein identifying further comprises determining a distance visibility from a given point within the grid to a stopping point, and wherein the stopping point identifies the start of one of the cavities or one of the protrusions.

10. The method of claim 6, wherein determining further comprises inverting the grid to identify the protrusions as a number of the cavities.

11. The method of claim 6 further comprising, grouping features of the protein module into similar types of structures which are identifiable from the grid.

12. The method of claim of claim 6 further comprising, ranking the cavities, wherein higher ranking cavities are more likely to be active sites within the given protein, and wherein the ranking is based on a size associated with each of the cavities.

13. A method, comprising:

acquiring features of a given protein from a data store;

using the features to present a three-dimensional graphical representation of the given protein on a display; and

interactively receiving instructions via an interface to zoom in or out or pan around to selective portions of the presented protein.

14. The method of claim 13 further comprising:

receiving a selected portion of the given protein as a search term;

querying a data store on the selected portion; and

returning candidate matches from the data store, which substantially conform to the search term.

15. The method of claim 14 further comprising:

clustering the candidate matches into one or more groupings; and

permitting each group to be browsed via the interface.

16. The method of claim 13 further comprising:

receiving text-based information as an additional search term; and

modifying the search term with the additional search term to perform a different query against the data store.

17. A method, comprising:

receiving a graphical search term for a portion of a given protein or for a drug molecule;

reducing the graphical search term into features associated with a vector;

searching a data store with the features to predict portions of previously indexed proteins or previously indexed drug molecules that bind with the portion of the given protein or the drug molecule; and

returning results for predicted binding proteins or predicted binding drug molecules.

18. The method of claim 17 further comprising, permitting the features of the graphical search term to be manually modified and re-performing the searching and the returning of the results.

19. The method of claim 17 further comprising, representing the results as percentages, wherein each percentage corresponds to a predictive value indicating to what degree a returned predicted binding protein or returned predicted binding drug will bind with the portion or with the drug molecule used with the graphical search term.

20. The method of claim 19 further comprising, ranking the results in response to the percentages.

21. A method, comprising:

mining a data store for similar portions of different proteins or different drug molecules;

clustering similar groupings found during mining into hierarchical arrangements; and

presenting a hierarchy associated with the similar groupings for browsing.

22. The method of claim 21 further comprising, performing targeted searches on one or more select ones of the similar groupings on a received graphical representation or a portion of a given protein or drug molecule.

23. The method of claim 21 further comprising:

presenting a selected similar grouping within a display; and

interactively zooming in or out or panning on selected portions of the similar grouping.

24. A method, comprising:

generating a node histogram for an active site or a binding site or a given protein;

determining distances between the active site or the binding site to other different active sites or binding sites; and

creating a model characterization for the active site or the binding site from the node histogram and the distances.

25. The method of claim 24 further comprising, substantially matching the model characterization to one or more different active sites or binding sites or a drug molecule.

26. The method of claim 24 further comprising, indexing and storing the model characterization for subsequent search and retrieval.

27. The method of claim 24 further comprising, using the model characterization as a search term for a search query to find a potential active site or a potential binding site that substantially matches the active site or the binding site.

28. A system, comprising:

one or more data stores to house feature'"'"'s of proteins and drug molecules;

a data abstraction service to graphically represent and present the features, the proteins, and the drug molecules;

one or more tools to mine, search, query, and predict reactions and interactions of the features, the proteins, and the drug molecules and to receive candidate features, candidate proteins, and the drug molecules for comparison and prediction against the one or more data stores; and

an interface to permit the tools to be accessed and the data abstraction service to be accessed in response to directives received through the interface.

29. The system of claim 28, wherein the one or more tools include a query service to query the data store, a mining service to analyze and mine the data store, a prediction or docking service to predict the reactions and interactions, and population service to populate the one or more data stores with the features, the proteins, and the drug molecules.

30. The system of claim 28, wherein at least a portion of the features are acquired from an external repository.

31. The system of claim 28, wherein at least one of the tools is a service that dynamically creates the candidate features from a graphical search term supplied via the interface.

32. The system of claim 28, wherein the interface issues searches and queries via select ones of the tools and receives results, which are presented on a display.

33. The system of claim 28, wherein the searches and queries are dynamically identified by selecting portions of the features presented on the display by the data abstraction service.

34. A system, comprising:

grid mapping service;

a feature extraction service, wherein the grid mapping service is to map a graphical portion or drug molecule into a grid of coordinate points, and wherein the feature extraction service is to identify features of the coordinate points and linkages between features, and wherein a number of the features include active and binding sites for the graphical protein or drug molecule.

35. The system of claim 34 further comprising, a visualization service to present the features and the linkages graphically on a display.

36. The system of claim 34 further comprising, a clustering service to dynamically cluster the features with other previously identified features associated with other proteins or other drug molecules.

37. The system of claim 34 further comprising, a search and query service to search and query a data store using one or more of the extracted features.

38. The system of claim 34 further comprising, a scoring service to score each of the linkages identified with the features.

39. The system of claim 34 further comprising, a prediction service to mine a data store for other portions of proteins or other portions of drug molecules that bind to one or more of the features.

40. A data store implemented in a machine-accessible medium for storage and retrieval of features associated with proteins and drug molecules, the data store comprising:

a protein or drug molecule identifier;

a plurality of coordinate points associated with various portions of each protein or drug molecule identifier;

neighboring information associated with each coordinate point;

a maximum span associated with each coordinate point; and

normal features associated with each coordinate point;

wherein the information associated with each drug molecule is indexed and accessible collectively or individually from the data store.

41. The data store of claim 40, wherein the neighboring information includes link attributes for coordinate points that are in proximity to one another within a configurable threshold distance.

42. The data store of claim 40 further comprising, presentation information to define presentation attributes of the neighboring information, the maximum span, the normal features, and the coordinate points.

43. The data store of claim 40, wherein a number of the coordinate points associated with each protein or drug molecule identifier is identified as receptor sites or ligands of a given protein or drug molecule.

44. The data store of claim 43, wherein the receptor sites or the ligands further include scores that identify the likelihood that they are legitimate receptor sites or legitimate ligands.

45. The data store of claim 40, wherein the information associated with each protein or drug molecule is provided to a data abstraction service and used to graphically present a protein or drug molecule in three dimensions.