US5526281A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Machine-learning approach to modeling biological activity for molecular design and to modeling other characteristics
Assignee: Nexus Biosystems Inc
•
Filing: 1994-10-28
•
Granted: 1996-06-11
•
Expires: 2014-10-28
Abstract
Abstract Explicit representation of molecular shape of molecules is combined with neural network learning methods to provide models with high predictive ability that generalize to different chemical classes where structurally diverse molecules exhibiting similar surface characteristics are treated as similar. A new machine-learning methodology that can accept multiple representations of objects and construct models that predict characteristics of those objects. An extension of this methodology can be applied in cases where the representations of the objects are determined by a set of adjustable parameters. An iterative process applies intermediate models to generate new representations of the objects by adjusting said parameters and repeatedly retrains the models to obtain better predictive models. This method can be applied to molecules because each molecule can have many orientations and conformations (representations) that are determined by a set of translation, rotation and torsion angle parameters.
Inventors
- 👤 David Chapman
- 👤 Roger Critchlow
- 👤 Ajay N. Jain
- 👤 Rick Lathrop
- 👤 Tomas L. Perez
- 👤 Tom Dietterich
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-11-04 01:08
CPC Classifications
C; C07; C07K; C07K1/00; G; G16; G16C; G16C20/00; G16C20/50; G16C20/30; G16C20/70; Y; Y10; Y10S; Y10S706/00; Y10S706/902; Y10S706/919; Y10S706/92
Claims Analysis
Total Claims: 54
Claim 1
Independent Claim
318 words
A method in a computer for predicting activity of molecules with respect to a chemical function based on known activities of a plurality of molecules, said molecules each having one or more conformations and orientations, each combination of a conformation and an orientation defining a pose of a molecule, said method comprising: (a) operating a computer to define one or more poses of each molecule as the initial poses of an initial training set; (b) operating a computer to construct a model in said computer with parameters for predicting activity of poses with respect to said chemical function and setting model parameter values; (c) operating a computer to predict the activities of at least some of said initial poses in the initial training set using the model and said model parameter values and comparing the predicted activities of initial poses of molecules to results of a set of actual tests for such activities of such molecules; (d) modifying said model parameter values in a computer based on a prior comparison between predicted activities of poses in the initial or an updated training set and said results of a set of actual tests for such activities to minimize the differences between the predicted activities of said at least some of the poses of molecules in such set and said results of a set of actual tests for such activities of such molecules, and modifying or selecting poses of the molecules to obtain an updated training set of enhanced poses with higher predictive value than poses prior to the modifying step; (e) operating said computer to use the model and modified model parameters values to predict the activity of a molecule not in the training set, and having unknown activity; and (f) visually displaying using computer graphics, in response to said conditionally modified model, a representation of a second new molecule not in the training set, and having unknown activity.
Claim 2
Independent Claim
318 words
A method in a computer for predicting activity of molecules with respect to a chemical function based on known activities of a plurality of molecules, said molecules each having one or more conformations and orientations, each combination of a conformation and an orientation defining a pose of a molecule, said method comprising: (a) operating a computer to define one or more poses of each molecule as the initial poses of an initial training set; (b) operating a computer to construct a model in said computer with parameters for predicting activity of poses with respect to said chemical function and setting model parameter values; (c) operating a computer to predict the activities of at least some of said initial poses in the initial training set using the model and said model parameter values and comparing the predicted activities of initial poses of molecules to results of a set of actual tests for such activities of such molecules; (d) modifying said model parameter values in a computer based on a prior comparison between predicted activities of poses in the initial or an updated training set and said results of a set of actual tests for such activities to minimize the differences between the predicted activities of said at least some of the poses of molecules in such set and said results of a set of actual tests for such activities of such molecules, and modifying or selecting poses of the molecules to obtain an updated training set of enhanced poses with higher predictive value than poses prior to the modifying step; (e) operating said computer to use the model and modified model parameters values to predict the activity of a molecule not in the training set, and having unknown activity; and (f) visually displaying using computer graphics, in response to said conditionally modified model, a representation of a second new molecule not in the training set, and having unknown activity.
Claim 3
Dependent Claim
79 words
The method of claim 1, further comprising: (f) predicting the activities of at least some of said enhanced poses in the updated set using the modified parameter values and comparing the predicted activities of enhanced poses of molecules to the known activities of such molecules; and (g) repeating steps (d) and (f) prior to step (e), wherein step (d) is repeated based on a prior comparison between predicted activities of poses in the updated set and their known activities.
Claim 4
Dependent Claim
26 words
The method of claim 2, wherein step (g) is carried out until there is no substantial change in the model parameter values and the enhanced poses.
Claim 5
Dependent Claim
21 words
The method of claim 1, wherein said modifying step employs gradient descent in modifying the poses and the model parameter values.
Claim 6
Dependent Claim
70 words
The method of claim 1, wherein said modifying step in (d) first iteratively modifies the model parameter values until the differences between the predicted activities of said at least some of the enhanced poses of molecules and the known activities of such molecules are minimized to arrive at a set of modified model parameter values and then iteratively modifies the poses to maximize their activities and to obtain enhanced poses.
Claim 7
Dependent Claim
82 words
The method of claim 5, wherein each time after poses of molecules in the training set have been modified, said modifying step in (d) iteratively modifies the model parameter values until the differences between the predicted activities of said at least some of the enhanced poses of molecules and the known activities of such molecules are minimized to obtain a set of modified parameter values, so that any pose modification thereafter will be in accordance with said set of modified parameter values.
Claim 8
Dependent Claim
95 words
The method of claim 1, wherein the pose of each molecule in the set that has higher predicted activity than other poses in the set of the same molecule defines the best pose of such molecule, and wherein said model parameter values are modified in step (d) based on a prior comparison between predicted activities of the best poses in the set and their known activities to minimize the differences between the predicted activities of said at least some of the best poses of molecules in the set and the known activities of such molecules.
Claim 9
Dependent Claim
44 words
The method of claim 1, said model constructing step including extracting a set of feature values from each of said initial poses related to said activity and setting an initial value for each of the features to be some of the model parameter values.
Claim 10
Dependent Claim
50 words
The method of claim 8, said model constructing step further including setting initial mean and standard deviations of a feature value and a Gaussian-like function representing a contribution to predicted activity of a pose as a function of said feature value in relation to its initial mean and standard deviations.
Claim 11
Dependent Claim
23 words
The method of claim 9, wherein said model constructing step further includes setting a positive or negative weighting factor for said Gaussian function.
Claim 12
Dependent Claim
73 words
The method of claim 1, wherein an error function is defined for each pose, said function being a difference between the predicted activity of such pose of a molecule and the known activity of such molecule, wherein said modifying step includes deriving a total error function indicating the sum of the individual error functions of each of some of the poses and changing the model parameter values to minimize the total error function.
Claim 13
Dependent Claim
18 words
The method of claim 11, wherein said modifying step employs gradient based steps to minimize the error function.
Claim 14
Dependent Claim
25 words
The method of claim 1, wherein said pose modifying step in step (d) modifies the poses as functions of parameters including orientation and conformation parameters.
Claim 15
Dependent Claim
26 words
The method of claim 13, wherein said functions are differentiable and said pose modifying step includes differentiating the functions with respect to orientation and conformation parameters.
Claim 16
Dependent Claim
79 words
The method of claim 1, further comprising setting a set of ordered numerical values according to a preset order to represent the known activities of said plurality of molecules, wherein said modifying step (d) also includes adjusting the set of ordered numerical values while retaining the preset order to reduce the differences between the predicted activities of said at least some poses of molecules in the initial or an updated training set and the known activities of such molecules.
Claim 17
Dependent Claim
34 words
The method of claim 1, further comprising, prior to the selecting step, searching for conformers of the molecules in the training set and aligning the conformers relative to one another to form possible poses.
Claim 18
Dependent Claim
34 words
The method of claim 1, wherein said using step (e) also indicates which conformer of said molecule not in the training set would have the highest predicted activity with respect to said chemical function.
Claim 19
Dependent Claim
34 words
The method of claim 1, wherein said using step also indicates which properties of said molecule not in the training set would have effects on its predicted activity with respect to said chemical function.
Claim 20
Dependent Claim
21 words
The method of claim 1, further comprising visually displaying relationship between poses of molecules and model parameter values using computer graphics.
Claim 21
Dependent Claim
26 words
The method of claim 19, further comprising modifying said poses with respect to the model parameter values displayed to modify the predicted activities of the poses.
Claim 22
Dependent Claim
22 words
The method of claim 1, wherein said using step includes searching a database of molecules with unknown activities and predicting their activities.
Claim 23
Dependent Claim
38 words
The method of claim 1, wherein said model constructing step includes setting a sigmoid function representing a contribution to predicted activity of a pose as a sum of the weighted Gaussians of one or more individual feature values.
Claim 24
Dependent Claim
32 words
The method of claim 22, wherein said model constructing step further includes setting another sigmoid function representing the overall predicted activity of a pose as a weighted sum of the sigmoid functions.
Claim 25
Independent Claim
182 words
A method of operating a computer for predicting activity of molecules with respect to a chemical function based on known activities of a training set of molecules, said molecules in the set each having one or more conformations and orientations, each combination of a conformation and an orientation defining a pose of a molecule, said method comprising: operating a computer to extract a set of feature values from each a set of results from actual tests performed for each of said poses of molecules in the training set, said feature values related to said activity, said extracting step including the following steps: (a) creating a surface representation in said computer of each of the poses of molecules in the training set; and (b) obtaining a feature value between at least one sampling point and a point on said surface representation in said computer of each of the poses constructing a model for predicting activity of poses with respect to said chemical function using said feature values and using the model to predict the activity of a molecule not in the training set.
Claim 26
Independent Claim
182 words
A method of operating a computer for predicting activity of molecules with respect to a chemical function based on known activities of a training set of molecules, said molecules in the set each having one or more conformations and orientations, each combination of a conformation and an orientation defining a pose of a molecule, said method comprising: operating a computer to extract a set of feature values from each a set of results from actual tests performed for each of said poses of molecules in the training set, said feature values related to said activity, said extracting step including the following steps: (a) creating a surface representation in said computer of each of the poses of molecules in the training set; and (b) obtaining a feature value between at least one sampling point and a point on said surface representation in said computer of each of the poses constructing a model for predicting activity of poses with respect to said chemical function using said feature values and using the model to predict the activity of a molecule not in the training set.
Claim 27
Dependent Claim
29 words
The method of claim 24, wherein said creating step creates said representations by finding van der Waals surface representations of atoms on the surface of each of the poses.
Claim 28
Dependent Claim
93 words
The method of claim 25, wherein said finding step includes: finding a first atom of the molecule with its center at a minimum distance to a sampling point using a squared distance function; finding a set of atoms in the vicinity of the first atom whose van der Waals radii are larger than that of the first atom, if any; and determining from a square root function which of the atoms in the set has a van der Waals surface closer to the sampling point than that of the first atom, if any.
Claim 29
Dependent Claim
27 words
The method of claim 24, wherein said creating step creates said representations by summing Gaussian surface representations of atoms on the surface of each of the poses.
Claim 30
Dependent Claim
25 words
The method of claim 24, said obtaining step including determining the distances between said at least one point and the surface representations along predetermined directions.
Claim 31
Dependent Claim
26 words
The method of claim 24, said obtaining step including: determining the minimum distance between said at least one point and the surface representations of the poses.
Claim 32
Dependent Claim
33 words
The method of claim 29, further comprising: selecting a plurality of points around said surface representations; and determining the minimum distance between each of said points and the surface representations of the poses.
Claim 33
Dependent Claim
32 words
The method of claim 30, wherein said points selecting step includes: forming an average surface representation of substantially all the poses; and selecting a plurality of points around said average surface representation.
Claim 34
Dependent Claim
14 words
The method of claim 24, said feature values including a steric or electrostatic value.
Claim 35
Dependent Claim
20 words
The method of claim 24, further comprising visually displaying relationship between poses of molecules and said model using computer graphics.
Claim 36
Dependent Claim
24 words
The method of claim 33, further comprising modifying said poses with respect to the model displayed to modify the predicted activities of the poses.
Claim 37
Independent Claim
294 words
A method in a computer for arriving at a model for predicting activity of molecules with respect to a chemical function based on known activities of a training set of molecules, said molecules in the set each having one or more conformations and orientations, each combination of a conformation and an orientation defining a pose of a molecule, said method comprising: (a) operating a computer to define one or more poses from possible poses of each molecule as the initial poses of a training set; (b) operating a computer to construct a model in said computer with parameters for predicting activity of poses with respect to said chemical function and setting model parameter values; (c) operating a computer to predict the activities of at least some of said initial poses in the training set using the model and said model parameter values and comparing the predicted activities of initial poses of molecules to a set of results from actual tests performed for activities of such molecules; (d) modifying said model parameter values in said computer based on a prior comparison between predicted activities of poses in the set and said set of results from actual tests performed for activities to minimize the differences between the predicted activities of said at least some of the poses of molecules in the set and said set of results from actual tests performed for activities of such molecules, and also modifying or selecting poses of the molecules to obtain an updated training set of enhanced poses with greater predicted activities than poses in the set prior to the modifying step; and (e) visually displaying using computer graphics, in response to said modified model, a representation of a new molecule not in the training set, and having unknown activity.
Claim 38
Independent Claim
294 words
A method in a computer for arriving at a model for predicting activity of molecules with respect to a chemical function based on known activities of a training set of molecules, said molecules in the set each having one or more conformations and orientations, each combination of a conformation and an orientation defining a pose of a molecule, said method comprising: (a) operating a computer to define one or more poses from possible poses of each molecule as the initial poses of a training set; (b) operating a computer to construct a model in said computer with parameters for predicting activity of poses with respect to said chemical function and setting model parameter values; (c) operating a computer to predict the activities of at least some of said initial poses in the training set using the model and said model parameter values and comparing the predicted activities of initial poses of molecules to a set of results from actual tests performed for activities of such molecules; (d) modifying said model parameter values in said computer based on a prior comparison between predicted activities of poses in the set and said set of results from actual tests performed for activities to minimize the differences between the predicted activities of said at least some of the poses of molecules in the set and said set of results from actual tests performed for activities of such molecules, and also modifying or selecting poses of the molecules to obtain an updated training set of enhanced poses with greater predicted activities than poses in the set prior to the modifying step; and (e) visually displaying using computer graphics, in response to said modified model, a representation of a new molecule not in the training set, and having unknown activity.
Claim 39
Independent Claim
252 words
A method of operating a computer for predicting characteristics of an object based on known characteristics of a plurality of other objects, said other objects each having one or more representations, said method comprising: (a) operating a computer to define one or more representations from possible representations of each of said other objects as the initial representations; (b) operating a computer to construct a model in said computer for predicting characteristics of the representations; (c) operating a computer to predict the characteristics of at least some of said initial representations or an updated set of representations using the model and comparing the predicted characteristics of initial representations of said other objects to a set of results from actual tests performed for characteristics of such other objects, wherein for each of said other objects, the representation that has better characteristics than other representations of the same object defines the best representation of such object; (d) modifying said model in said computer based on a prior comparison between predicted characteristics of the best representations of said other objects and said set of results from actual tests performed for characteristics to minimize the differences between the predicted characteristics of said best representations of the other objects and said set of results from actual tests performed for characteristics of such objects; (e) operating said computer to use the modified model to predict the characteristics of an object, said object not being one of said other objects for which possible representations were chosen as the initial representations.
Claim 40
Independent Claim
252 words
A method of operating a computer for predicting characteristics of an object based on known characteristics of a plurality of other objects, said other objects each having one or more representations, said method comprising: (a) operating a computer to define one or more representations from possible representations of each of said other objects as the initial representations; (b) operating a computer to construct a model in said computer for predicting characteristics of the representations; (c) operating a computer to predict the characteristics of at least some of said initial representations or an updated set of representations using the model and comparing the predicted characteristics of initial representations of said other objects to a set of results from actual tests performed for characteristics of such other objects, wherein for each of said other objects, the representation that has better characteristics than other representations of the same object defines the best representation of such object; (d) modifying said model in said computer based on a prior comparison between predicted characteristics of the best representations of said other objects and said set of results from actual tests performed for characteristics to minimize the differences between the predicted characteristics of said best representations of the other objects and said set of results from actual tests performed for characteristics of such objects; (e) operating said computer to use the modified model to predict the characteristics of an object, said object not being one of said other objects for which possible representations were chosen as the initial representations.
Claim 41
Dependent Claim
47 words
The method of claim 36, further comprising: (f) repeating steps (c) and (d) prior to step (e), wherein step (c) is repeated based on a prior comparison between predicted characteristics of at least some of the representations of each object and the known characteristics of those objects.
Claim 42
Independent Claim
323 words
A method of operating a computer for predicting characteristics of an object based on known characteristics of a plurality of other objects, said other objects each having many representations, each representation called a pose and being defined by the object and values of one or more parameters defining pose parameters, said method comprising: operating a computer to define one or more poses from the possible representations of each of said other objects as the initial poses; (b) operating a computer to construct a model in said computer for predicting characteristics of the objects from one or more poses of the objects; (c) operating a computer to predict characteristics of at least some of said initial poses or enhanced poses using the model and comparing the predicted characteristics of the initial poses of said other objects to a set of results from actual tests performed for characteristics of such other objects, wherein for each of said other objects the pose that has better predicted characteristics than other poses of the same object defines the best pose of such object; (d) modifying said model in said computer based on a prior comparison between predicted characteristics of the best poses of said other objects and said set of results from actual tests performed for characteristics to minimize the differences between the predicted characteristics of said best poses of the other objects and said set of results from actual tests performed for characteristics of such objects; (e) operating a computer to define an updated set of enhanced poses by computing new pose parameters for each object such that the resulting pose is predicted by said model to have improved characteristics; (f) repeating steps (c), (d), and (e) one or more times; and (g) operating a computer to apply the modified model to predict the characteristics of an object, said object not being one of said other objects for which possible representations were chosen as the initial poses.
Claim 43
Independent Claim
323 words
A method of operating a computer for predicting characteristics of an object based on known characteristics of a plurality of other objects, said other objects each having many representations, each representation called a pose and being defined by the object and values of one or more parameters defining pose parameters, said method comprising: operating a computer to define one or more poses from the possible representations of each of said other objects as the initial poses; (b) operating a computer to construct a model in said computer for predicting characteristics of the objects from one or more poses of the objects; (c) operating a computer to predict characteristics of at least some of said initial poses or enhanced poses using the model and comparing the predicted characteristics of the initial poses of said other objects to a set of results from actual tests performed for characteristics of such other objects, wherein for each of said other objects the pose that has better predicted characteristics than other poses of the same object defines the best pose of such object; (d) modifying said model in said computer based on a prior comparison between predicted characteristics of the best poses of said other objects and said set of results from actual tests performed for characteristics to minimize the differences between the predicted characteristics of said best poses of the other objects and said set of results from actual tests performed for characteristics of such objects; (e) operating a computer to define an updated set of enhanced poses by computing new pose parameters for each object such that the resulting pose is predicted by said model to have improved characteristics; (f) repeating steps (c), (d), and (e) one or more times; and (g) operating a computer to apply the modified model to predict the characteristics of an object, said object not being one of said other objects for which possible representations were chosen as the initial poses.
Claim 44
Independent Claim
324 words
A method for predicting activity of molecules with respect to a chemical function based on activities of a plurality of molecules, said method comprising operating a computer to define a plurality of molecules, each one of said molecules having at least one pose; operating a computer to define a training set having at least one pose for each one of said molecules; operating a computer to construct a model in a memory for predicting a result of an assay for a measure of activity relating to a set of desired physical properties for said molecule; operating a computer using said model to produce a prediction for a first pose of a first molecule in said training set; operating a computer using said model to produce a prediction for a second pose of said first molecule; conditionally modifying said model in said memory in response to a difference between said prediction and a result of an actual said assay conducted for said first pose; conditionally modifying said training set in said memory to replace said first pose with said second pose in response to a difference between said prediction for said first pose and said prediction for said second pose; repeating said steps of operating a computer using said model for said first and second poses, and conditionally modifying said model and said training set in memory, until a predetermined condition is reached; operating a computer using said model to produce a prediction for a pose of a new molecule, said new molecule not being in said training set; conducting an actual said assay for said new molecule; comparing said prediction for said pose of said new molecule with a result of said actual assay for said new molecule; and repeating said steps of operating a computer using said model for said pose of said new molecule, conducting an actual said assay for said new molecule, and said prediction, until a predetermined condition is reached.
Claim 45
Independent Claim
324 words
A method for predicting activity of molecules with respect to a chemical function based on activities of a plurality of molecules, said method comprising operating a computer to define a plurality of molecules, each one of said molecules having at least one pose; operating a computer to define a training set having at least one pose for each one of said molecules; operating a computer to construct a model in a memory for predicting a result of an assay for a measure of activity relating to a set of desired physical properties for said molecule; operating a computer using said model to produce a prediction for a first pose of a first molecule in said training set; operating a computer using said model to produce a prediction for a second pose of said first molecule; conditionally modifying said model in said memory in response to a difference between said prediction and a result of an actual said assay conducted for said first pose; conditionally modifying said training set in said memory to replace said first pose with said second pose in response to a difference between said prediction for said first pose and said prediction for said second pose; repeating said steps of operating a computer using said model for said first and second poses, and conditionally modifying said model and said training set in memory, until a predetermined condition is reached; operating a computer using said model to produce a prediction for a pose of a new molecule, said new molecule not being in said training set; conducting an actual said assay for said new molecule; comparing said prediction for said pose of said new molecule with a result of said actual assay for said new molecule; and repeating said steps of operating a computer using said model for said pose of said new molecule, conducting an actual said assay for said new molecule, and said prediction, until a predetermined condition is reached.
Claim 46
Dependent Claim
37 words
A method as in claim 39, comprising conditionally modifying said model in response to a difference between said prediction for said pose of said new molecule and said result of said actual assay for said new molecule.
Claim 47
Dependent Claim
38 words
A method as in claim 39, comprising conditionally modifying said training set in response to a difference between said prediction for said pose of said new molecule and said result of said actual assay for said new molecule.
Claim 48
Dependent Claim
156 words
A method as in claim 39, comprising operating said model for at least one pose for each one of a plurality of new molecules, said plurality of new molecules not being in said training set, to produce a prediction for each one of said plurality of new molecules; selecting a choice one of said plurality of new molecules in response to said prediction for each one of said plurality of new molecules; conducting an actual said assay for said choice one of said plurality of new molecules; comparing a result of said actual assay for said choice one of said plurality of new molecules with at least a part of said prediction for each one said plurality of new molecules; and repeating said steps of selecting a choice one, conducting an actual said assay for said choice one, and comparing a result of said actual assay for said choice one, until a predetermined condition is reached.
Claim 49
Dependent Claim
46 words
A method as in claim 42, comprising conditionally modifying said model in response to a difference between said prediction for said choice one of said plurality of new molecules and said result of said actual assay for said choice one of said plurality of new molecules.
Claim 50
Dependent Claim
47 words
A method as in claim 42, comprising conditionally modifying said training set in response to a difference between said prediction for said choice one of said plurality of new molecules and said result of said actual assay for said choice one of said plurality of new molecules.
Claim 51
Dependent Claim
122 words
A method as in claim 39, wherein said model comprises a neural network, said neural network comprising a plurality of layers each having at least one node; said plurality of layers including a first layer having at least one node coupled to an input value; a second layer having at least one node coupled to a plurality of nodes of said first layer; said first layer having at least one node with a first transfer function, said first transfer function comprising a Gaussian function with parameters m, s, and z, said parameters being modifiable; and said second layer having at least one node with a second transfer function, said second transfer function comprising a sigmoid function with parameters, said parameters being modifiable.
Claim 52
Dependent Claim
17 words
A method as in claim 39, wherein said new molecule is designed in response to said model.
Claim 53
Dependent Claim
38 words
A method as in claim 39, wherein said new molecule is a catalyst, a carbohydrate, a coating agent, a cosmetic, an explosive, an industrial chemical, a paint, a perfuming agent, a petroleum derivative, a polymer, or a polypeptide.
Claim 54
Dependent Claim
13 words
A method as in claim 39, wherein said new molecule is a pharmaceutical.
References Cited (Backward Citations)
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 1999-06-11 | 1999-12-11 | 2000-06-11 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2003-06-11 | 2003-12-11 | 2004-06-11 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2007-06-11 | 2007-12-11 | 2008-06-11 | $7700.00 | Window Closed / Lapsed |