DETAILED ACTION
Notice of Pre-AIA or AIA Status
1. The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claim status
2. Claims 1-16 are currently pending and under exam herein.
Claims 1-16 are rejected.
Claim 4 is objected to.
Priority
3. This application is a 371 of International Application No. PCT/EP2022/062389 which designated the United States and was filed 09 May 2022. Claimed benefit of foreign priority to the prior European Patent Application No. EP21173067.6, filed on 10 May 2021, is acknowledged. A certified copy of the priority document has been received. In this action, all claims are examined as though they had an effective filing date 10 May 2021. In future actions, the effective filing date of one or more claims may change, due to amendments to the claims, or further analysis of the disclosure(s) of the priority application(s).
Information Disclosure Statement
4. The information disclosure statement (IDS) submitted on 09 November 2023 is being considered by the examiner.
Drawings
5. The drawing submitted on 09 November 2023 are accepted by the examiner.
Claim Objections
6. Claim 4 is objected to because of the following informalities:
Claim 4 recites: “The method according to claim 1, wherein as at least one excipient from the list which changes the viscosity of the new formulation (8) the most sufficient a combination of two or more excipients from the list is used”, which is grammatically incorrect. A possible correction is: “The method according to claim1, wherein of the at least one excipient from the list which changes the viscosity of the new formulation (8), the most sufficient is wherein a combination of two or more excipients from the list is used.” For the purposes of examination, the claim will be interpreted to be the suggested version.
Appropriate correction is required.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
7. Claims 6-9 and 15-16 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claims 15 and 16 recite “wherein the Gaussian process is replaced”. This expression lacks antecedent basis because “a Gaussian process” has not been introduced in the claim or in the preceding claim family (i.e. claim 14) or in the referenced claim 5. For the purpose of examination, ‘the Gaussian process’ will be interpreted to mean ‘a Gaussian process’. It should also be noted that if ‘the Gaussian process’ had antecedent basis, claims 15 and 16 may instead be rejected as improper dependent claims under 35 U.S.C 112(d) for failing to include all the limitation of the claim upon which they depend (see MPEM 7.36.01).
Claims 7 and 8 recite “the formulation specialist” but this expression lacks antecedent basis because “a formulation specialist” has not been introduced in the claim or in the preceding claim family. Claim 9 is also rejected based on its dependency on claim 8 and for failing to resolve the indefiniteness issue.
Claim 8 recites “proposing the new set of excipients (2) or the combination thereof to the formulation specialist (9), who then conducts the respective experiments (10) in the lab to determine the resulting viscosities”. Similarly, claim 5 recites “proposed to a formulation specialist (9), who conducts the at least one respective experiment (10) in a lab to validate the predicted viscosities (3)”. It is unclear if the conducting respective experiments is a required part of the invention. With broadest reasonable interpretation and for the purpose of examination, the conducting the respective experiments step will not be considered to be a necessary step of the invention. Claim 9 is also rejected based on its dependency on claim 8 and for failing to resolve the indefiniteness issue. Claims 14-16 are also rejected based on their dependency on claim 5 and for failing to resolve the indefiniteness issue.
Claim 6 recites “wherein the Machine Learning Model (5) is created and trained”. It is unclear if the training of the Machine Learning Model is a required part of the invention. With broadest reasonable interpretation and for the purpose of examination, the training will not be considered to be a necessary step of the invention. Claims 7-9 are also rejected based on its dependency on claim 6 and for failing to resolve the indefiniteness issue.
Claim 7 recites “wherein the viscosity values of a given formulation (8) are modelled via the Machine Learning Model (5)”. It is unclear if the modelling via the Machine Learning Model is a required part of the invention. With broadest reasonable interpretation and for the purpose of examination, the modelling will not be considered to be a necessary step of the invention.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
8. Claims 14-16 are rejected under 35 U.S.C. 101 because the claimed invention is directed to non-statutory subject matter.
In accordance with MPEP § 2106, claims are analyzed to determine if they recite statutory subject matter. In the instant application, the following claims do not recite statutory subject matter:
Claim 14 recites: A Machine Learning Model performed on a computer which is created and trained according to claims 5
Claim 15 recites: The Machine Learning Model according to claim 14, wherein the Gaussian Process is replaced with any other model architecture fulfilling the same purpose, in particular other types of stochastic processes, deep Bayesian networks, generalized linear models, neural networks, support vector machines, tree based models, ensemble models, etc.
Claim 16 recites: The Machine Learning Model according to claim 14, wherein the Gaussian Process is replaced with another type of stochastic process, a deep Bayesian network, a generalized linear model, a neural network, a support vector machine, a tree-based model, or an ensemble model.
The claimed subject matter of claims 14-16 do not fall within at least one of the four categories of patent eligible subject matter because they are not directed to a method, a manufacture, a composition of matter or a process. Claims 14-16 are directed to a machine learning model, which is software per se because there are no methods stated. The limitations directed to how the model is generated and trained (claim 14) and the type of algorithm used in the model (claims 15 and 16) simply further define the model but do not change its position as software per se and its categorization as non-statutory subject matter.
Therefore, claims 14-16 are rejected for claiming non-statutory subject matter (STEP 1: NO).
9. Claims 1-13 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 2A, Prong 1
In accordance with MPEP § 2106, claims found to recite statutory subject matter (Step 1: YES) are then analyzed to determine if the claims recite any concepts that equate to an abstract idea, law of nature or natural phenomenon (Step 2A, Prong 1). In the instant application, the claims recite the following limitations that equate to an abstract idea:
Claim 1 recites: Generating representations of at least one excipient (2) from a list of excipients by the computer (6) via In-Silico-simulations
Claim 1 recites: to recognize patterns in the data set (1)
Claim 1 recites: to evaluate the viscosity changing effect of at least one viscosity changing excipient (2) chosen from the list of excipients to a new formulation (8) containing at least one unknown protein (11) and the at least one viscosity changing excipient (2)
Claim 1 recites: applying the recognized patterns on provided data of the at least one unknown protein (11)
Claim 1 recites: Selecting, depending on the evaluation result, the at least one excipient from the list according to an acquisition criterion and applying it to the unknown protein (11), wherein the provided data of the at least one unknown protein (11) are data describing the viscosity of a protein composition containing the at least one unknown protein (11) and optionally with at least one viscosity changing excipient (2)
Claim 2 recites: the method according to claim 1, wherein the evaluation of the viscosity changing effect of at least one viscosity changing excipient (2) is done by predicting the viscosity (3) of the new formulation (8) containing at least one unknown protein (11) and at least one viscosity changing excipient (2)
Claim 4 recites: The method according to claim 1, wherein as at least one excipient from the list which changes the viscosity of the new formulation (8) the most sufficient a combination of two or more excipients from the list is used.
Claim 5 recites: The method according to claim 1, wherein at least one specific experimental measurement (10) is proposed to a formulation specialist (9), who conducts the at least one respective experiment (10) in a lab to validate the predicted viscosities (3) and trains the Machine Learning Model (5) with the validated results by adding them to the provided data set (1) in the database via the computer (6).
Claim 6 recites: The method according to claim 1, wherein the Machine Learning Model (5) is created and trained by combining the data set (1) describing the viscosity of at least one prototypical protein formulation (8) with the representations of the at least one viscosity changing excipient (2) or a combination thereof.
Claim 7 recites: The method according to claim 6, wherein the viscosity values of a given formulation (8) are modelled via the Machine Learning Model (5) in the form of a Gaussian process and the model predictions (3) are used to guide the formulation specialist (9) by means of a Bayesian optimal experimental design.
Claim 8 recites: The method according to claim 6, wherein the training of the Machine Learning Model (5) on the computer (6) is done by performing at least once the following steps: a Optimizing the Machine Learning Model Parameters with training data from the data set (1) by maximizing the marginal likelihood of the training data
Claim 8 recites: Evaluating a posterior distribution of viscosity values for untested excipients (2) or a combination thereof based on the Machine Learning Model (5)
Claim 8 recites: predicting a viscosity (3)
Claim 8 recites: Selecting a new set of excipients (2) or a combination thereof
Claim 8 recites: optimizing an acquisition score obtained from the computed posterior distribution
Claim 8 recites: Proposing the new set of excipients (2) or the combination thereof to the formulation specialist (9), who then conducts the respective experiments (10) in the lab to determine the resulting viscosities
Claim 8 recites: Adding the obtained measurements (10) to the training data.
Claim 9 recites: The method according to claim 8, wherein the prediction of the viscosity (3) obtained from the posterior distribution of viscosity values (3) is based on a pH-dependent feature vector characterizing the excipients (2) used in the considered formulation (8) and on the used excipient concentration levels.
Claim 10 recites: The method according to claim 1 wherein the acquisition criterion is which viscosity changing excipient (2) reduces the viscosity the most
Claim 11 recites: The method according to claim 1, wherein the representations of excipients (2) are generated by the computer (6) in the form of physical parameters as well as molecular fingerprints.
Claim 12 recites: The method according to claim1, wherein the generated representations of excipients (2) are cross-validated experimentally
Claim 13 recites: The method according to claim 1, wherein the generated representations of excipients (2) include quantum mechanic features, optionally complemented with a set of topological molecular fingerprints
The limitations regarding ‘generating representations’ (which is interpreted to comprise generating string representations of excipients, which relies on mathematical graph-traversal rules and text-processing algorithms), ‘optimizing maximizing the marginal likelihood of the training data’ and ‘optimizing an acquisition score’, are verbal equivalents that describe a mathematical calculation that is performed as the limitation and are so simple that they could be performed in the human mind or with pen and paper. Therefore, these limitations fall under the "Mathematical concepts" and "Mental processes" groupings of abstract ideas.
The limitations regarding ‘recognize patterns in the dataset’, ‘selecting the at least one excipient’, ‘evaluate the viscosity changing effect’ and ‘applying the recognized patterns’ (which is claimed as part of the process of evaluating viscosity changing effects), ‘predicting the/a viscosity’, ‘evaluating a posterior distribution of viscosity values’, ‘selecting a new set of excipients’, ‘proposing the new set of excipients’ and ‘adding the obtained measurements to the training data’ under broadest reasonable interpretation encompasses copying values into a list, recognizing patterns, making selections, evaluating effects and making proposals and thus can be performed in the human mind because the human mind is capable of providing values and determining information from other values, and making decisions based on the values thus it falls into the “mental process” grouping of abstract ideas.
The limitations that further limit the criteria used to make the selection of excipients (claim 1), the number of excipients selected (claim 4), what the experimenter does with the proposed excipients (claim 5), how the machine learning model that is used was trained (in claim 6), the algorithm used in model and in the prediction (claim 7), what information the predictions of viscosity are based on (in claim 9), the acquisition criterion used to make the selection (claim 10), data or features represented in the representations (claims 11 and 13), whether the representations were cross-validated experimentally (claim 12) merely further limit the judicial exceptions but do not inherently change their positions as abstract ideas.
While claims 1-13 recite performing some aspects of the analysis with a computer, there are no additional limitations that indicate that this computer requires anything other than carrying out the recited mental process or mathematical concept in a generic computer environment. Merely reciting that a mental process is being performed in a generic computer environment does not preclude the steps from being performed practically in the human mind or with pen and paper as claimed. If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components, then if falls within the "Mental processes" grouping of abstract ideas.
As such, claims 1-13 recite an abstract idea (Step 2A, Prong 1: YES).
Step 2A, Prong 2
Claims found to recite a judicial exception under Step 2A, Prong 1 are then further analyzed to determine if the claims as a whole integrate the recited judicial exception into a practical application or not (Step 2A, Prong 2). This judicial exception is not integrated into a practical application because the claims do not recite an additional element that reflects an improvement to technology or applies or uses the recited judicial exception in some other meaningful way. Rather, the instant claims recite additional elements that amount to mere instructions to implement the abstract idea in a generic computing environment or insignificant extra-solution activity. Specifically, the claims recite the following additional elements:
Claim 1 recites: a method for selecting at least one viscosity changing excipient (2) for a formulation (8) containing at least one unknown protein (11) via a computer (6)
Claim 1 recites: comprising the following steps: Providing a data set (1) from a database that describes the viscosity of several known formulations containing at least one protein and optionally at least one viscosity changing excipient (2)
Claim 1 recites: Using a Machine Learning Model (5) executed on the computer (6) that uses the generated representations of at least one excipient (2)
Claim 3 recites: The method according to claim 1, wherein the data set (1) has been generated by experimental measurements (10)
Claim 3 recites: is stored in the database via the computer (6).
The additional element directed to ‘providing a dataset’ merely serves to gather data that is used an input for the judicial exception. The activities are incidental to the primary process and are merely a nominal or tangential addition to the claim. As set forth in MPEP 2106.05(g), mere data gathering activities are insignificant extra-solution activity that does not provide a practical application. The limitation of data types (i.e. the viscosity of several known formulations containing at least one protein and optionally at least one viscosity changing excipient) is an additional element, but it amounts to specifying a particular field-of-use, which does not provide a practical application (MPEP 2106.05(h)). The limitation ‘wherein the data set (1) has been generated by experimental measurements (10)’ merely describes how the provided dataset was generated but does not change the position of ‘providing a dataset’ as a data gathering activity.
Similarly storing data that is input to or output from the judicial exception, are pre-solution and post-solution "extra-solution activity" as they are activities that are incidental to the primary process or product that are merely a nominal or tangential addition to the claim (MPEP 2106.05(g)).
There are no limitations that indicate that the computer requires anything other than a generic computing system. As such, these limitations equate to mere instructions to implement the abstract idea on a generic computer that the courts have stated does not render an abstract idea eligible in Alice Corp., 573 U.S. at 223, 110 USPQ2d at 1983. See also 573 U.S. at 224, 110 USPQ2d at 1984.
Step 2B
Claims found to be directed to a judicial exception are then further evaluated to determine if the claims recite an inventive concept that provides significantly more than the judicial exception itself (Step 2B).
The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception because the claims recite additional elements that equate to mere instructions to apply the recited exception in a generic computing environment or well-understood, and conventional activity.
As discussed above, there are no additional limitations to indicate that the claimed computer requires anything other than generic computer components in order to carry out the recited abstract idea in the claims. Claims that amount to nothing more than an instruction to apply the abstract idea using a generic computer do not render an abstract idea eligible. Alice Corp., 573 U.S. at 223, 110 USPQ2d at 1983. See also 573 U.S. at 224, 110 USPQ2d at 1984.
As set forth in MPEP section 2106.05(g), the courts have decided that limitations that merely add an insignificant extra-solution activity, do not amount to an inventive concept, particularly when the activities are well-understood and conventional. Parker v. Flook, 437 U.S. 584, 588-89, 198 USPQ 193, 196 (1978). As set forth in MPEP section 2106.05(d), the courts have recognized that limitations directed to data gathering and output that are claimed as insignificant extra-solution activity are routine, well understood and conventional, including receiving and sending information over a network to gather data (buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014) (computer receives and sends information over a network)) and storing and retrieving information in memory, Versata Dev. Group, Inc. v. SAP Am., Inc., 793 F.3d 1306, 1334, 115 USPQ2d 1681, 1701 (Fed. Cir. 2015); OIP Techs., 788 F.3d at 1363, 115 USPQ2d at 1092-93).
The additional elements do not comprise an inventive concept when considered individually or as an ordered combination that transforms the claimed judicial exception into a patent-eligible application of the judicial exception. Therefore, the claims do not amount to significantly more than the judicial exception itself (Step 2B: No). As such, claims 1-13 are not patent eligible.
Claim Rejections - 35 USC § 102
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
10. Claims 1-3, 6, 10 and 12 are rejected under 35 U.S.C. 102(a)(1) as being unpatentable over Ren et al. (WO2021041384A1, published 4 March 2021, IDS 09 November 2023), The italicized text corresponds to the instant claim limitations.
Regarding claim 1, an ‘unknown protein’ is interpreted, based on the specification, to mean proteins that are to be tested by the described method whose properties and/or characteristics, like specific protein descriptors, are not necessarily known at the time when the disclosed method performed (p. 5, lines 27-34).
Regarding claim 1, Ren et al. discloses a computer system and method to facilitate protein formulation optimization for drug product development. Ren et al. further discloses that the protein can be a new molecule, for which training data may not exist. Ren et al. further discloses that after a specific molecule is chosen for a given indication, a number of different candidate formulations (e.g., different solution pH levels, different excipient types, etc.) are considered for use with the molecule, and screened to find the top/best candidate formulations to optimize for viscosity, as well as product quality attributes relating to activity and chemical and physical stability (e.g., in terms of chemical structure/properties) of the molecule/formulation combination over time and at various temperatures (para. 0004; 0007; 0036; Fig. 1-2; A method for selecting at least one viscosity changing excipient (2) for a formulation (8) containing at least one unknown protein (11) via a computer (6)).
Regarding claim 1, Ren et al. discloses that a set of formulation descriptors may include descriptors of a candidate product molecule (i.e. a protein), descriptors of a candidate excipient, and descriptors of a candidate solution. Ren et al. further disclose that a data collection unit may receive the descriptors from a user (e.g., a user of client device), or may read the descriptors from a file or database. Ren et al. further discloses that the formulation ‘descriptors’ are categorized into groups and that a group of descriptors can correspond to a range of expected values for viscosity (para. 0023 0007; Providing a data set (1) from a database that describes the viscosity of several known formulations containing at least one protein and optionally at least one viscosity changing excipient (2);
Pertaining to claim 1, Ren et al. discloses using a list of excipients to generate a list of canonical SMILES of molecules (stored as File EX) and using ChemoPy software to get a list of physical and chemical descriptors of excipients (Fig. 5; para. 0044-0045; Generating representations of at least one excipient (2) from a list of excipients by the computer (6) via In-Silico-simulations)
Pertaining to claim 1, Ren eta l. discloses that the list of physical and chemical descriptors of excipients are further analyzed by dimensionality reduction and pattern recognition algorithms, followed by using PCAs or tSNEs for clustering, and finally using the clusters of excipients to improve model predictions. Ren et al. further discloses a computing system configured to train machine learning models using historical data corresponding to product molecule/protein descriptors, and use those models to predict viscosity and/or other protein formulation properties based on different sets of formulation descriptors (including product molecule/protein descriptors, excipient descriptors, and/or solution descriptors). Ren et al. further discloses that prediction models may be trained using numerous training data sets that each include (1) a formulation descriptor set which is used to predict viscosity and (2) a label indicating the actual/historical measured value for that formulation descriptor set. Ren et al. further discloses that the predictions can be made for a “new” antibody or new product XYZ147 or new molecule ( para. 0007; 0018; 0037-0038; ; 0040; 0042; Fig. 5; Fig. 6A-B; Using a Machine Learning Model (5) executed on the computer (6) that uses the generated representations of at least one excipient (2) to recognize patterns in the data set (1) to evaluate the viscosity changing effect of at least one viscosity changing excipient (2) chosen from the list of excipients to a new formulation (8) containing at least one unknown protein (11) and the at least one viscosity changing excipient (2) by applying the recognized patterns on provided data of the at least one unknown protein (11)).
Pertaining to claim 1, Ren et al. discloses that after a specific molecule is chosen for a given indication, a number of different candidate formulations (e.g., different solution pH levels, different excipient types, etc.) are considered for use with the molecule, and screened to find the top/best candidate formulations that have an optimized viscosity, as well as product quality attributes relating to activity and chemical and physical stability (para. 0004; Fig. 1; Selecting, depending on the evaluation result, the at least one excipient from the list according to an acquisition criterion and applying it to the unknown protein (11)).
Pertaining to claim 1, Ren et al. discloses that the provided data for the new formulation predicted to be optimal for the product (protein) includes the protein name, a description of the viscosity (i.e. predicted viscosity value), a viscosity class/range that was analyzed (e.g. 0-15cP), a pH, the excipient components, and a graph of predicted viscosity versus concentration in platform buffer at a specified temperature. Ren et al. futher discloses that the data can be provided using a graphical user interface (GUI) (para. 0046; Fig. 6A-B; wherein the provided data of the at least one unknown protein (11) are data describing the viscosity of a protein composition containing the at least one unknown protein (11) and optionally with at least one viscosity changing excipient (2)).
Pertaining to claim 2, Ren et al. discloses that the output provided on the graphical user interface presents to the user displays predictions provided by the prediction unit and/or other information derived from those predictions. Ren et al. further provides examples displaying the protein name, the excipient name, the viscosity value predicted, the pH and other information (para. 0046-0047; Fig. 6A-B; The method according to claim 1, wherein the evaluation of the viscosity changing effect of at least one viscosity changing excipient (2) is done by predicting the viscosity (3) of the new formulation (8) containing at least one unknown protein (11) and at least one viscosity changing excipient (2)).
Regarding claim 3, Ren et al. discloses that the models are trained using historical data stored in a training database and that the excipients may be known excipients for which a large amount of historical data exists. Ren et al. further discloses that values of formulation descriptors of the training datasets were determined based on actual historical measurements (para. 0031; 0037; The method according to claim 1, wherein the data set (1) has been generated by experimental measurements (10) and is stored in the database via the computer (6)).
Pertaining to claim 6, Ren et al. discloses that a computing system trains (and validates) machine learning models using historical data stored in training database. Ren et al. further discloses that for new molecules, this training data may not exist; therefore, the patterns in the amino acid sequence, relative composition of amino acids and homology modeling (based on the amino acid sequence of the product molecule) may be used to determine more detailed categorical descriptors for the new molecule (e.g., particular types of chemical and/or structural properties, etc.), and/or to determine numerical descriptors for the new molecule (e.g., hydrophobic surface area or charge on specific regions, etc.). Ren et a. further discloses that for new molecules, a user can enter the amino acid sequence of the molecule into the homology modeling software, and selecting an appropriate molecule template. The homology modeling software may then attempt to “fit” the amino acid sequence to the selected template. Ren et al. discloses that for the excipient, the ‘formulation descriptor set’ for training the machine learning model may use categorical and/or numerical descriptors including category/type of the excipient, such as “NaCI,” “Creatinine,” “Arginine,” “Sarcosine,” molecular weight (MW), charge, solubility, bonds, general category (e.g., surfactant vs. amino acid, etc.), and/or other suitable descriptors. Ren et al. further discloses that prediction models may be trained using numerous training data sets that each include (1) a formulation descriptor and (2) a label indicating the actual/historical measured value for that formulation descriptor set. Ren et al. further discloses that a ‘formulation descriptor set’ is input into a model to predict a value of the protein formulation property (such as viscosity) (para. 0031; 0038-0039; 0040; 0042; Fig. 7, abstract; The method according to claim 1, wherein the Machine Learning Model (5) is created and trained by combining the data set (1) describing the viscosity of at least one prototypical protein formulation (8) with the representations of the at least one viscosity changing excipient (2) or a combination thereof).
Regarding claim 10, Ren et al. discloses that a visualization (e.g., in a separate GUI) may display a rank-ordering of excipient/solution formulations for a specific molecule (e.g., ranked according to predicted viscosity, with lower viscosity values corresponding to a higher rank (para. 0050; The method according to claim 1, wherein the acquisition criterion is which viscosity changing excipient (2) reduces the viscosity the most).
Regarding claim 12, “cross-validated experimentally will be interpreted to mean “validated experimentally”.
Regarding claim 12, Ren et al. discloses that the protein formulation property (PFP) prediction application predicts the value of a protein formulation property for a particular set of inputs (formulation descriptors) and that this may be performed as a part of a formulation screening and optimization process (e.g., to provide a faster and more focused alternative to stage 14 (empirical testing to measure viscosity, stability and activity for each formulation) and stage 16 (large-scale production of empirically tested candidates) in FIG. 1), or for research, development, validation, and/or other purposes (para. 0030; para. 0005-0006; Fig. 1; The method according to claim1, wherein the generated representations of excipients (2) are cross-validated experimentally).
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
11. Claim 4 is rejected under 35 U.S.C. 103 as being unpatentable over Ren et al. (WO2021041384A1, filed 25 August 2020, IDS 09 November 2023), as applied to claims 1-3, 6, 10 and 12 above, in view of Whitaker et al. (Journal of Pharmaceutical Sciences, Vol. 106, 2017, p. 3230-3241). The italicized text corresponds to the instant claim limitations.
The limitations of claims 1-3, 6, 10 and 12 are taught by Ren et al. above.
Pertaining to claim 4, Ren et al. is silent to: the method according to claim 1, wherein as at least one excipient from the list which changes the viscosity of the new formulation (8) the most sufficient a combination of two or more excipients from the list is used. However, this limitation was known in the art at the time of the effective filing date of the invention as taught by Whitaker et al.
Pertaining to claim 4, Whitaker et al. teaches an approach to identify and select stable ultra-high-concentration monoclonal antibody formulations with reduced viscosities. Whitaker et al. further teaches that various categories and individual types of pharmaceutical excipients and other additives (56 in total) were screened for their viscosity reducing effects on 2 different mAbs. Whitaker et al. further discloses the identification of combinations of excipients with optimal viscosity in solution with monoclonal antibodies mAb A and mAb (abstract; Table 2; The method according to claim1, wherein as at least one excipient from the list which changes the viscosity of the new formulation (8) the most sufficient a combination of two or more excipients from the list is used).
An invention would have been prima facie obvious to one of ordinary skill in the art at the effective filing date of the invention if some motivation in the prior art would have led that person to combine the prior art teachings to arrive at the claimed invention. Whitaker et al. taught that their approach of using combinations of excipients considered not only viscosity effects but also known protein stabilization effects of certain additives. Whitaker et al. further taught that the sugar sucrose
was included in these formulations for its known stabilizing effect on monoclonal antibodies, especially during freeze-thaw and that its addition at 5% to the existing excipient had no notable effects on viscosity for either mAb A or mAb C when formulated with 75 mM NaCl and 50 mM arginine or lysine, respectively, plus 10 mM histidine (p. 3234, col 1, para. 2 – col. 2, para. 1; Table 2). Therefore, one of ordinary skill in the art would have been motivated to utilize the excipient combination approach taught by Whitaker et al. in the ML modeling approach to optimize excipient-drug viscosities taught by Ren et al., in order to improve to improve stability of monoclonal antibodies while optimizing viscosity. Furthermore, one of ordinary skill in the art would predict that the data and data acquisition methods taught by Whitaker et al. could be readily added to the method and system of Ren et al. with a reasonable expectation of success because they both pertain to optimizing viscosity of excipient drug combinations. The invention is therefore prima facie obvious.
12. Claims 5 and 14-16 are rejected under 35 U.S.C. 103 as being unpatentable over Ren et al. (WO2021041384A1, filed 25 August 2020, IDS 09 November 2023), as applied to claims 1-3, 6, 10 and 12 above, in view of Lee et al. (US20240013866A1, EFD 25 March 2021), and Narayanan et al. (Trends in Pharmacological Sciences, 2021, Vol. 42, p. 151 - 166). The italicized text corresponds to the instant claim limitations.
The limitations of claims 1-3, 6, 10 and 12 are taught by Ren et al. above.
Pertaining to claim 5, Ren et al. discloses that the protein formulation property (PFP) prediction application predicts the value of a protein formulation property for a particular set of inputs (formulation descriptors) and that this may be performed as a part of a formulation screening and optimization process (e.g., to provide a faster and more focused alternative to stage 14 (empirical testing to measure viscosity, stability and activity for each formulation) and stage 16 (large-scale production of empirically tested candidates) in FIG. 1), or for research, development, validation, and/or other purposes (para. 0030; para. 0005-0006; Fig. 1; The method according to claim1, wherein at least one specific experimental measurement (10) is proposed to a formulation specialist (9), who conducts the at least one respective experiment (10) in a lab to validate the predicted viscosities (3)) .
Pertaining to claims 14, Ren et al. discloses that the protein formulation property (PFP) prediction application predicts the value of a protein formulation property for a particular set of inputs (formulation descriptors) and that this may be performed as a part of a formulation screening and optimization process (e.g., to provide a faster and more focused alternative to stage 14 (empirical testing to measure viscosity, stability and activity for each formulation) and stage 16 (large-scale production of empirically tested candidates) in FIG. 1), or for research, development, validation, and/or other purposes. Ren et al. further discloses one or more processors, each of which may be a programmable microprocessor that executes software instructions stored in memory unit to execute some or all of the functions of computing system (para. 0030; para. 0005-0006; Fig. 1; para. 0020; A Machine Learning Model performed on a computer which is created and trained according to claim 5 [which recites: the method according to claim1, wherein at least one specific experimental measurement (10) is proposed to a formulation specialist (9), who conducts the at least one respective experiment (10) in a lab to validate the predicted viscosities (3)]) .
Regarding claims 5 and 14, Ren et al. is silent to: trains the Machine Learning Model (5) with the validated results by adding them to the provided data set (1) in the database via the computer (6)) (claim 5) and A Machine Learning Model performed on a computer which is created and trained according to claim 5 [which recites: trains the Machine Learning Model (5) with the validated results by adding them to the provided data set (1) in the database via the computer (6)] (claim 14). However, these limitations were known in the art at the time of the effective filing date of the invention as taught by Lee et al. and Narayanan et al.
Regarding claim 5, Lee et al. teaches a method and system for machine learning based prediction of the properties of chemical formulations including viscosity. Lee et al. discloses that in some implementations, the machine learning models can be used in active learning techniques. Lee et al. further discloses results from the testing can be used as training data to re-train the machine-learned model. After re-training, predictions from the model can then again be used to identify certain molecules or mixtures for testing. Thus, an iterative pipeline can be evaluated where a model is used to select candidates and then testing results for the candidates can be used to re-train the model, and so on (para. 0128; abstract; trains the Machine Learning Model (5) with the validated results by adding them to the provided data set (1) in the database via the computer (6)).
Regarding claims 14, Lee et al. teaches a method and system for machine learning based prediction of the properties of chemical formulations including viscosity. Lee et al. discloses that in some implementations, the machine learning models can be used in active learning techniques. Lee et al. further discloses results from the testing can be used as training data to re-train the machine-learned model. After re-training, predictions from the model can then again be used to identify certain molecules or mixtures for testing. Thus, an iterative pipeline can be evaluated where a model is used to select candidates and then testing results for the candidates can be used to re-train the model, and so on (para. 0128; abstract; A Machine Learning Model performed on a computer which is created and trained according to claim 5 [which recites: trains the Machine Learning Model (5) with the validated results by adding them to the provided data set (1) in the database via the computer (6)]).
An invention would have been prima facie obvious to one of ordinary skill in the art at the effective filing date of the invention if some motivation in the prior art would have led that person to combine the prior art teachings to arrive at the claimed invention. Narayanan et al. teaches applications for machine learning techniques in biologics discovery and development, focusing on protein engineering, early biophysical screening, and formulation, and discusses the power of machine learning (ML) in extracting information from complex datasets. Narayanan et al. further teaches that active learning, which involves updating training datasets and thus the models and using them iteratively to design new experiments, ensures the efficient learning of input–output relationships in the entire design space. In comparison to classical design of experiments (DoE), active learning methods present the ability to learn the relevant input–output information more resource-efficiently (Fig. 2b; p. 154, para. 1). Therefore, one of ordinary skill in the art would have been motivated to utilize the active learning technique taught by Lee et al. in the ML-based viscosity prediction method and system taught by Ren et al. in order to use laboratory and computational resources more efficiently. Furthermore, one of ordinary skill in the art would predict that the active learning methods taught by Lee et al. could be readily added to the system and method of Ren et al. with a reasonable expectation of success because they both pertain to ML-based prediction of properties of formulation including viscosity using large-scale data sets combined with experimental validation. The invention is therefore prima facie obvious.
Regarding claim 15, Ren et al. disclose that some or all of the machine learning models at the second stage (predicting) and possibly also the first stage (training and feature selection) may be regression models, such as random forest models or support vector regression model, or any other suitable type of machine learning model (para. 0007; 0053; The Machine Learning Model according to claim 14, wherein the Gaussian Process is replaced with any other model architecture fulfilling the same purpose).
Pertaining to claim 16, Ren et al. disclose that some or all of the machine learning models at the second stage (predicting) and possibly also the first stage (training and feature selection) may be regression models, such as random forest models or support vector regression model, or any other suitable type of machine learning model (para. 0007; 0053; The Machine Learning Model according to claim 14, wherein the Gaussian Process is replaced with another type of stochastic process, a deep Bayesian network, a generalized linear model, a neural network, a support vector machine, a tree-based model, or an ensemble model).
13. Claim 7 is rejected under 35 U.S.C. 103 as being unpatentable over Ren et al. (WO2021041384A1, filed 25 August 2020, IDS 09 November 2023), in view of Jin et al. (Advances in Polymer Technology, 2020, Vol. 2020, p. 1-14), and Lunney et al. (J. Pharm Innov., 2008, Vol. 3, p. 188-203). The italicized text corresponds to the instant claim limitations.
The limitations of claims 1-3, 6, 10 and 12 are taught by Ren et al. above.
Pertaining to claim 7, Ren et al. is silent to: The method according to claim 6, wherein the viscosity values of a given formulation (8) are modelled via the Machine Learning Model (5) in the form of a Gaussian process and the model predictions (3) are used to guide the formulation specialist (9) by means of a Bayesian optimal experimental design. However, this limitation was known in the art at the time of the effective filing date of the invention as taught by Jin et al. and Lunney et al.
Pertaining to claim 7, Jin et al. teaches using ensemble just-in-time learning with Gaussian process regression to predict viscosity. Jin et al. further teaches for a new input, the training outputs and the new viscosity output follow a joint Gaussian distribution and from that calculates a predicted viscosity and variance associated with that prediction (i.e. a conditional Gaussian distribution which is a posterior predictive distribution for viscosity) (abstract; p. 6, col. 1, para. 2-4; Fig. 2; The method according to claim 6, wherein the viscosity values of a given formulation (8) are modelled via the Machine Learning Model (5) in the form of a Gaussian process and the model predictions (3) are used to guide the formulation specialist (9) by means of a Bayesian optimal experimental design).
An invention would have been prima facie obvious to one of ordinary skill in the art at the effective filing date of the invention if some motivation in the prior art would have led that person to combine the prior art teachings to arrive at the claimed invention. Jin et al. taught that just in time learning (JIT) has strong capability of handling nonlinearity, time-varying behavior, and multiphase and multimode process characteristics. Jin et al. further taught that unlike others ensemble JIT approaches, their method can consider multiple perturbations together (e.g. perturbations to training data, input variables, similarity measure, local modeling technique or model parameters) and it uses adaptive learning for improved prediction performance in predicting viscosity (p. 2, col. 1, para. 2 – col. 2, para. 3). Therefore, one of ordinary skill in the art would have been motivated to utilize the Gaussian process regression taught by Jin et al. in the ML-based viscosity prediction method and system taught by Ren et al. in order to handle nonlinearity and multiphase process characteristics and to improve prediction performance in predicting viscosity by using active learning. Furthermore, one of ordinary skill in the art would predict that the Gaussian process methods taught by Jin et al. could be readily added to the system and method of Ren et al. with a reasonable expectation of success because they both pertain to ML-based prediction of viscosity. The invention is therefore prima facie obvious.
Pertaining to claim 7, Ren et al. and Jin et al. are silent to wherein the modeled predictions are used to guide the formulation specialist (9) by means of a Bayesian optimal experimental design. However, this would be obvious to try. As disclosed by Lunney et al., given the complexity of pharmaceutical processes and the number of variables to consider in developing operational ranges, only a well-executed program of experiments, supported by appropriate statistical analyses, could provide the necessary information to truly capture a design space. At the time of the invention, there were a finite number of experimental design strategies (including one-factor-at a time approach, orthogonal array experiments, response surface methods and Bayesian optimal design). However, typical approaches that employ one factor at a time are ineffective because they cannot characterize interactions among factors and are subject to incidental confounding, and other approaches (orthogonal array experiments and response surface methods) are restrictive, inefficient and costly. In contrast, Bayesian optimal design effectively addresses and characterizes complex design space of pharmaceutical processes and is highly flexible without restriction imposed by reliance on textbook designs (abstract; p. 202, col. 2, para. 2-3; p. 190, col. 2, para. 4 – p. 198, col. 1, para. 1). A person having ordinary skill in the art could have pursued using Bayesian optimal design with a reasonable expectation of success because as disclosed by Lunney et al. there are numerous references concerning the development and construction of Bayesian design in the literature in the field of pharmaceutical manufacturing (p. 202, col. 1, para. 1 – col. 2, para. 1).
14. Claims 8-9 are rejected under 35 U.S.C. 103 as being unpatentable over Ren et al. (WO2021041384A1, filed 25 August 2020, IDS 09 November 2023), as applied to claims 1-3, 6, 10 and 12 above, in view of Romero et al. (ASC Synthetic Biology, 2012, Vol. 1, p. 221-228), and further in view of Lee et al. (US20240013866A1, EFD 25 March 2021) and Narayanan et al. (Trends in Pharmacological Sciences, 2021, Vol. 42, p. 151 - 166). The italicized text corresponds to the instant claim limitations.
The limitations of claims 1-3, 6, 10 and 12 are taught by Ren et al. above.
Regarding claim 8, Ren et al. discloses that the protein formulation property (PFP) prediction application predicts the value of a protein formulation property for a particular set of inputs (formulation descriptors) and that this may be performed as a part of a formulation screening and optimization process (e.g., to provide a faster and more focused alternative to stage 14 (empirical testing to measure viscosity, stability and activity for each formulation) and stage 16 (large-scale production of empirically tested candidates) in FIG. 1), or for research, development, validation, and/or other purposes (para. 0030; para. 0005-0006; Fig. 1; Proposing the new set of excipients (2) or the combination thereof to the formulation specialist (9), who then conducts the respective experiments (10)).
Regarding claim 8, Ren et al. is silent to: the method according to claim 6, wherein the training of the Machine Learning Model on the computer is done by performing at least once the following steps: Optimizing the Machine Learning Model Parameters with training data from the data set (1) by maximizing the marginal likelihood of the training data; Evaluating a posterior distribution of viscosity values for untested excipients (2) or a combination thereof based on the Machine Learning Model (5) and thereby predicting a viscosity (3)); Selecting a new set of excipients (2) or a combination thereof by optimizing an acquisition score obtained from the computed posterior distribution. However, this limitation was known in the art at the time of the effective filing date of the invention as taught by Romero et al.
Regarding claim 8, Romero et al. teaches a method to predict properties of therapeutic proteins including stability and catalysis using a Bayesian linear regression model. Romero further teaches that For regression models, the independent variable corresponded to chimera sequences and is represented with a binary vector x, where xi indicates the parent identity at block i. Romero et al. further teaches using regression analysis wherein a chimera’s binary functional status was modeled with a Bayesian logistic regression model, which contains a Bernoulli likelihood function and a zero-mean, isotropic Gaussian prior on coefficients.45 The resulting posterior distribution was approximated using Laplace’s method and prior variance was estimated from the data by maximizing the marginal likelihood function. (abstract, p. 227, col. 1, para. 2; wherein the training of the Machine Learning Model (5) on the computer (6) is done by performing at least once the following steps: Optimizing the Machine Learning Model Parameters with training data from the data set (1) by maximizing the marginal likelihood of the training data).
Regarding claim 8, Romero et al. teaches that a chimera’s (i.e. proteins) binary functional status was modeled with a Bayesian logistic regression model, which contains a Bernoulli likelihood function and a zero-mean, isotropic Gaussian prior on coefficients. Romero et al. further teach that the resulting posterior distribution was approximated using Laplace’s method and prior variance was estimated from the data by maximizing the marginal likelihood function. Romero discloses that using Newton’s method, they found the maximum a posteriori (MAP) estimates for each chimera block’s contribution to functionality and the probability that a chimera is functional was estimated by applying the MAP parameter estimates to the logistic model (p. 227, col. 1, para. 3; Evaluating a posterior distribution of viscosity values for untested excipients (2) or a combination thereof based on the Machine Learning Model (5) and thereby predicting a viscosity (3)).
Regarding claim 8, Romero et al. discloses that logarithm of a chimera’s long-term stability (AUC) was modeled with a Bayesian linear regression model, which consists of a Gaussian likelihood function with a zero-mean, isotropic Gaussian prior on coefficients. Romero et al. further discloses that the measurement noise and prior variance were estimated from the data by maximizing the marginal likelihood function. With these hyperparameters, MAP estimates for each chimera block’s contribution to long-term stability were found in closed-form. Romero et al. further discloses that active learning algorithm selects sequences from the initial library of 7 chimeric enzymes that are both highly informative and functional (abstract; p. 227; col. 1, para. 4; Selecting a new set of excipients (2) or a combination thereof by optimizing an acquisition score obtained from the computed posterior distribution).
An invention would have been prima facie obvious to one of ordinary skill in the art at the effective filing date of the invention if some motivation in the prior art would have led that person to combine the prior art teachings to arrive at the claimed invention. Romero et al. taught that the regression analysis approach with Bayesian parameter estimation, is good for studies with limited datasets and that it outperforms maximum likelihood estimation for small data sets (p. 227, col. 1, para. 2). Therefore, one of ordinary skill in the art would have been motivated to utilize the regression-based modeling approach taught by Romero et al. in the ML modeling to predict viscosity from excipient descriptors and molecule descriptors taught by Ren et al., in order to get good performance on small datasets. Furthermore, one of ordinary skill in the art would predict that the excipient representations taught by Romero et al. could be readily added to the system of Ren et al. with a reasonable expectation of success because they both pertain to machine learning analysis to predict properties of protein therapeutics. The invention is therefore prima facie obvious.
Regarding claim 8, Ren et al. and Romero et al. are silent to: Adding the obtained measurements (10) to the training data. However, this limitation was known in the art at the time of the effective filing date of the invention as taught by Lee et al. and Narayanan et al.
Regarding claim 8, Lee et al. teaches a method and system for machine learning based prediction of the properties of chemical formulations including viscosity. Lee et al. discloses that in some implementations, the machine learning models can be used in active learning techniques. Lee et al. further discloses results from the testing can be used as training data to re-train the machine-learned model. After re-training, predictions from the model can then again be used to identify certain molecules or mixtures for testing. Thus, an iterative pipeline can be evaluated where a model is used to select candidates and then testing results for the candidates can be used to re-train the model, and so on (para. 0128; abstract; Adding the obtained measurements (10) to the training data).
An invention would have been prima facie obvious to one of ordinary skill in the art at the effective filing date of the invention if some motivation in the prior art would have led that person to combine the prior art teachings to arrive at the claimed invention. Narayanan et al. teaches applications for machine learning techniques in biologics discovery and development, focusing on protein engineering, early biophysical screening, and formulation, and discusses the power of machine learning (ML) in extracting information from complex datasets. Narayanan et al. further teaches that active learning, which involves updating training datasets and thus the models and using them iteratively to design new experiments, ensures the efficient learning of input–output relationships in the entire design space. In comparison to classical design of experiments (DoE), active learning methods present the ability to learn the relevant input–output information more resource-efficiently (Fig. 2b; p. 154, para. 1). Therefore, one of ordinary skill in the art would have been motivated to utilize the active learning technique taught by Lee et al. in the ML-based viscosity prediction method and system taught by Ren et al. and Romero et al. in order to use laboratory and computational resources more efficiently. Furthermore, one of ordinary skill in the art would predict that the active learning methods taught by Lee et al. could be readily added to the system and method of Ren et al. and Romero et al. with a reasonable expectation of success because they both pertain to ML-based prediction of properties of formulation including viscosity using large-scale data sets combined with experimental validation. The invention is therefore prima facie obvious.
Pertaining to claim 9, Ren et al. teaches that after a specific molecule is chosen for a given indication, a number of different candidate formulations (e.g., different solution pH levels, different excipient types, etc.) are considered for use with the molecule, and screened to find the top/best candidate formulations by optimizing for viscosity, as well as product quality attributes relating to activity and chemical and physical stability (e.g., in terms of chemical structure/properties) of the molecule/formulation combination over time and at various temperatures (para. 0004; The method according to claim 8, wherein the prediction of the viscosity (3) obtained from the posterior distribution of viscosity values (3) is based on a pH-dependent feature vector characterizing the excipients (2) used in the considered formulation (8) and on the used excipient concentration levels).
15. Claim 11 is rejected under 35 U.S.C. 103 as being unpatentable over Ren et al. (WO2021041384A1, filed 25 August 2020, IDS 09 November 2023), as applied to claims 1-3, 6, 10 and 12 above, in view of Reker et al. (Nature Nanotechnology, Vol. 16, 2021, p. 725 - 733 published 25 March 2021) and Cloutier et al. (Mol. Pharmaceutics, 2019, Vo. 16, p. 3657-3662, 9 November 2023 IDS). The italicized text corresponds to the instant claim limitations.
The limitations of claims 1-3, 6, 10 and 12 are taught by Ren et al. above.
Regarding claim 11, physical parameters will be interpreted to describe the excipients and their properties including the charge distribution, dipole moment, quadrupole moment trace and anisotropy, polarizability, molecular London dispersion coefficient (C6), log P water/hexane distribution coefficient, solvent accessible surface area, molecular orbital energy HOMO-LUMO gap (as indicated in the specification on p. 10, lines 14-24). Molecular fingerprints will be interpreted to be a compact, computable representation of a molecule’s structural or property features, such as atom connectivity, functional groups, or 3D shape, enabling rapid comparison and analysis (i.e. a vector that represents the chemical structure). According to the specification, molecular fingerprints can include topological features that may be generated using RDKit secondary, structured collection of chemical descriptors used to enrich the baseline structural data of excipients (p. 19, lines 18-25).
Pertaining to claim 11, Ren et al. discloses that models are trained using excipient descriptors, molecule descriptors and solution descriptors to predict viscosity in a particular formulation described by a set of formulation descriptors. Ren et al. disclose that the excipient formulation descriptor set may include a category/type of the excipient, such as “NaCI,” “Creatinine,” “Arginine,” “Sarcosine,” as well as molecular weight (MW), charge, solubility, bonds, general category (e.g., surfactant vs. amino acid, etc.). Ren et al. further discloses that ChemoPy (a Python package) reads and input file and produces a list of physical and chemical descriptors of the excipients are used for modeling. (para. 0007; 0039; 0044-0045; The method according to claim 1, wherein the representations of excipients (2) are generated by the computer (6) in the form of physical parameters).
Regarding claim 11, Ren et al. is silent to wherein the representations of excipients include molecular fingerprints. However, this limitation was known in the art at the time of the effective filing date of the invention as taught by Reker et al. and Cloutier.
Regarding claim 11, Reker et al. discloses a computational method for high-throughput design of self-assembling drug nanoparticles (i.e. formulations that are pairings of drugs and excipients). Reker et al. further discloses that MD simulations of drug–excipient systems were computationally analysed to quantify non-covalent interaction potentials. These potentials serve as input for our machine learning model. The machine learning model uses both these interaction potentials and the molecular properties of drugs and excipients to predict which drug–excipient pairs will most likely lead to nanoparticle formation. Reker et al. further discloses that for ML modeling, input for chemical structures of all drugs were simplified molecular-input line-entry system (SMILES) representations and data for excipients were described according to radial chemical substructures (Morgan Fingerprint, radius 4, 2,048 bits, rdkit.org) and calculated physicochemical properties (rdkit.Chem.Descriptors._descList, rdkit.org) based on previously published evaluations (Fig. 1; p. 734, col. 1, para. 2; wherein the representations of excipients include molecular fingerprints).
An invention would have been prima facie obvious to one of ordinary skill in the art at the effective filing date of the invention if some motivation in the prior art would have led that person to combine the prior art teachings to arrive at the claimed invention. Reker et al. taught a method to predict physical properties of nanoparticles in excipients and that their approach can improve the selection of specific nanoparticles for various applications. Reker et al. further show that machine learning modeling with both physical parameters and molecular fingerprints can be used to anticipate nanoparticle stability and suggest that application to anticipate more complex properties would be appropriate to further facilitate nanoparticle selection. Reker et al. further discloses that their method and extensions thereof will be an important step towards personalized drug delivery (p. 732, col. 2, para. 2). Furthermore, Cloutier et al., who developed ML models of antibody-excipient interactions, used their approach to predict both experimental aggregation (i.e. stability) and viscosity behavior (abstract), demonstrating that both properties are driven by the same overlapping molecular forces like protein-protein interaction sand surface charge distribution. Therefore, one of ordinary skill in the art would have been motivated to utilize the approach to represent excipient features in the form of both physical parameters and molecular fingerprints as taught by Reker et al. in the ML modeling to predict viscosity from excipient descriptors and molecule descriptors taught by Ren et al., in order to improve selection of nanoparticle drugs by predicting solubility and/or other more complex properties such as viscosity. Furthermore, one of ordinary skill in the art would predict that the excipient representations taught by Reker et al. could be readily added to the system of Ren et al. with a reasonable expectation of success because they both pertain to machine learning analysis to predict properties of excipient-drug combinations. The invention is therefore prima facie obvious.
16. Claim 13 is rejected under 35 U.S.C. 103 as being unpatentable over Ren et al. (WO2021041384A1, filed 25 August 2020, IDS 09 November 2023), as applied to claims 1-3, 6, 10 and 12 above, in view of by Pozarska et al. (European Journal of Pharmaceutical Sciences, Vol. 49, 2013, p. 505-511), and Nordness et al. (Journal of Molecular Liquids, Vol., 334, 2021, p. 116019-116028). The italicized text corresponds to the instant claim limitations.
The limitations of claims 1-3, 6, 10 and 12 are taught by Ren et al. above.
Regarding claim 13 Ren et al. is silent to: The method according to claim 1, wherein the generated representations of excipients (2) include quantum mechanic features, optionally complemented with a set of topological molecular fingerprints. However, this limitation was known in the art at the time of the effective filing date of the invention as taught by Pozarska et al. and Nordness et al.
Regarding claim 13, Pozarska et al. taught the application of the quantum chemistry program, Conductor Screening Model for Real Solvents (COSMO-RS) to help with the selection of excipients for formulation development in early discovery. The excipient solubility predictions obtained from COSMO-RS were compared to experimentally obtained solubilities. The results showed that in general, COSMO-RS was able to help formulators with the selection of the most appropriate excipients to solubilise the model compound (abstract; The method according to claim 1, wherein the generated representations of excipients (2) include quantum mechanic features, optionally complemented with a set of topological molecular fingerprints).
An invention would have been prima facie obvious to one of ordinary skill in the art at the effective filing date of the invention if some motivation in the prior art would have led that person to combine the prior art teachings to arrive at the claimed invention. Pozarska et al. taught that the benefits of COSMO-RS include reducing the resource and time spent on formulation development while making effective use of the limited bulk available in early discovery (p. 511, col. 1, para. 1). Therefore, one of ordinary skill in the art would have been motivated to utilize the approach to use quantum mechanic features to predict properties of excipient formulations as taught by Pozarska et al. in the ML modeling to predict viscosity from excipient descriptors and molecule descriptors taught by Ren et al., in order to reduce resources and time spent on formulation development while making effective use of limited bulk available in early discovery. Furthermore, one of ordinary skill in the art would predict that the quantum mechanics features taught by Pozarska et al. could be readily added to the system of Ren et al. with a reasonable expectation of success because they both pertain to machine learning analysis to predict excipient properties. Furthermore, as disclosed by Nordness et al., the same quantum mechanics features (COSMO-RS) in machine learning models can be used to predict viscosity (abstract). The invention is therefore prima facie obvious.
Conclusion
17. No claims are allowed.
E-mail Communications Authorization
18. Per updated USPTO Internet usage policies, Applicant and/or applicant's representative is encouraged to authorize the USPTO examiner to discuss any subject matter concerning the above application via Internet e-mail communications. See MPEP 502.03. To approve such communications, Applicant must provide written authorization for e-mail communication by submitting the following statement via EFS-Web (using PTO/SB/439) or Central Fax (571-273-8300): "Recognizing that Internet communications are not secure, / hereby authorize the USPTO to communicate with the undersigned and practitioners in accordance with 37 CFR 1.33 and 37 CFR 1.34 concerning any subject matter of this application by video conferencing, instant messaging, or electronic mail. / understand that a copy of these communications will be made of record in the application file."
Written authorizations submitted to the Examiner via e-mail are NOT proper. Written authorizations must be submitted via EFS-Web (using PTO/SB/439) or Central Fax (571-273- 8300). A paper copy of e-mail correspondence will be placed in the patent application when appropriate. E-mails from the USPTO are for the sole use of the intended recipient, and may contain information subject to the confidentiality requirement set forth in 35 USC § 122. See also MPEP 502.03.
Inquiries
19. Any inquiry concerning this communication or earlier communications from the examiner should be directed to JENNIFER J SMITH whose telephone number is (571)272-7801. The examiner can normally be reached Monday-Friday 7:30 AM - 4:00 PM.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Olivia Wise can be reached at (571) 272-2249. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/J.J.S./Examiner, Art Unit 1685
/OLIVIA M. WISE/Supervisory Patent Examiner, Art Unit 1685
Prosecution Insights