DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
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 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.
Claim Status
Claims 1-20 are pending.
Claim 7 is objected to.
Claims 1-20 are rejected.
Priority
This application is a CON of 17/960,755 filed Oct 5 2022 (now USP 11,742,076), which claims priority to US provisional applications 63/252523, 63/252539, 63/252562, 63/252500, 63/292115, 63/294751, 63/328189, 63/337753, 63/400250, and 63/413150, filed as early as Oct 5 2021.
Accordingly, each of claims 1-20 are afforded the effective filing date of Oct 5 2021.
Information Disclosure Statement
The information disclosure statement (IDS) filed on Apr 19 2024 is in compliance with the provisions of 37 CFR 1.97 and has therefore been considered. A signed copy of the IDS document is included with this Office Action.
Drawings
The Drawings submitted Jun 21 2023 are accepted.
Claim Objections
The claims are objected to for the following informalities:
Claim 7 recites “wherein drug signature”, which should be amended to recite “wherein the drug signature”.
Claim Rejections - 35 USC § 112
35 U.S.C. 112(b)
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.
Claims 2-6 and 10-11 are rejected under 35 U.S.C. 112(b) as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor regards as the invention.
Claim 2, final limitation, recites “the entities”. However, claim 1, 3rd limitation, recites “one or more entities”. It is therefore not clear if claim 2 intends to require more than one entities, or if the claim continues to require only one entity and should be amended to recite “one or more entities”, which is assumed for compact examination. Claims 3-4 repeat the issue. Claims 5-6 are rejected based on their dependency from claims 2-4.
Claim 10 recites “wherein each patient embedding corresponds to a respective patient and is generated by processing multi-modal data characterizing the patient using the encoder neural network”. It is unclear whether the wherein clause is intended to require generating the patient embeddings within the metes and bounds of the claimed invention, or if it is only further limiting the patient embeddings such that performing their generation is not required within the metes and bounds of the invention. As set forth in MPEP 2111.04.I, “wherein” clauses raise the question as to the limiting effect of the language in a claim. As the claims do not recite an active generation of the embeddings, the metes and bounds of the claims are unclear. For compact examination, it is assumed that the generation is not required to be performed. The rejection may be overcome by clarifying what steps are required to be performed. Claim 11 is rejected based on its dependency from claim 10.
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.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to one or more judicial exceptions without significantly more.
MPEP 2106 organizes judicial exception analysis into Steps 1, 2A (Prongs One and Two) and 2B as follows below. MPEP 2106 and the following USPTO website provide further explanation and case law citations: uspto.gov/patent/laws-and-regulations/examination-policy/examination-guidance-and-training-materials.
Framework with which to Evaluate Subject Matter Eligibility:
Step 1: Are the claims directed to a process, machine, manufacture, or composition of matter;
Step 2A, Prong One: Do the claims recite a judicially recognized exception, i.e. a law of nature, a natural phenomenon, or an abstract idea;
Step 2A, Prong Two: If the claims recite a judicial exception under Prong One, then is the judicial exception integrated into a practical application (Prong Two); and
Step 2B: If the claims do not integrate the judicial exception, do the claims provide an inventive concept.
Framework Analysis as Pertains to the Instant Claims:
Step 1
With respect to Step 1: yes, the claims are directed to a method, a system, and non-transitory computer storage media, i.e., a process, machine, or manufacture within the above 101 categories [Step 1: YES; See MPEP § 2106.03].
Step 2A, Prong One
With respect to Step 2A, Prong One, the claims recite judicial exceptions in the form of abstract ideas. The MPEP at 2106.04(a)(2) further explains that abstract ideas are defined as:
mathematical concepts (mathematical formulas or equations, mathematical relationships and mathematical calculations);
certain methods of organizing human activity (fundamental economic practices or principles, managing personal behavior or relationships or interactions between people); and/or
mental processes (procedures for observing, evaluating, analyzing/ judging and organizing information).
The claims also recite a law of nature or a natural phenomenon. The MPEP at 2106.04(b) further explains that laws of nature and natural phenomena include naturally occurring principles/relations and nature-based products that are naturally occurring or that do not have markedly different characteristics compared to what occurs in nature.
With respect to the instant claims, under the Step 2A, Prong One evaluation, the claims are found to recite abstract ideas that fall into the grouping of mental processes (in particular procedures for observing, analyzing and organizing information) and mathematical concepts (in particular mathematical relationships and formulas) as well as a law of nature or a natural phenomenon are as follows:
Independent claims 1 and 19-20: generating a drug signature for a drug, wherein: the drug signature comprises a respective impact score for each of a plurality of features; and the impact score for a feature characterizes an impact, caused by administering a drug to one or more entities, on a value of the feature measured for the one or more entities;
generating a network input to an encoder neural network based on the drug signature; and
processing: (i) the embedding of the drug signature in the latent space, and (ii) data defining a plurality of patient categories, to generate a plurality of response scores, wherein each response score corresponds to a respective patient category and characterizes a predicted response of patients included in the patient category to the drug.
Dependent claim 12: determining a ranking of the plurality of patient categories based on the response scores.
Dependent claim 13: determining that a new patient is included in a patient category of the plurality of patient categories;
identifying the response score for the patient category of the new patient; and
… generating a recommendation for whether the new patient should receive the drug based at least in part on the response score for the patient category of the new patient.
Dependent claims 2-11 and 14-16 recite further steps that limit the judicial exceptions in independent claim 1 and, as such, also are directed to those abstract ideas. For example, claim 2 further limits generating the drug signature to being based on the pre-treatment and post-treatment feature representations of the entities; claims 3-6 further limits generating the drug signature based on a differential feature representation of the pre-treatment and post-treatment feature representations; claims 7-8 further limit the impact score to being based on gene or protein expression level; claim 9 further limits the network input to including the drug signature; claim 10 further limits the plurality of patient categories to a cluster of patient embeddings in latent space; claim 11 further limits generating the response score for the patient category; and claims 14-16 further limit the entities of claim 1.
The abstract ideas recited in the claims are evaluated under the Broadest Reasonable Interpretation (BRI) and determined to each cover performance either in the mind and/or by mathematical operation because the method only requires a user to manually generate a plurality of response scores based on drug signatures and patient categories. Without further detail as to the methodology involved in “generating”, “processing”, “determining”, and “identifying”, under the BRI, one may simply, for example, use pen and paper to: generate a drug signature with impact scores that characterizes an impact of administering a drug to entities which can be cells or patients; generate a network input based on the drug signature (the specification as published discloses that the network input can include an ordered collections of features indexed by a set of feature dimensions at [0588], which is considered to be able to be performed mentally or with pen and paper); process an embedded drug signature score produced by an encoder neural network and patient category data to generate response scores; rank the patient categories based on their response scores; determine that a new patient is included; identify the response score for the patient category with the new patient; and generate a recommendation for whether the new patient should receive the drug. Those steps directed to generating a drug signature, an impact score, and a response score as in claims 1 and 19-20, as well as determining differential feature representations, dividing, and averaging in claims 3-6, require mathematical techniques as the only supported embodiments, as is disclosed in the specification as published at least at [0575], which describes that the drug signature can be represented as an ordered collection of numerical values of impact score, as a vector, matrix, or other tensor of numerical values. These limitations therefore require mathematical techniques as the only supported embodiments, as they express mathematical relationships in words.
Further, the claims recite a natural relationship between the response of one or more entities to the administered drug.
Therefore, claims 1 and 19-20 and those claims dependent therefrom recite an abstract idea and a law of nature/natural phenomenon [Step 2A, Prong 1: YES; See MPEP § 2106.04].
Step 2A, Prong Two
Because the claims do recite judicial exceptions, direction under Step 2A, Prong Two, provides that the claims must be examined further to determine whether they integrate the judicial exceptions into a practical application (MPEP 2106.04(d)). A claim can be said to integrate a judicial exception into a practical application when it applies, relies on, or uses the judicial exception in a manner that imposes a meaningful limit on the judicial exception. This is performed by analyzing the additional elements of the claim to determine if the judicial exceptions are integrated into a practical application (MPEP 2106.04(d).I.; MPEP 2106.05(a-h)). If the claim contains no additional elements beyond the judicial exceptions, the claim is said to fail to integrate the judicial exceptions into a practical application (MPEP 2106.04(d).III).
Additional elements, Step 2A, Prong Two
With respect to the instant recitations, the claims recite the following additional elements:
Independent claims 1 and 19-20: generating an embedding of the drug signature in a latent space, comprising… processing the network input generated based on the drug signature using the encoder neural network to generate the embedding of the drug signature in the latent space.
Dependent claim 2: obtaining, for each of the entities: (i) a pre-treatment feature representation of the entity that comprises, for each of the plurality of features, a respective pre-treatment value of the feature that is measured for the entity prior to the drug being administered to the entity; and (ii) a post-treatment feature representation of the entity that comprises, for each of the plurality of features, a respective post-treatment value of the feature that is measured for the entity after the drug is administered to the entity.
Dependent claim 13: automatically generating a recommendation….
Dependent claims 17-18 recite steps that further limit the recited additional elements in the claims. For example, claim 17-18 further limit how the encoder neural network has been trained.
The claims also include non-abstract computing elements. For example, independent claim 1 includes that the method is performed by one or more computers; independent claim 19 includes a system comprising: one or more computers; and one or more storage devices communicatively coupled to the one or more computers, wherein the one or more storage devices store instructions that, when executed by the one or more computers, cause the one or more computers to perform operations; and independent claim 20 includes one or more non-transitory computer storage media storing instructions that when executed by one or more computers cause the one or more computers to perform operations.
Considerations under Step 2A, Prong Two
With respect to Step 2A, Prong Two, the additional elements of the claims do not integrate the judicial exceptions into a practical application for the following reasons. Those steps directed to data gathering, such as “generating” an embedding by “processing” the network input using an encoder neural network and “obtaining” a pre- and post-treatment feature representation, and to data outputting, such as “automatically generating” a recommendation, perform functions of collecting and outputting the data needed to carry out the judicial exceptions. It is further noted that the limitations directed to obtaining data and automatically generating an output are considered to recite normal functions of a computer for receiving and outputting data. Data gathering and outputting do not impose any meaningful limitation on the judicial exceptions, or on how the judicial exceptions are performed. Data gathering and outputting steps are not sufficient to integrate judicial exceptions into a practical application (MPEP 2106.05(g)).
Further steps directed to additional non-abstract computing elements and an encoder neural network do not describe any specific computational steps by which the “computer parts” perform or carry out the judicial exceptions, nor do they provide any details of how specific structures of the computer, such as the computer-readable recording media, are used to implement these functions. The claims state nothing more than a generic computer which performs the functions that constitute the judicial exceptions. Hence, these are mere instructions to apply the judicial exceptions using a computer, and therefore the claim does not integrate that judicial exceptions into a practical application. The courts have weighed in and consistently maintained that when, for example, a memory, display, processor, machine, etc.… are recited so generically (i.e., no details are provided) that they represent no more than mere instructions to apply the judicial exception on a computer, and these limitations may be viewed as nothing more than generally linking the use of the judicial exception to the technological environment of a computer (MPEP 2106.05(f)).
The specification does not provide a clear explanation for how the additional elements provide these improvements. Therefore, the additional elements do not clearly improve the functioning of a computer, or comprise an improvement to any other technical field. Further, the additional elements do not clearly affect a particular treatment; they do not clearly require or set forth a particular machine; they do not clearly effect a transformation of matter; nor do they clearly provide a nonconventional or unconventional step (MPEP2106.04(d)).
Thus, none of the claims recite additional elements which would integrate a judicial exception into a practical application, and the claims are directed to one or more judicial exceptions [Step 2A, Prong 2: NO; See MPEP § 2106.04(d)].
Step 2B (MPEP 2106.05.A i-vi)
According to analysis so far, the additional elements described above do not provide significantly more than the judicial exception. A determination of whether additional elements provide significantly more also rests on whether the additional elements or a combination of elements represents other than what is well-understood, routine, and conventional. Conventionality is a question of fact and may be evidenced as: a citation to an express statement in the specification or to a statement made by an applicant during prosecution that demonstrates a well-understood, routine or conventional nature of the additional element(s); a citation to one or more of the court decisions as discussed in MPEP 2106(d)(II) as noting the well-understood, routine, conventional nature of the additional element(s); a citation to a publication that demonstrates the well-understood, routine, conventional nature of the additional element(s); and/or a statement that the examiner is taking official notice with respect to the well-understood, routine, conventional nature of the additional element(s).
With respect to the instant claims, the prior art to Wei et al. (IEEE Access, 2020, 9:4939-4956; newly cited) discloses that using an encoder neural network to generate an embedding of data in a latent space is a data gathering element that is routine, well-understood and conventional in the art. Said portions of the prior art are, for example, (p. 4948, col. 2, par. 4; entire document is relevant). Further, the courts have found that receiving and outputting data are well-understood, routine, and conventional functions of a computer when claimed in a merely generic manner or as insignificant extra-solution activity (see Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information), 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), Versata Dev. Group, Inc. v. SAP Am., Inc., 793 F.3d 1306, 1334, 115 USPQ2d 1681, 1701 (Fed. Cir. 2015), and OIP Techs., 788 F.3d at 1363, 115 USPQ2d at 1092-93, as discussed in MPEP 2106.05(d)(II)(i)). As such, the claims simply append well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception (MPEP2106.05(d)). The data gathering steps as recited in the instant claims constitute a general link to a technological environment which is insufficient to constitute an inventive concept which would render the claims significantly more than the judicial exception (MPEP2106.05(g)&(h)).
With respect to claims 1 and 19-20 and those claims dependent therefrom, the computer-related elements or the general purpose computer do not rise to the level of significantly more than the judicial exception. It is further noted that the limitations directed to obtaining data and automatically generating an output are considered to recite normal functions of a computer for receiving and outputting data. The claims state nothing more than a generic computer which performs the functions that constitute the judicial exceptions. Hence, these are mere instructions to apply the judicial exceptions using a computer, which the courts have found to not provide significantly more when recited in a claim with a judicial exception (Alice Corp., 573 U.S. at 225-26, 110 USPQ2d at 1984; see MPEP 2106.05(A)). The specification as published also notes that computer processors and systems, as example, are commercially available or widely used at [0675]. The additional elements are set forth at such a high level of generality that they can be met by a general purpose computer. Therefore, the computer components constitute no more than a general link to a technological environment, which is insufficient to constitute an inventive concept that would render the claims significantly more than the judicial exceptions (see MPEP 2106.05(b)I-III).
Taken alone, the additional elements do not amount to significantly more than the above-identified judicial exception(s). Even when viewed as a combination, the additional elements fail to transform the exception into a patent-eligible application of that exception. Thus, the claims as a whole do not amount to significantly more than the exception itself [Step 2B: NO; See MPEP § 2106.05].
Therefore, the instant claims are not drawn to eligible subject matter as they are directed to one or more judicial exceptions without significantly more. For additional guidance, applicant is directed generally to the MPEP § 2106.
Claim Rejections - 35 USC § 102
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.
Claims 1, 7, 9-16, and 19-20 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by He et al. (bioRxiv, May 2021doi.org/10.1101/2021.05.20.445055, p. 1-18; newly cited).
Claim 1 discloses a method performed by one or more computers. Claim 19 discloses a system comprising: one or more computers; and one or more storage devices communicatively coupled to the one or more computers, wherein the one or more storage devices store instructions that, when executed by the one or more computers, cause the one or more computers to perform operations. Claim 20 discloses one or more non-transitory computer storage media storing instructions that when executed by one or more computers cause the one or more computers to perform operations.
The prior art to He discloses a novel Context-aware Deconfounding Autoencoder (CODE-AE) that can extract common biological signals masked by context-specific patterns and confounding factors (abstract). He teaches that CODE-AE is an algorithm (p. 3, par. 3), the code of which are available online (p. 15, par. 5), which reads on instructions that, when executed by the one or more computers, cause the one or more computers to perform operations as instantly claimed. It is therefore considered the He inherently teaches a method performed by a computer, as in claim 1, a system, as in claim 19, and a non-transitory computer storage media, as in claim 20. He, indicated by the open circles, teaches the instant features, indicated by the closed circles, as follows. Instantly claimed elements which are considered to be equivalent to the prior art teachings are described in bold for all claims.
The operations of claims 1 and 19-20 comprise:
generating a drug signature for a drug, wherein: the drug signature comprises a respective impact score for each of a plurality of features; and the impact score for a feature characterizes an impact, caused by administering a drug to one or more entities, on a value of the feature measured for the one or more entities;
He teaches labeled cell line drug response data (i.e., a drug signature) composed of gene responses (i.e., impact score for each of a plurality of features) in a cell line (i.e., an entity) to a drug (Figure 1; p. 2, par. 3 through p. 3, par. 1; p. 11, par. 1 through p. 13, par. 1 and 4).
generating an embedding of the drug signature in a latent space, comprising: generating a network input to an encoder neural network based on the drug signature; and processing the network input generated based on the drug signature using the encoder neural network to generate the embedding of the drug signature in the latent space; and
He teaches mapping the input of gene expression profile of cell lines into an embedding space (Figure 1). He teaches that the embedding function of CODE-AE projects a high-dimensional expression profile of each cell line to a low-dimensional vector (p. 3, par. 2; p. 11, par. 1 through p. 13, par. 1). He teaches that the embeddings are latent embeddings (p. 7, par. 2) that are in latent space (p. 11, par. 1 through p. 12, par. 2). As He teaches that CODE-AE is an autoencoder (p. 2, par. 3) which contains a neural network (p. 3, par. 3; p. 10 through p. 11, par. 1), it is considered that He fairly teaches an encoder neural network for processing a drug signature input as instantly claimed.
processing: (i) the embedding of the drug signature in the latent space, and (ii) data defining a plurality of patient categories, to generate a plurality of response scores, wherein each response score corresponds to a respective patient category and characterizes a predicted response of patients included in the patient category to the drug.
He teaches that CODE-AE uses a supervised trained model on the embedding of the genes of the cell lines (i.e., (i)), as described above, and an embedding of the genes of patients (i.e., (ii)) to predict drug responses (i.e., response scores) (Figure 1; Table 2). He teaches that for the clinical model, patients were split into two groups, as responders and non-responders (i.e., each response score corresponds to a respective patient category and characterizes a predicted response of patients included in the patient category to the drug) (p. 14, par. 4-6).
Regarding claim 7, He teaches claim 1 as described above. Claim 7 further adds that the drug signature comprises one or more impact scores that each characterize an impact, caused by administering the drug to the one or more entities, on a level of expression of a respective gene in the one or more entities.
He teaches examining labeled cell line drug response data that is the gene expression level (i.e., impact score) of genes in the cell line (Figure 1; p. 13, par. 4).
Regarding claim 9, He teaches claim 1 as described above. Claim 9 further adds that the network input to the encoder neural network includes the drug signature.
He teaches mapping the input of gene expression profile of cell lines (i.e., drug signature) into an embedding space (Figure 1). He teaches that the embedding function of CODE-AE projects a high-dimensional expression profile of each cell line to a low-dimensional vector (p. 3, par. 2; p. 11, par. 1 through p. 13, par. 1).
Regarding claim 10, He teaches claim 1 as described above. Claim 10 further adds that each of the plurality of patient categories is defined by a cluster of patient embeddings in the latent space, wherein each patient embedding corresponds to a respective patient and is generated by processing multi-modal data characterizing the patient using the encoder neural network.
He teaches for the clinical model, patients are split into groups (i.e., clusters) of responders and non-responders to chemotherapy agents (p. 14, par. 4-6). As He teaches embedding the gene expression values of patients into latent space (Figure 1), it is considered that He fairly teaches clusters of embeddings of individual patients as instantly claimed. He also teaches ranking drugs based on their mean predicted responses for each patient cluster (p. 9, par. 2).
Regarding claim 11, He teaches claims 1 and 10 as described above. Claim 11 further adds that for each of the plurality of patient categories, generating the response score for the patient category comprises: determining a respective similarity measure between: (i) the embedding of the drug signature, and (ii) each of one or more patient embeddings in the cluster of patient embeddings defining the patient category; and determining the response score for the patient category based on the similarity measures.
He teaches performing a deconfounding of the distributions of the cell line (i.e., drug signature) and patient gene expression embeddings (Figure 1), which is considered to read on groups or clusters of patient categories as described above. He teaches that the distribution of common embeddings of patients is aligned with that of cell lines (Figure 1) according to different methods. He teaches that CODE-AE-MMD uses maximum mean discrepancy as a distance measurement between the latent representation of cell line and tissue samples, and that CODE-AE-ADV uses adversarial training to the push the representations of the cells line and tissue to be similar to each (p. 12, par. 2 through p. 13, par. 1). Both measures of similarity for the common, deconfounded embeddings read on the instantly claimed similarity measures.
Regarding claim 12, He teaches claim 1 as described above. Claim 12 further adds determining a ranking of the plurality of patient categories based on the response scores.
He teaches ranking drugs based on their mean predicted responses for each patient cluster (p. 9, par. 2), which reads on ranking patient categories as instantly claimed.
Regarding claim 13, He teaches claim 1 as described above. Claim 13 further adds determining that a new patient is included in a patient category of the plurality of patient categories; identifying the response score for the patient category of the new patient; and automatically generating a recommendation for whether the new patient should receive the drug based at least in part on the response score for the patient category of the new patient.
He teaches that when a new patient comes, the drug response can be predicted from the trained cell line model (p. 4, par. 1). He teaches ranking drugs based on their mean predicted responses for each patient cluster (p. 9, par. 2), which reads on generating a recommendation for whether the new patient should receive a drug. As He teaches that CODE-AE is an algorithm (p. 3, par. 3), it is considered that such an algorithm would inherently be performed to do the steps automatically as instantly claimed.
Regarding claims 14-16, He teaches claim 1 as described above. Claim 14 further adds that each of the one or more entities comprises a cell. Claim 15 further adds that each of the one or more entities comprises a collection of cells. Claim 16 further adds that each of the one or more entities is a patient.
He teaches that cell line (i.e., entity comprising a cell as in claim 14; entity comprises a collection of cells in claim 15) and patient (i.e., claim 16) gene expression values are the input to CODE-AE (Figure 1).
Claim Rejections - 35 USC § 103
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.
A. Claims 2-4 and 6 are rejected under 35 U.S.C. 103 as being unpatentable over He, as applied to claim 1 in the above 35 USC 102 rejection, and in view of Shayakhmetov et al. (Frontiers in Pharmacology, 2020, 11(269):1-11; newly cited).
Regarding claims 2-4, He teaches claim 1 as described above. Claim 2 further adds, for each of the entities: (i) a pre-treatment feature representation of the entity that comprises, for each of the plurality of features, a respective pre-treatment value of the feature that is measured for the entity prior to the drug being administered to the entity; and (ii) a post-treatment feature representation of the entity that comprises, for each of the plurality of features, a respective post-treatment value of the feature that is measured for the entity after the drug is administered to the entity; and generating the drug signature based on the pre-treatment and post-treatment feature representations of the entities. Claim 3 further adds generating, for each of the plurality of entities, a respective differential feature representation of the entity based on a difference between: (i) the pre-treatment feature representation of the entity, and (ii) the post-treatment feature representation of the entity; and generating the drug signature based on the differential feature representations of the entities. Claim 4 further adds generating a respective entity-specific drug signature for each of the entities based on the differential feature representation of the entity; and generating the drug signature by combining the entity-specific drug signatures.
He teaches only using labeled cell line drug response data (Figure 1), and therefore does not explicitly teach the claim as recited.
However, Shayakhmetov discloses a generative autoencoder model that infers drug molecules that could induce a desired change in gene expression by separating cellular processes captured in gene expression changes into two feature sets, those related and unrelated to the drug incubation (abstract). Shayakhmetov teaches that the gene expression profiles dataset includes measurements of gene expression before (i.e., (i)) and after (i.e., (ii)) cells react with a molecule at a given concentration (p. 6, col. 2, par. 1). Shayakhmetov teaches encoding Δge, where Δge=gea-geb, gea being the perturbation-induced gene expression profile and geb being the control (p. 6, col. 1, par. 2 and p. 7, col. 1, par. 3). Shayakhmetov teaches a Latent Supervised Adversarial Autoencoder model which uses a differential expression encoder to pass the latent code of gene expression changes (i.e., drug signature) to the decoder of the model (p. 3, col. 1, par. 3 through p. 5, col. 2, par. 2; Figures 2-5). As each of the genes in the gene expression profiles is considered an entity, it is considered that Shayakhmetov fairly teaches generating the differential features for each entity as in claim 3 and generating the drug signature by combining the entity-specific drug signatures in claim 4.
Regarding claim 6, He teaches claim 1 and, in view of Shayakhmetov, claims 2-4 as described above. Claim 6 further adds generating the drug signature by combining the entity- specific drug signatures comprises: averaging the entity-specific drug signatures, which He does not teach.
Shayakhmetov teaches using the gene expression change averaged over tokens log p(x|Δge,η) (p. 7, col. 1, par. 4 through col. 2, par. 1), which reads on averaging the entity-specific drug signatures as instantly claimed.
Regarding claims 2-4 and 6, it would have been prima facie obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine, in the course of routine experimentation and with a reasonable expectation of success, He and Shayakhmetov because each reference discloses methods for using autoencoders to analyze genetic response to drugs. The motivation to examine differential gene expression would have been to examine the changes in gene expression that are common to the transcriptomes and drugs, as taught by Shayakhmetov (p. 2, col. 1, par. 1). Therefore, it would have been obvious to substitute representations of gene expression changes in response to a drug rather than just gene expressions of cell lines with known drug responses into the model of He.
B. Claim 5 is rejected under 35 U.S.C. 103 as being unpatentable over He, as applied to claim 1 in the above 35 USC 102 rejection, in view of Shayakhmetov, as applied to claims 2-4 above, and in view of Regev et al. (US/20220180975; newly cited).
Regarding claim 5, He teaches claim 1 and, in view of Shayakhmetov, claims 2-4 as described above. Claim 5 further adds that generating the entity-specific drug signature for the entity comprises: element-wise dividing the differential feature representation for the entity by the pre-treatment feature representation of the entity, which neither He nor Shayakhmetov teach.
However, the prior art to Regev discloses systems and methods train machine learning methods with intrinsic and extrinsic features of a cell and/or tissue to define transcriptomic profiles of the cell and/or tissue (abstract). Regev teaches that, for each cell type (i.e., entity), provided mean gene-wise expression values were divided by the sum of all mean gene-wise expression values for this cell type [0567].
It would have been prima facie obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine, in the course of routine experimentation and with a reasonable expectation of success, He in view of Shayakhmetov with Regev because each reference discloses methods for using machine learning to analyze transcriptomic data. The motivation would have been to ensure that per cell type, the gene-wise relative frequencies across all genes summed up to 1, as taught by Regev [0567]. It would have been obvious to one of ordinary skill in the art to apply the teachings of Regev for dividing gene-wise expression values by a sum of all mean gene-wise expression values to those of He in view of Shayakhmetov to the pre-treatment feature representation for the entity because one of ordinary skill in the art would understand that such an action would result in a normalization of the values.
C. Claims 8 and 17-18 is rejected under 35 U.S.C. 103 as being unpatentable over He, as applied to claim 1 in the above 35 USC 102 rejection, and in view of Oskooei et al. (US/20200365270; newly cited).
Regarding claim 8, He teaches claim 1 as described above. Claim 8 further adds that the drug signature comprises one or more impact scores that each characterize an impact, caused by administering the drug to the one or more entities, on a level of expression of a respective protein in the one or more entities, which He does not teach.
However, the prior art to Oskooei discloses
Oskooei teaches that neural network models for drug efficacy prediction are generated by processing training data derived from a dataset which correlates measured biomolecular data for disease-cell samples with measured drug efficacy values for a plurality of drugs tested in relation to samples in the dataset, where such datasets may include various types of biomolecular data for disease cells, e.g. gene expression data, protein levels, genomic data, copy number alteration and/or other molecular biomarkers, which provide a biomolecular profile of a disease cell [0037]. Oskooei teaches using the training data pairs and the drug efficacy values corresponding to respective pairs to train weights of the model including a first attention-based encoder which encodes the biomolecular data of each pair, to produce encoded biomolecular data, a second attention-based encoder which encodes the string representation of each pair, to produce encoded string data, and a set of neural network layers which process the encoded biomolecular and string data for each pair, to produce a model output corresponding to drug efficacy (abstract).
Regarding claim 17, He teaches claim 1 as described above. Claim 17 further adds that the encoder neural network has been trained to process multi-modal data characterizing patients, which He does not teach.
However, Oskooei teaches that neural network models for drug efficacy prediction are generated by processing training data derived from a dataset which correlates measured biomolecular data for disease-cell samples with measured drug efficacy values for a plurality of drugs tested in relation to samples in the dataset, where such datasets may include various types of biomolecular data for disease cells, e.g. gene expression data, protein levels, genomic data, copy number alteration and/or other molecular biomarkers, which provide a biomolecular profile of a disease cell [0037] (see also Figure 2-3). Oskooei teaches that the biomolecular data used in training data pairs may comprise data items corresponding to one or more types of biomolecular data, e.g. gene expression values, protein levels, etc., for samples in the training dataset [0038]. As Oskooei teaches that the biomolecular data can be more than one type of data, it is considered that Oskooei fairly teaches multi-modal data as instantly claimed.
Regarding claim 18, He teaches claim 1 and, in view of Oskooei, claim 17 as described above. Claim 18 further adds that the encoder neural network has been trained by operations comprising: obtaining a plurality of training examples, wherein each training example corresponds to a respective patient and includes multi-modal data that characterizes the patient; jointly training the encoder neural network along with a decoder neural network on the plurality of training examples, comprising, for each training example: processing the multi-modal data from the training example using the encoder neural network to generate an embedding of the multi-modal data from the training example; processing the embedding of the multi-modal data from the training example using the decoder neural network to generate a reconstruction of the multi-modal data from the training example; and updating current values of a set of encoder parameters and current values of a set of decoder parameters using gradients of a reconstruction loss function that measures an error in the reconstruction of the multi-modal data from the training example.
He teaches that datasets for training include patient tumor tissue gene expression profiles (p. 13, par. 4). He teaches that the encoder encodes embeddings of the patient tissue data, which is reconstructed by a decoder (p. 11, par. 2). He teaches measuring the quality of the reconstruction through the mean squared error (p. 11, section 4.1.1). He teaches applying a penalty term for difference loss, and calculating the loss for the model using the error loss described above and the difference loss (p. 11, section 4.1.1). He teaches using a Wasserstein generative adversarial networks with gradient penalty (p. 12, par. 3), which read on gradients of a reconstruction loss function as instantly claimed. He teaches training (i.e., updating parameters of the model, which includes the encoder and decoder) CODE-AE with the gene data from cell lines and patients (Figure 1).
He does not teach jointly training using multi-modal data.
However, Oskooei teaches that the biomolecular data used in training data pairs may comprise data items corresponding to one or more types of biomolecular data, e.g. gene expression values, protein levels, etc., for samples in the training dataset [0038], which reads on multi-modal data as instantly claimed as described above. As Oskooei teaches that multiple types of biomolecular data can be encoded and analyzed at once, it is considered that Oskooei fairly teaches jointly training the neural network as instantly claimed.
Regarding claims 8 and 17-18, it would have been prima facie obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine, in the course of routine experimentation and with a reasonable expectation of success, He in view of Oskooei because each reference discloses methods for using encoder based neural networks to analyze biomolecular data to predict drug responses. The motivation to examine other or multiple types of biomolecular data would have been to provide personalized (or precision) medicine where biomolecular data, such as gene expression data or other molecular biomarkers obtained from a patient, may be used to design a personalized course of treatment, as taught by Oskooei [0003].
Conclusion
No claims are allowed.
Inquiries
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JANNA NICOLE SCHULTZHAUS whose telephone number is (571)272-0812. The examiner can normally be reached on Monday - Friday 8-4.
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 on (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 an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see https://ppair-my.uspto.gov/pair/PrivatePair. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/JANNA NICOLE SCHULTZHAUS/Examiner, Art Unit 1685
Prosecution Insights