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 .
Claim Status
Claims 10-11, 14-17, 19-21, 25-26, 28-30, 32, 35-38, and 40 are pending.
Claims 26, 28-32, 35-38, and 40 are withdrawn.
Claims 1-9, 12-13, 18, 22-24, 27, 31, 33-34, 39, and 41-47 are canceled.
Claims 10-11, 14-17, 19-21, and 25 are examined on the merit.
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in
37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 09/30/2025 has been entered.
Priority
Applicant's claim for the benefit of a prior-filed U.S. Provisional Application No. 6 63/013,480 field 04/21/2020 is acknowledged. Accordingly, the effective filing date of the claimed invention is 04/21/2020.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 02/09/2026 is in compliance with the provisions of 37 CFR 1.97. Accordingly, the list of cited references was considered in full by the examiner. A signed copy of the corresponding 1449 form has been included with this Office action.
The following rejections and/or objections are either maintained or newly applied. They constitute the complete set presently being applied to the instant application.
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 10-11, 14-17, 19-21, and 25 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
The Supreme Court has established a two-step framework for this analysis, wherein a claim does not satisfy § 101 if (1) it is “directed to” a patent-ineligible concept, i.e., a law of nature, natural phenomenon, or abstract idea, and (2), if so, the particular elements of the claim, considered “both individually and as an ordered combination,” do not add enough to “transform the nature of the claim into a patent-eligible application.” Elec. Power Grp., LLC v. Alstom S.A., 830 F.3d 1350, 1353 (Fed. Cir. 2016) (quoting Alice, 134 S. Ct. at 2355). Applicant is also directed to MPEP 2106.
Step 1: The instantly claimed invention (claim(s) 10-11, 14-17, 19-21, and 25 being representative) is directed to a method. Therefore, the instantly claimed invention falls into one of the four statutory categories. [Step 1: YES]
Step 2A: First it is determined in Prong One whether a claim recites a judicial exception, and if so, then it is determined in in Prong Two if the recited judicial exception is integrated into a practical application of that exception.
Step 2A, Prong 1: Under the MPEP § 2106.04, the Step 2A (Prong 1) analysis requires determining whether a claim recites an abstract idea, law of nature, or natural phenomenon.
Claim(s) 10-11, 14-17, 19-21, and 25 recite the following steps which fall under the mathematical concepts, mental processes, and/or certain methods of organizing human activity groupings of abstract ideas:
Claim 10 recites filtering, from the dextramer sequence data, based on the single cell sequence data, data associated with low-quality cell; filtering, from the dextramer sequence data, based on the single cell TCR- data, data according to a presence or an absence of an a-chain or a p-chain; and identifying data remaining in the normalized filtered dextramer sequence data as associated with reliable TCR-pMHC binding events; identifying, based on the data remaining in the filtered dextramer sequence data, a plurality of TCR sequences; the limitations filtering low quality data, filtering data according to a presence or an absence of an a-chain or a p-chain is considered a mathematical calculation, as disclosed in specification [00114], “Filter methods include, for example, Pearson's correlation, linear discriminant analysis, analysis of variance (ANOVA), chi-square, combinations thereof, and the like”; the limitation identifying as associated with reliable TCR-pMHC binding events can be practically performed in human mind (mental process) because human mind is able to process/ assess/ filter data in order to reject unwanted data. Furthermore, the steps directed to “identifying” are steps that can be performed mentally because given the plain meaning of the term “identify”, a human mind is capable of looking at data and making an identification based on given parameters.
Claim 10 further recites adjusting, based on a measure of background noise; the limitation adjusting is considered a mathematical calculation, as disclosed in the present specification [0016], adjusting background noise by subtracting Max (P99.9, argmax Ds,u). A such, the adjusting step falls into mathematical concepts grouping of abstract ideas.
Claim 10 further recites generating a one-dimensional input vector comprising an encoded paired αβ chain CDR3, V gene, and J gene segment sequence; the limitation generating a one-dimensional input vector is considered a mathematical calculation, as disclosed in instant specification [00108] “The V and J gene identifiers may be one-hot encoded”, and as such, falls into mathematical concepts groupings of abstract ideas.
Claim 10 further recites generating a training dataset comprising at least the one-dimensional input vectors; the limitation generating a training dataset involves mathematical calculation of generating one-dimensional input vectors, and as such, falls into mathematical concepts groupings of abstract ideas. also, see instant specification [0105-00106] normalizing adjusted dextramer binding signals using an equation that is used as portion of the training dataset. Said limitation also falls into mental process of generating/creating a dataset since human mind is capable of creating dataset after series of data manipulation steps.
Claim 10 further recites training, based on the training dataset, a machine learning model; the limitation training a model is considered a mathematical calculation, as claimed in instant claim, training dataset comprising at least the one-dimensional input vectors, and as disclosed in instant specification [00135] “The model may be trained by minimizing the cross-entropy loss …, [0114]: Training includes backward elimination and greedy optimization algorithm…, [0119]: model trained using a machine learning approach such as discriminant analysis; decision tree; a nearest neighbor (NN) algorithm (e.g., k-NN models, replicator NN models, etc.); statistical algorithm (e.g., Bayesian networks, etc.); clustering algorithm (e.g., k-means, mean-shift, etc.); neural networks (e.g., reservoir networks, artificial neural networks, etc.); support vector machines (SVMs); logistic regression algorithms; linear regression algorithms; Markov models or chains; principal component analysis (PCA) (e.g., for linear models); multi-layer perceptron (MLP) ANNs (e.g., for non-linear models); replicating reservoir networks (e.g., for non-linear models, typically for time series); random forest classification; a combination thereof and/or the like”. As such, said limitation falls into mathematical concepts groupings of abstract ideas.
Claim 10 further recites training a machine learning model to predict one or more TCR-pMHIC binding events; the claim does not provide any details about how the trained model operates; As stated above, the limitation of training, as claimed in instant claim and in light of specification, encompasses specific mathematical calculation to perform the training and therefore, encompasses mathematical concepts, see also the instant specification [00139]: the machine learning classifier may predict a likelihood that the TCR sequence will bind to one or more specific peptides; further, the limitation to predict events, given the plain meaning of “predict” encompasses mental observations or evaluations, e.g. a person predicting an event based on the result of a mental and/or mathematical analysis, and as such, falls into mental processes of abstract ideas. See MPEP 2106.04(a)(2), subsection III.
Claim 11 recites calculating a number of genes, and calculating a fraction of mitochondrial gene expression; the limitation calculating falls into mathematical concepts groupings of abstract ideas.
Claim 11 further recites removing genes outside a threshold, and removing genes expression data that exceeds a gene expression threshold; the limitations removing is considered a mathematical relationship between variables and numbers, and as such, falls within mathematical concepts grouping of abstract ideas.
Claim 15 recites calculating a maximum negative control dextramer signal, calculating a maximum sorted dextramer signal; calculating a maximum unsorted dextramer signal; said limitations are considered mathematical calculation, as disclosed in present specification [0092-0093]. As such, said limitations fall within mathematical concepts grouping of abstract ideas.
Claim 16 recites estimating dextramer binding background noise; estimating a dextramer sorting gate efficiency; determining measure of background noise; and subtracting the measure of background noise from a dextramer signal associated with each cell; the limitations estimating, calculating, and subtracting are considered mathematical calculations, as disclosed in the present specification [0091-0093], and as such, fall within mathematical concepts groupings of abstract ideas.
Claim 19 recites performing, for each cell represented in the dextramer sequence data, cell-wise and normalization on the dextramer signals associated with each cell; and performing, for each cell represented in the dextramer sequence data, pMHC-wise normalization; the limitations performing normalization is considered mathematical calculations, as disclosed in present specification [0104-00105], and as such, falls within mathematical concepts groupings of abstract ideas.
Claim 20 recites identifying a presence or an absence of at least one a-chain and at least one B-chain; the limitations identifying, given the plain meaning of “identifying” can be practically performed in human mind (mental process), since human mind is capable of identifying presence or absence of data.
Claim 21 recites predicting a binding status of a newly presented receptor sequence according to the trained machine learning model; determining, by the machine learning model, based on the subject TCR sequence data, a subject TCR binding pattern; and determining, based on a repository of antigen locations and the subject TCR binding pattern, a likelihood that a subject associated with the TCR sequence data has traveled to one or more locations; the limitations predicting and determining can be practically performed in human mind, since human mind is capable of predict a status based on the result of an algorithm/model, presenting data to an algorithm/model, determine a pattern based on the result of a mathematical calculation.
Claim 21 further recites determining a likelihood that a subject associated with the TCR sequence data has traveled to one or more locations; the limitation determining a likelihood is considered a mathematical calculation of a calculating a probability of an event, and as such, falls within mathematical concepts groupings of abstract ideas.
Claim 25 recites generating, based on the data remaining in the filtered dextramer sequence data associated with reliable TCR-pMHC binding events, a TCR binding pattern for a subject; identifying, based on the second single cell sequence data, second dextramer sequence data, and second single cell T Cell Receptor (TCR) sequence data for the subject, a second TCR binding pattern; and identifying, based on a comparison of the TCR binding pattern for the subject and the second TCR binding pattern, the subject; the limitations generating, determining, and identifying can be practically performed in human mind (mental process), since human mind is capable of generating a pattern, determining a second data, identifying based on a comparison.
The identified claims recite a law of nature, a natural phenomenon (product of nature) or fall into one of the groups of abstract ideas of mathematical concepts, mental processes, and/or certain methods of organizing human activity for the reasons set forth above. Therefore, the claims are directed to a judicial exception and require further analysis in Prong Two. [Step 2A, Prong 1: YES]
Step 2A: Prong 2: Under the MPEP § 2106.04, the Step 2A, Prong 2 analysis requires identifying whether there are any additional elements recited in the claim beyond the judicial exception(s), and evaluating those additional elements to determine whether they integrate the exception into a practical application of the exception. This judicial exception is not integrated into a practical application for the following reasons.
The additional elements of claims include the following:
Claim 1 recites receiving single cell sequencing data comprising single cell sequence data, dextramer sequence data, and single cell T-Cell Receptor (TCR) sequence data; generating/outputting, by the trained machine learning model, a TCR-pMHC binding affinity map.
Claim 21 recites the trained machine learning model; presenting, to the machine learning model, subject TCR sequence data.
Claim 25 recites receiving, at a subsequent point in time, second single cell sequence data, second dextramer sequence data, and second single cell T Cell Receptor (TCR) sequence data for the subject.
The additional elements of receiving data serve to collect the information for use by the abstract idea and do not impose any meaningful limitation on the abstract idea, or on how the abstract idea is performed. Data gathering steps are not sufficient to integrate an abstract idea into a practical application. Similarly, the additional element of presenting/inputting, to the trained machine learning model, a TCR sequence, and generating/outputting a binding affinity map amount to necessary data gathering and outputting, therefore, considered extra solution activities. the additional elements of receiving and generating data are activities incidental to the primary process or product (all uses of the judicial exception require such data gathering or data output) that are merely a nominal or tangential addition to the claim and they amount to necessary data gathering and outputting. These additional elements are considered insignificant extra-solution activities. As explained by the Supreme Court, the addition of insignificant extra-solution activity does not amount to an inventive concept. See MPEP 2106.05(g)(3).
Furthermore, the additional element of the trained machine learning model of claim 10 and 21 merely indicates a field of use or technological environment in which the judicial exception is performed. See MPEP 2106.05(h). Said limitation can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to a particular field of use or a technological environment.
Therefore, these additional elements represent mere instructions to implement an abstract idea or other exception on a computer and amount to insignificant extra-solution activity, which is not sufficient to integrate the recited judicial exception into a practical application. See MPEP 2106.05(g). [Step 2A, Prong 2: NO]
Step 2B: In the second step it is determined whether the claimed subject matter includes additional elements that amount to significantly more than the judicial exception. An inventive concept cannot be furnished by an abstract idea itself. See MPEP § 2106.05.
The claims do not include any additional steps appended to the judicial exception that are sufficient to amount to significantly more than the judicial exception.
As set forth in the MPEP, determinations of whether or not additional elements (or a combination of additional elements) may provide significantly more and/or an inventive concept rests in whether or not the additional elements (or combination of elements) represents well-understood, routine, conventional activity. Said assessment is made by a factual determination stemming from a conclusion that an element (or combination of elements) is widely prevalent or in common use in the relevant industry, which is determined by either : (1) 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); (2) 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); (3) a citation to a publication that demonstrates the well-understood, routine, conventional nature of the additional element(s); and/or (4) a statement that the examiner is taking official notice with respect to the well-understood, routine, conventional nature of the additional element(s).
The additional elements of claims include the following:
Claim 1 recites receiving single cell sequencing data comprising single cell sequence data, dextramer sequence data, and single cell T-Cell Receptor (TCR) sequence data; generating/outputting, by the trained machine learning model, a TCR-pMHC binding affinity map.
Claim 21 recites the trained machine learning model; presenting, to the machine learning model, subject TCR sequence data.
Claim 25 recites receiving, at a subsequent point in time, second single cell sequence data, second dextramer sequence data, and second single cell T Cell Receptor (TCR) sequence data for the subject.
The additional elements of receiving first and second data, are conventional computer components and/or processes. The courts have found the use of a computer or other machinery in its ordinary capacity for economic or other tasks (e.g., to receive, store, or transmit data) or simply adding a general-purpose computer or computer components after the fact to an abstract idea (e.g., a fundamental economic practice or mathematical equation) does not provide significantly more. See Affinity Labs v. DirecTV, 838 F.3d 1253, 1262, 120 USPQ2d 1201, 1207 (Fed. Cir. 2016) (cellular telephone); TU Communications LLC v. AV Auto, LLC, 823 F.3d 607,613,118 USPQ2d 1744, 1748 (Fed. Cir. 2016) (computer server and telephone unit). See MPEP 2106.05 (d) II. i.
Furthermore, the additional elements of receiving, presenting/inputting and generating/outputting amount to necessary data gathering and outputting. The courts have found the limitations that amount to necessary data gathering and outputting are insignificant extra-solution activity that do not amount to significantly more (see MPEP 2106.05(g)).
Furthermore, the additional element of generating, by the trained machine learning model of claim 10 and the trained machine model of claim 21 merely indicates a field of use or technological environment in which the judicial exception is performed. See MPEP 2106.05(h). Said limitation can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to a particular field of use or a technological environment.
Therefore, the additional element is not sufficient to amount to significantly more than the judicial exception. See MPEP 2106.05(g).
Applicant is reminded that the judicial exception alone cannot provide the inventive concept or the practical application and that the identification of whether the additional elements amount to such an inventive concept requires considering the additional elements individually and in combination to determine if they provide significantly more than the judicial exception. (MPEP 2106.05.A i-vi).
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].
Therefore, the instantly rejected claims are not drawn to eligible subject matter as they are directed to an abstract idea without significantly more.
Response to Applicant’s Arguments
Applicant's arguments filed 09/30/2025 have been fully considered but they are not persuasive.
Applicant states:
The training of a machine learning model is not an abstract idea per the MPEP
Similar to the example from MPEP § 2106.04(a)(1), claim 10 recites steps for receiving input data: receiving single cell sequencing data comprising single cell sequence data, dextramer sequence data, and single cell T-Cell Receptor (TCR) sequence data;
Also similar to the example from MPEP § 2106.04(a)(1), claim 10 recites steps for training the model: training, based on the training data set, a machine learning model to predict one or more TCR-pMHC binding events;
Thus, similar to the example method of training a neural network in the MPEP, claim 10 recites a method of training a "machine learning model." That is, the example claim and present claim 10 both use machine learning concepts to train machine learning models. Neither the example claim nor present claim 10 recite (set forth or describe) an abstract idea.
In the Response to Arguments section at page 12, the Office Action states "Example 39 is directed to training a neural network for facial detection. In contrast, the instant claims do not recite a neural network nor do the claims include any analogous steps directed to a practical application of analyzing facial images." Applicant respectfully disagrees.
Firstly, while claim 10 does not recite "a neural network," claim 10 does recite "a machine learning model.
Secondly, contrary to the assertion of the Office Action, the example in the MPEP does not "include... steps directed to a practical application of analyzing facial images" and thus, it is improper to require claim 10 to include such features. As stated above, the MPEP recites this example as an example claim which does not recite an abstract idea. Id. The example is "eligible at Step 2A Prong One" and never reaches the question of a practical application at Prong Two.
Additionally, Applicant notes that the Deputy Commissioner of Patents issued a Memorandum to Technology Centers 2100, 2600, and 3600 on August 4, 2025 ("Deputy Commissioner Memorandum") further detailing examination guidelines for 35 U.S.C. § 101, which is attached as Appendix A for your reference. In the Deputy Commissioner Memorandum, Examiners are reminded to "distinguish claims that recite an exception (which require further eligibility analysis) from claims that merely involve an exception (which are eligible and do not require further eligibility analysis)." Deputy Commissioner Memorandum, p. 3. The Deputy Commissioner Memorandum (p. 3) explicitly states that Example 39 "illustrates claim limitations that merely involve an abstract idea" and does not recite a judicial exception.
It is respectfully submitted that the above statements are not persuasive. First, in contrast to MPEP § 2106.04(a)(1) training a machine learning model example’s input data of digital facial image (not an abstract idea), instant claims input data comprises sequencing data (data consists of letters and/or numbers) (abstract idea). Second, the instant claims require only data manipulation/preprocessing of such data (mantal processes and/or mathematical calculation/mathematical concepts), for example filtering and adjusting, … and generating a one-dimensional input vector (mathematical calculation/mathematical concepts), and inputting such preprocessed data into a machine learning model to generate an affinity map/ output. The process of learning is an iterative process of performing mental and/or mathematical steps for the purpose of evaluating information. Additionally, as stated above, the recitation of using a “trained machine learning model” to generate an output, merely indicates a field of use or technological environment in which the judicial exception is performed. See MPEP 2106.05(h). Said limitation can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to a particular field of use or a technological environment.
It is noted that the Office is currently applying updated framework assessment as per the 2019 Guidance and its updates as recently as July 2024. Instant application is more similar to Example 47, which is directed to training a neural network that is used to determine and evaluate anomaly in data. In Example 47, claim 2, the claimed discretizing, detecting, and analyzing steps were found to encompass mental choices or evaluations, and the claimed discretizing and training using a backpropagation algorithm and gradient descent algorithm were found to encompasses performing mathematical calculations. Similar to example 47, instant claimed filtering data, adjusting data, identifying data, generating data, and training a model encompass performing mental processes and/or mathematical calculations (concepts).
With regards to Applicant referring to August 2025 Memorandum, stating that Examiners are reminded to "distinguish claims that recite an exception (which require further eligibility analysis) from claims that merely involve an exception (which are eligible and do not require further eligibility analysis), Examiner submits that instant claims clearly recite the exceptions of filtering, adjusting, generating a one-dimensional input vector, and training the model, which are words to describe mathematical calculations, as well as mental processes of identifying and predicting (See above rejection for more details), which are all considered abstract ideas.
Applicant further states:
The steps of generating training data and generating, by the trained model, a TCR-pMHC binding affinity map are not an abstract idea
The Office Action alleges that "generating, for each TCR sequence of the plurality of TCR sequences, a one-dimensional input vector comprising an encoded paired ap chain CDR3 amino acid sequence, V gene segment sequence, and J gene segment sequence," "generating a training data set comprising at least the one-dimensional input vectors," and "training, based on the training data set, a machine learning model to predict one or more TCR-pMHC binding events," are limitations which fall under the mathematical calculations grouping of abstract ideas. Applicant notes that claim 10 as amended further recites "presenting, to the trained machine learning model, a TCR sequence," and "generating, by the trained machine learning model and based on the TCR sequence, a TCR-pMHC binding affinity map."
As noted above, the Deputy Commissioner Memorandum reminds Examiners to
"distinguish claims that recite an exception (which require further eligibility analysis) from claims that merely involve an exception (which are eligible and do not require further eligibility analysis." The above listed elements of claim 10 may "involve" mathematical concepts, but they do not "recite" mathematical concepts. None of these elements are per se mathematical relationships, mathematical formulas or equations, or mathematical calculations. As reasoned by the Deputy Commissioner Memorandum at page 3, the present claims do not "require specific mathematical calculations" and do not "refer[] to [any] mathematical calculations by name," and therefore do not recite a judicial exception.
It is respectfully submitted that this statement is not persuasive. Instant claims clearly recite the exceptions of filtering, adjusting, generating a one-dimensional input vector (for example, specific mathematical calculations), and training the model, which are word to describe mathematical calculations, as well as mental processes of identifying and predicting (See above rejection for more details), which are all considered abstract ideas.
Applicant further states:
In addition, these steps cannot reasonably be performed in the human mind. For example, "generating, for TCR sequence of the plurality of TCR sequences, a one-dimensional input vector comprising an encoded paired ap chain CDR3 amino acid sequence, V gene segment sequence, and J gene segment sequence," "training, based on the training data set, a machine learning model to predict one or more TCR-pMHC binding events," "presenting, to the trained machine learning model, a TCR sequence," and "generating, by the trained machine learning model and based on the TCR sequence, a TCR-pMHC binding affinity map," cannot reasonably be performed in the human mind. Indeed, the elements "training... a machine learning model to predict one or more TCR-pMHC binding events," "presenting, to the trained machine learning model, a TCR sequence," and "generating, by the trained machine learning model and based on the TCR sequence, a TCR-pMHC binding affinity map," necessarily requires a machine learning model, which, by definition, cannot be reduced to pen and paper. Furthermore, none of the elements of the claims recite any method of organizing human activity, such as a fundamental economic concept or managing interactions between people.
At least these elements of claim 10, when considered together as a combination, are beyond the judicial exception of mathematical concepts or mental processes. Thus, claim 10 of the present application is patent-eligible because it does not recite, nor is it directed to, a judicial exception. Accordingly, for at least these reasons, Applicant submits claim 10 is directed to patent-eligible subject matter. Therefore, Applicant respectfully requests that the rejection of claim 10, as well as its dependent claims, be withdrawn and that such claims be allowed.
It is respectfully submitted that the above statements are not persuasive. As stated above, the steps of generating and training, as noted by the applicant above, are considered mathematical calculations, not mental processes.
Additionally, the steps of presenting/inputting and generating/outputting, as noted by the applicant above, are considered additional elements that are considered extra solution activities. Moreover, the recitation of using a “trained machine learning model” to generate an output, merely indicates a field of use or technological environment in which the judicial exception is performed. See MPEP 2106.05(h). Said limitation can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to a particular field of use or a technological environment. These additional elements do not integrate the judicial exception(s) into a practical application in Step 2A Prong Two, nor they amount to significantly more in Step 2B of 101 analysis.
Further, with regard to the Applicant assertion that these steps cannot reasonably be performed in the human mind, Examiner submits that whether the human mind is equipped to perform a task is not linked to the scope of the task. There is not a threshold at which point performing mental and/or mathematical processes graduates from what can be performed by the mind to not performable by the human mind. While this may take a long time, the use of a physical aid, such as a pen-and-paper or computer, may accelerate this process, and this does not negate the mental and/or mathematical nature of the limitation. The courts do not distinguish between mental processes that are performed entirely in the human mind and mental processes that require a human to use a physical aid (e.g., pen and paper or a computer) to perform the claim limitation. See MPEP 2106.04(a)(2) III. B. & C.
Applicant will further note that complexity of operations does not equate to eligibility. The fact remains that the steps are directed to operations that are mental and/or mathematical as stated above.
Therefore, the steps of filtering, adjusting, identifying, generating a one-dimensional input vector, generating a dataset comprising one-dimensional input vector, and to predict are abstract ideas.
Applicant further states:
b. The claims integrate any alleged judicial exception into a practical
application (Prong Two of Sten 2A)
"Receiving single cell sequencing data," "generating... a one- dimensional input vector," "generating a training data set," "training... a machine learning model," "presenting to the trained machine learning model a TCR sequence," and "generating... a TCR-pMHC binding affinity map," are additional elements beyond the alleged judicial exception.
As noted above and for at least the reasons stated in the Deputy Commissioner Memorandum, even if these elements of claim 10 "involve" mathematical concepts, they do not "recite" mathematical concepts. Thus, at least these elements of claim 1, when considered together as a combination, are beyond the judicial exception of mathematical concepts or mental processes and constitute additional elements to be considered in the determination of whether the claim integrates any alleged judicial exception into a practical application.
For at least these reasons, Applicant respectfully submits that the above-mentioned elements should be considered additional elements beyond the alleged judicial exception.
It is respectfully submitted that the above statements are not persuasive. As stated above, the steps of "generating... a one- dimensional input vector" ,"generating a training data set" and "training... a machine learning model" are abstract ideas, not additional elements. Furthermore, the additional elements of receiving data, presenting/inputting data, and generating a map/outputting are considered extra solution activities that do not integrate the judicial exception(s) into a practical application in Step 2A Prong Two, nor they amount to significantly more in Step 2B of 101 analysis.
Applicant further states:
The claims provide improvements to technology that relate to the identification and prediction of reliable TCR-pMHC bindings.
Applicant submits the present claims provide improvements to technology that relate to the identification and prediction of reliable TCR-pMHC bindings. The present Specification explains that "highly multiplexed dextramer binding data are often associated with low signal-to- noise ratios [making] it bioinformatically challenging to reliably identify TCR-pMHC binding events using such large-scale binding datasets." US20210335447A1, at [0004]. The Specification also notes that "[w]hile the results from initial TCR-pMHC binding classifiers are encouraging, they were only trained using CDR loop sequences and thus unable to learn the overall complex sequence patterns from full-length TCR sequences, resulting in sub-optimal prediction accuracy for highly diverse pMHC binding TCRs." Id, at [0005].
Thus, as Example 48 (Speech Separation, Claim 2) was found eligible due to its
structured data processing pipeline and generation of data, the present claim should also be deemed eligible because it provides a structured and standardized training dataset for training a
model and using that model to generate a TCR-pMHC binding affinity map that significantly enhances TCR binding affinity predictions. The training data structure itself as well as the resultant map reflect a technical advancement, ensuring the claim is not directed to an abstract idea and instead integrates a practical application into biological AI-driven predictions. Accordingly, for at least these reasons, Applicant submits claim 10 is directed to patent-eligible subject matter. Therefore, Applicant respectfully requests that the rejection of claim 10, as well as its dependent claims, be withdrawn and that such claims be allowed.
It is respectfully submitted that this is not persuasive. The Applicant remarks are directed to Step 2A Prong Two of 101 analysis, specifically whether the additional elements integrate the recited judicial exception into a practical application of the exception.
As noted by the Applicant, instant claimed invention is directed to the identification and prediction of reliable TCR-pMHC bindings (abstract ideas). It is important to note, the judicial exception alone cannot provide the improvement (See MPEP 2106.04(d) III). The improvement must be provided by one or more additional elements. See the discussion of Diamond v. Diehr, 450 U.S. 175, 187 and 191-92, 209 USPQ 1, 10 (1981)) in subsection II, below. In addition, the improvement can be provided by the additional element(s) in combination with the recited judicial exception. See MPEP § 2106.04(d) (discussing Finjan, Inc. v. Blue Coat Sys., Inc., 879 F.3d 1299, 1303-04, 125 USPQ2d 1282, 1285-87 (Fed. Cir. 2018)).
With regards to applicant stating analogy to claim 2 of Example 48, Examiner submits that: First, Example 48 is directed to a speech separation method. Second, claim 2 of Example 48 was found eligible because in addition to judicial exceptions of steps (b)-(e), the additional elements of steps (f) synthesizing speech waveforms, and step (g) combining the speech waveforms to generate a mixed speech signal reflected the technical improvement. In contrast to claim 2 of Example 48, the only additional elements of the instant application is “receiving … sequence data”, presenting/inputting, and generating a map/outputting, which amount to mere data gathering incidental to the judicial exceptions and thus are insignificant extra-solution activity. See MPEP 2106.05(g).
Applicant further states:
b. The claim as a whole provides improvements to technology that relate to the identification and prediction of reliable TCR-pMHC bindings
The present claims leverage advanced computational techniques to process and analyze high-dimensional biological data specifically, single cell sequencing data that includes T-cell receptor (TCR) sequencing data and dextramer sequencing data - to generate a TCR-pMHC binding affinity map for more rapid and accurate prediction of immune response outcomes. Moreover, the claims reflect an improvement in training machine learning models that utilize such high-dimensional biological data but require that the data be reliable in order for the resultant model predictions to be biologically meaningful.
Accordingly, claim 10 recites the creation of a training data set using TCR sequences from the reliable data that are represented as one-dimensional input vectors, training a machine learning model based on that training data set, and using the model to generate a binding affinity map.
As described in the present specification, a computational improvement realized by the present invention includes "lower dimensional representations" for genes which results in "improved model regularization" and "stronger gene associations." US20210335447A1, 128.
When data is represented as a one-dimensional vector, convolutional operations, such as those used in a Convolutional Neural Network (CNN), can be applied more efficiently. 1D convolutions specifically benefit from this, as they involve fewer parameters and computations compared to higher-dimensional convolutions. This reduction in computational overhead directly contributes to faster model training and prediction ("improved model regularization and force the model to learn stronger gene associations" Id.). In addition, one-dimensional vectors consume less memory compared to higher-dimensional arrays. Efficient memory usage not only speeds up computation but also allows the model to handle larger datasets and more complex models without running into memory constraints. By reducing the data to a one-dimensional representation, the model is forced to focus on the most relevant features, improving regularization. Better regularization leads to a model that generalizes well to new data and is thus able to generate more accurate binding affinity maps in response to receipt of a new TCR sequence. This reduces the need for extensive computational resources during training.
Another technological improvement realized by the present claims resulting in a practical application is the identification and prediction of reliable TCR-pMHC bindings which are critical for developing targeted immunotherapies but are hindered by high-dimensional data complexity and low signal-to-noise ratios.
The present claims solve this technical problem by implementing a method that includes specific steps to remove noise and improve the reliability of a TCR-pMHC binding prediction model which is then able to generate a TCR-pMHC binding affinity map with improved accuracy. This is not a trivial application of computing technology but a sophisticated integration of computational biology techniques specifically designed to handle the complexities of immunological data. These steps are specified in the claims and directly address the noted technical problems by improving data integrity and analysis accuracy, which are essential for reliable biological insights and therapeutic developments.
The use of tailored algorithms and data filtering techniques allows for handling and analyzing biological data in a way that is both computationally efficient and scientifically meaningful, leading to more accurate predictions of immune interactions.
" Generating one-dimensional input vectors ensures consistency and standardization of training data, facilitating improved pattern recognition and generalization across diverse TCR sequences. Further, training the machine learning model on these one-dimensional vectors allows the model to generate improved predictions for TCR sequences.
At least these features of claim 10 clearly indicate the alleged judicial exception has been integrated into a practical application by applying/using the judicial exception in a manner that imposes a meaningful limit on its use such that the claim is more than a drafting effort designed to monopolize the judicial exception. The features of claim 10 impose meaningful limits on the use of the alleged judicial exception, because their recitation in the claim ensures that no mathematical concept or steps of evaluating, analyzing, or organizing information are being monopolized by the claim. In other words, the elements of claim 10 noted above prevent the claim from anticipating or foreclosing all uses of machine learning in the realm of TCR-pMIC interaction data.
It is respectfully submitted that this is not persuasive. Taken as a whole, the claims are directed to judicial exceptions of preprocessing data (filtering, adjusting, identifying) and generating a dataset and training a model to predict events based on the training dataset. As stated above, the steps of filtering, adjusting, identifying, generating a one-dimensional input vector, generating a dataset, and training are judicial exceptions. It is important to note, the judicial exception alone cannot provide the improvement (See MPEP 2106.04(d) III). The improvement must be provided by one or more additional elements. See the discussion of Diamond v. Diehr, 450 U.S. 175, 187 and 191-92, 209 USPQ 1, 10 (1981)) in subsection II, below. In addition, the improvement can be provided by the additional element(s) in combination with the recited judicial exception. See MPEP § 2106.04(d) (discussing Finjan, Inc. v. Blue Coat Sys., Inc., 879 F.3d 1299, 1303-04, 125 USPQ2d 1282, 1285-87 (Fed. Cir. 2018)). The additional element of claim 10 is receiving sequence data, presenting/inputting data, generating a map/outputting, which merely amount to data gathering and outputting, and as such, extra solution activity that does not integrate any judicial exceptions into a practical application. further the additional element of using “the trained machine learning model” merely indicates a field of use or technological environment in which the judicial exception is performed.
With regards to Applicant stating that “the one-dimensional vectors consume less memory compared to higher-dimensional arrays”, Examiner states that instant claims do not recite any memory or computer components to be evaluated in the rejection.
With regards to applicant stating that “The use of tailored algorithms and data filtering techniques allows for handling and analyzing biological data in a way that is both computationally efficient and scientifically meaningful, leading to more accurate predictions of immune interactions”, Examiner states that “tailored algorithms and data filtering techniques” are judicial exceptions. At best, the instant claim invention can be viewed as an improved judicial exception(s) that cannot provide the improvement (see above).
With regards to Applicant stating that “training the machine learning model on these one-dimensional vectors allows the model to generate improved predictions for TCR sequences. At least these features of claim 10 clearly indicate the alleged judicial exception has been integrated into a practical application by applying/using the judicial exception in a manner that imposes a meaningful limit on its use such that the claim is more than a drafting effort designed to monopolize the judicial exception.”, Examiner submits that the steps of presenting/inputting and generating/outputting are considered additional elements that are considered extra solution activities. Moreover, the recitation of using a “trained machine learning model” to generate an output, merely indicates a field of use or technological environment in which the judicial exception is performed. See MPEP 2106.05(h). Said limitation can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to a particular field of use or a technological environment. Therefore, these additionally recited elements do not meaningfully limit the recited abstract ideas. These additional elements do not integrate the judicial exception(s) into a practical application in Step 2A Prong Two of 101 analysis.
Therefore, the instantly rejected claims are not drawn to eligible subject matter as they are directed to an abstract idea (and/or natural correlation) without significantly more.
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.
Claims 10, 14-15, 19 are rejected under 35 U.S.C. 103 as being unpatentable over Fischer et al. (Predicting antigen specificity of single T cells based on TCR CDR3 regions, bioRxiv, August 14, 2019, pages 1-26), as evidence by 10X Genomics (A new way of exploring immunity: linking highly multiplexed antigen recognition to immune repertoire and phenotype, 10xgenomics, 2019, pages 1-16) in view of Bulik-Sullivan et al. (US 11,264,117 B2).
Regarding claim 10, Fischer teaches a high throughput single-cell, multiple sequence-data deep learning method to assign TCR to epitope specificity (abstract). Fischer further teaches a data set based on single-cell pMHC capture in which paired ɑ and β-chain reconstructed for 10,000s of cells and binding-specificity measured for 44 distinct pMHC complexes. Fischer further teaches receiving dextramer sequencing data as evidenced by 10X Genomics (Section: Results: pg. 2, para. 2); reading on limitations of receiving single cell sequencing data comprising single cell sequence data, dextramer sequence data, and single cell T-Cell Receptor (TCR) sequence data.
Fischer further teaches removing putative doublets from the data set (Section: Cell-specific covariates improve binding event prediction, pg. 2, para. 3); reading on limitations of filtering, from the dextramer sequence data, based on the single cell sequence data, data associated with low-quality cells
Fischer further disclose that an alternative interpretation of the improved performance of multi-task models is their ability to learn better de-noised low-dimensional representations of TCR sequence (pg. 4, para. 2); reading on limitations of adjusting, based on a measure of background noise, the dextramer sequence data.
Fischer further discloses designing a model to predict TCR-antigen binding based on ɑ- and β-chain sequences and cell-specific covariates accounting for the variability found in data including chains (Fig. 1b) (Results). Fischer further discloses a feedforward neural network to predict an antigen specificity of a T-cell based on the alpha and beta chain sequences, which “filters” for the specific amino acid sequences within the CDR3 regions of alpha and beta chains of T-cell receptor (Figure 1, pg. 8, para. 1). Fischer further discloses using paired TCR ɑ- and β-chain while determining the T-cell specificity to train multiple deep learning (pg. 1, last para.; see also, pg. 2 Results). Fischer further discloses identifying TCR sequences after applying filters and by focusing on beta chain and CDR3 sequences (pg. 23, L494-500). Fischer further discloses removing all cellular barcodes that contain more than one ɑ- or β-chain as mature CD8+ T cells are expected to only have a single functional ɑ- and β-chain allele (pg. 22, L. 455-460); reading on limitations of filtering, from the dextramer sequence data, based on the single cell TCR- data, data according to a presence or an absence of an a-chain or a b-chain by removing data associated with cells having one or more alpha chains, one or more beta chains, or multiple alpha or beta chains.
Fischer further discloses that a newly developed single-cell technology that enables the simultaneous sequencing of the paired TCR ɑ- and β-chain while determining the T-cell specificity to train multiple deep learning architectures modeling the TCR-pMHC interaction including both chains (Introduction, first page, last para.); reading on limitations of identifying data remaining in the filtered dextramer sequence data as associated with reliable TCR-pMHC binding events.
Fischer further discloses identifying TCR sequences after applying filters and by focusing on beta chain and CDR3 sequences and using such data as training data set (pg. 23, L494-503). Fischer further discloses predicting antigen specificity based on TCR sequences from single-cell data (pg. 5, L. 160-163); reading on limitations of identifying, based on data remaining in in the filtered dextramer sequence data, plurality of TCR sequences.
Fischer further discloses models fit on both the CDR3 loop of ɑ- and β-chain of the TCR (Fig. 1b) and models fit on the CDR3 loop of the β-chain and the antigen sequence (Fig. 2a). Fischer further discloses that the input vector comprises V and J gene segment sequence (using 10x Genomics dataset inherently includes VDJ junction information; see 10x Genomics, pg. 3, col. 2: “The Chromium Single Cell Immune Profiling workflow with Feature Barcode technology generates Single Cell 5’Gene Expression, V(D)J, and Cell Surface Protein libraries… Chromium Single Cell V(D)J enriched libraries, 5’ Gene Expression libraries, and Cell Surface Protein libraries were quantified, normalized, and sequenced according to the User Guide”). Fischer further discloses integrating two sequences and using separate sequence-embedding layer stacks for each sequence (all models presented in Fig. 1 and models indicated as “separate” in Fig. 2) or by appending the two padded sequences and using a single sequence-embedding layer stack (models indicated as “concatenated” in Fig. 2) (pg. 20, L. 378-386). Fischer further discloses using one-hot encoding and using 1 by 1 convolution, to represent sequences numerically and dimensionality reduction of multi-dimensions (Supp. Figure 3). Fischer further discloses training antigen specificity models based on TCR sequences from single-cell data (pg. 5, L. 142-155 and 160-170). Fischer further discloses padding each CDR3 sequence to a length of 40 amino acids and concatenated these padded chain observations to a sequence of length 80 for models that were trained on both chains (pg. 22, L. 464-466). Fischer further discloses generating negative samples for both training and test set separately by generating unobserved pairs of TCR and antigens (pg. 23, L. 498-500); reading on limitations of generating a one-dimensional input vector comprising an encoded paired alpha-beta chain CDR3 amnio acid sequence and generating a training dataset and training a model.
Fischer further discloses that an alternative interpretation of the improved performance of multi-task models is their ability to learn better de-noised low-dimensional representations of TCR sequences, through the integration of more diverse training data (pg. 4, L. 120-124); Fischer further discloses predicting a binding events and binding affinity based on TCR CDR3 sequences and pMHC count (Fig. 2, P.9, Pg. 3, last para.). Fischer further discloses TCR-pMHC modeling including both chains (pg. 1, last para.); reading on limitations of training, based on the training data set, a machine learning model to predict one or more TCR-pMHIC binding events: presenting, to the trained machine learning model, a TCR sequence; and generating, by the trained machine learning model and based on the TCR sequence, a TCR-pMHIC binding affinity map.
Further regarding claim 10, Fischer does not expressly disclose a binding affinity map. Bulik-Sullivan discloses a method for identifying neoantigens where peptide sequences and the associated k-mer blocks are input into a machine-learned presentation model to generate presentation likelihoods for the tumor neoantigens, each presentation likelihood representing the likelihood that a neoantigen is presented by an MHC allele on the surfaces of the tumor cells of the subject (abstract). Bulik-Sullivan further discloses that the peptide sequence of each of the neoantigens is encoded into a corresponding numerical vector, each numerical vector containing information about a plurality of amino acids constituting the peptide sequence and the set of positions of those amino acids in the peptide sequence; using a computer processor, the numerical vectors and one or more hotspot features are input into a machine learning presentation model to generate a presentation probability set of the neoantigen set; the training peptide sequence is encoded into a numerical vector containing information about the plurality of amino acids that make up the peptide and the location set of the amino acids in the peptide; the method of inputting the numerical vector into the machine learning presentation model includes: applying the machine learning presentation model to the peptide sequence of the neoantigen to generate a dependency score for each of the one or more MHC pairs, transforming the combination of the dependency scores to generate the presentation probability; The training dataset also includes at least one of the following: (a) data relating to peptide-MHC binding affinity measurements of at least one of the peptides; and (b) data relating to peptide-MHC binding stability measurements of at least one of the peptides (claims 1-11). Bulik-Sullivan further discloses that the sequence data is TCR sequence (col. 85-86; Example 12, FIG. 26).
Bulik-Sullivan further discloses using 10x software and a custom bioinformatics pipeline, the sequencing output was analyzed to identify T-cell receptor (TCR) α and β pairs, as shown in Supplementary Table 6. Supplementary Table 6 also lists the most common TCR homologous α and β variable (V) regions, linker (J) regions, constant (C) regions, and β-diversity (D) regions, as well as the CDR3 amino acid sequence. A homologous phenotype was defined as an α and β pair with a unique CDR3 amino acid sequence. For single α and single β pairs present at frequencies of two or more cells, homologous phenotypes were filtered to generate a final list of homologous/target peptides in patient CU04 (col. 86; Supplementary Table 6). Bulik-Sullivan further discloses encoding the peptide sequence comprises encoding the peptide sequence using a (claim 12, col. 245) one-hot encoding scheme. Bulik-Sullivan further discloses peptide-MHC binding affinity prediction (col. 83, para. 1).
Regarding claim 14, Fischer teaches a data set based on single-cell pMHC capture in which paired ɑ- and β-chains could be successfully reconstructed for 10,000s of cells and binding-specificity measured for 44 distinct pMHC complexes. Fischer utilizes the dataset as described in Fischer at page 2, para. 2 with reference to the 10x dataset. Fischer further teaches the dextramer is sorted (Figure 1 of 10x Genomics) and that the panel contains 6 dCODE Dextramer® reagents with irrelevant negative control peptides to assist in the detection of non-specific binding events (pg. 3, col. 1, first para. Of 10X Genomics). Fischer further teaches that after the final centrifugation, the cells were resuspended and cell suspension was reserved for a non-sorted cell population, as evidenced by 10X Genomics (pg. 3, col. 1, last para.); reading on limitations of determining, based on the dextramer sequence data, sorted dextramer sequence data wherein the sorted dextramer sequence data comprises sorted test dextramer sequence data and negative control dextramer sequence data and unsorted dextramer sequence data, wherein the unsorted dextramer sequence data comprises unsorted test dextramer sequence data.
Regarding claim 15, Fischer discloses setting a threshold such that a specific binding event required a UMI count greater than 10 that was also greater than five times the highest negative control UMI count for that cell (pg. 5, col. 1). Fischer teaches that the dCODE dextramer reagents with the negative control peptides were combined prior to sorting, as such negative control applied to both sorted and non-sorted signals (pg. 3, col. 1, first para. Of 10X Genomics); reading on limitations of determining a maximum negative control dextramer signal; determining, a maximum sorted dextramer signal and determining, a maximum unsorted dextramer signal.
Regarding claim 19, Fischer teaches that in standard single-cell RNA-seq processing, such effects are often rectified through normalization. Fischer further states that such normalization factors and negative control pMHC counts (for example, pMHC normalization) are useful predictors of a false negative binding event: We compared models only considering the donor identity covariate and models that also included a scaled total mRNA count (for example, cell-wise) covariate and ones that contained negative control count covariates (section: Cell-specific covariates improve binding event prediction, pg. 2, para. 3); reading on limitations of normalizing the dextramer sequence data, wherein normalizing the dextramer sequence data comprises: performing, for each cell represented in the dextramer sequence data, cell-wise and normalization on the dextramer signals associated with each cell; and performing, for each cell represented in the dextramer sequence data, pMHC-wise normalization.
In KSR Int 'l v. Teleflex, the Supreme Court, in rejecting the rigid application of the teaching, suggestion, and motivation test by the Federal Circuit, indicated that “The principles underlying [earlier] cases are instructive when the question is whether a patent claiming the combination of elements of prior art is obvious. When a work is available in one field of endeavor, design incentives and other market forces can prompt variations of it, either in the same field or a different one. If a person of ordinary skill can implement a predictable variation, § 103 likely bars its patentability.” KSR Int'l v. Teleflex lnc., 127 S. Ct. 1727, 1740 (2007).
Applying the KSR standard to Fischer and Bulik-Sullivan, the examiner concludes that the combination of Fischer and Bulik-Sullivan represents applying known technique to a known method. Both Fischer and Bulik-Sullivan are directed to predicting antigen specify of T cells using high throughput single cell repertoire profiling experiments. Fischer only disclosed data preprocessing of generating training dataset of multi-task models, given multiple initial amino acid embeddings, for TCR sequences and inherently by using 10x Genomics dataset V and J segment sequences as a one-dimensional input vector comprising paired alpha-beta chain CDR3 sequence. In the same field of research, Bulik-Sullivan provided incorporating V and J segment sequences utilizing a one-dimensional input vector and binding affinity model predictor. One ordinary skilled in the art would have recognized that aplying the known technique of Bulik-Sullivan would have yielded predictable results and resulted in an improved method. Combining the method of Fischer with input vector of Bulik-Sullivan would have successfully allowed for assembling a meaningful dataset and more accurate predictive results by accessing a broader range of sequence information. One ordinary skilled in the art before he effective filing data of the claimed invention would have had a reasonable expectation of success at combining the method of Fischer and Bulik-Sullivan and the combination would have been successful. This combination would have been expected to have provided a more meaningful assembly of dataset for a more accurate binding site prediction. Therefore, the invention would have been prima facie obvious to one of skill in the art before the effective filing date of the claimed invention, absent evidence to the contrary.
Claim 11 is rejected under 35 U.S.C. 103 as being unpatentable over Fischer et al. (Predicting antigen specificity of single T cells based on TCR CDR3 regions, bioRxiv, August 14, 2019, pages 1-26), as evidence by 10X Genomics (A new way of exploring immunity: linking highly multiplexed antigen recognition to immune repertoire and phenotype, 10xgenomics, 2019, pages 1-16) in view of Bulik-Sullivan et al. (US 11,264,117 B2), as applied to claim 10, 14-15, 19 above, in view of Dash et al. (Quantifiable predictive features define epitope specific T cell receptor repertoires, July 2017, 7 Macmillan Publishers Limited, part of Springer Nature pages 89-109), and further in view of Wolock (Scrublet: Computational Identification of Cell Doublets in Single Cell Transcriptomic Data, Cell Syst. 2019 April 24; 8(4): 281–291).
Regarding claim 11, Fischer teaches removing putative doublets from the data set (Section: Cell-specific covariates improve binding event prediction, pg. 2, para. 3); reading on limitations of filtering, from the dextramer sequence data, based on the single cell sequence data, data associated with low-quality cells
Fischer further teaches removing putative doublets from the data set (Section: Cell-specific covariates improve binding event prediction pg. 2, para. 3).
Further regarding claim 11, Fischer and Bulik-Sullivan do not expressly disclose determining a number of genes outside a gene threshold range. Dash is directed to the determinants of epitope specificity of T cell receptor repertoires (Abstract). Dash teaches quantifying V and J segment usage within a chain and across chains characterizing by an overrepresentation of individual genes as well as significant gene pairing preferences (pg. 89, col. 2, para. 2). Dash further teaches setting upper and lower bound thresholds and quantifying covariations between gene usage (pg. 94, col. 2, last para. - page 95, col. 1, first para.); reading on limitations of determining, for each cell represented in the dextramer sequence data, based on the single cell sequence data, a number of genes; removing, from the dextramer sequence data, data associated with cells having a number of genes outside of a gene threshold range.
Further regarding claim 11, Fischer, Bulik-Sullivan, and Dash do not disclose determining a fraction of mitochondrial gene expression that exceeds a gene expression threshold. However, Wolock is directed to computational identification of cell doublets in single cell transcriptomic data and introduces Single-Cell Remover of Doublets (Scrublet), a framework for predicting the impact of multiplets in a given analysis and identifying problematic multiplets (Summary). Wolock further teaches excluding mitochondrial gene counts of higher than a threshold in the doublet detection step (pg. 23, para. 3).
Applying the KSR standard to Fischer, Bulik-Sullivan, and Dash, the examiner concludes that the combination of Fischer, Bulik-Sullivan, and Dash represents the use of known techniques to improve similar methods. Fischer, Bulik-Sullivan, and Dash are directed to predicting antigen specify of T cells using high throughput single cell repertoire profiling experiments. Fischer and Bulik-Sullivan only disclosed data preprocessing and dataset assembly and training a model. In the same field of research, Dash provided quantifying V and J segment usage within a chain and across chains characterizing by an overrepresentation of individual genes according to defined thresholds, for the purpose of assessing the diversity of the immune receptor repertoire, indicating which gene segments are more common in antigen recognition providing a higher resolution of epitope specificity. Combining the method of Fischer and Bulik-Sullivan with gene quantification of Dash would have allowed for identifying potential antigen-specific TCRs. One ordinary skilled in the art before he effective filing data of the claimed invention would have had a reasonable expectation of success at combining the method of Fischer and Bulik-Sullivan with methods of Dash. This combination would have been expected to have provided a more meaningful assembly of dataset for a more accurate binding site prediction. Therefore, the invention would have been prima facie obvious to one of skill in the art before the effective filing date of the claimed invention, absent evidence to the contrary
Applying the KSR standard to Fischer, Bulik-Sullivan, Dash and Wolock, the examiner concludes that the combination of Fischer, Bulik-Sullivan, Dash and Wolock, represents the use of known techniques to improve similar methods. Fischer, Bulik-Sullivan, Dash and Wolock use high throughput single cell repertoire profiling experiments. Fischer, Bulik-Sullivan, and Dash only disclosed data preprocessing and dataset assembly and training a model and related gene quantifications. Wolock provided a framework for predicting the impact of multiplets in a given analysis and identifying problematic multiplets by excluding mitochondrial gene counts of higher than a threshold in the doublet detection step. Combining the method of Fischer, Bulik-Sullivan, and Dash with mitochondrial gene expression analysis of Wolock would have allowed for a more accurate identification of doublets. One ordinary skilled in the art before he effective filing data of the claimed invention would have had a reasonable expectation of success at combining the method of Fischer, Dash, and Wolock. This combination would have been expected to have provided a more meaningful assembly of dataset by removing problematic multiplets. Therefore, the invention would have been prima facie obvious to one of skill in the art before the effective filing date of the claimed invention, absent evidence to the contrary.
Claim 16 is rejected under 35 U.S.C. 103 as being unpatentable over Fischer et al. (Predicting antigen specificity of single T cells based on TCR CDR3 regions, bioRxiv, August 14, 2019, pages 1-26), as evidence by 10X Genomics (A new way of exploring immunity: linking highly multiplexed antigen recognition to immune repertoire and phenotype, 10xgenomics, 2019, pages 1-16) in view of Bulik-Sullivan et al. (US 11,264,117 B2), as applied to claim 10, 14-15, 19 above, in view of Tungatt et al. (Antibody Stabilization of Peptide–MHC Multimers Reveals Functional T Cells Bearing Extremely Low-Affinity TCRs, J Immunol (2015) 194 (1): 463–474)
Regarding claim 16, Fischer and Bulik-Sullivan teach the limitations of claims 11, and 14-15 above. Further, Fischer discloses that an alternative interpretation of the improved performance of multi-task models is their ability to learn better de-noised low-dimensional representations of TCR sequence through the integration of more diverse training data (Continuous binding affinities can be predicted based on pMHC counts, pg. 3, last para.). Fischer further discloses using fluorescent antibodies to stain reactions and gating cells, as evidenced by 10X Genomics (pg. 3, col. 1, para. 3). Fischer further discloses using the binarization described in 10X Genomics for the raw counts to receive binary outcome labels: A total pMHC UMI count larger than 10 and at least five times as high as the highest observed UMI count across all negative control pMHCs was required for a binding event. If more than one pMHC passed these criteria, the pMHC with the largest UMI count was chosen as the single binder (Binarization of 10x CD8+ T-cell pMHC counts into bound and unbound states); reading on limitations of estimating, based on the maximum negative control dextramer signals, a dextramer binding background noise; estimating, based on the maximum sorted dextramer signals and the maximum unsorted dextramer signals, a dextramer sorting gate efficiency; determining, based on the dextramer binding background noise and the dextramer sorting gate efficiency measure of background noise.
Fischer and Bulik-Sullivan do not expressly disclose subtracting, for each cell represented in the dextramer sequence data, the measure of background noise from a dextramer signal associated with each cell. However, Tungatt discloses an improved staining technique using anti-fluorochrome unconjugated primary Abs followed by secondary staining with anti-Ab fluorochrome-conjugated Abs to amplify fluorescence intensity. Tungatt further discloses a technique resulting an improved fluorescence intensity with both pMHC tetramers and dextramers (abstract). Tungatt further discloses subtracting background noise from signals tetramer/dextramer signals (Figure 2 and 4).
Applying the KSR standard to Fischer, Bulik-Sullivan, and Tungatt, the examiner concludes that the combination of Fischer, Bulik-Sullivan, and Tungatt represents the use of known techniques to improve similar methods. Fischer, Bulik-Sullivan, and Tungatt are directed to characterization of antigen specify of T cells using dextramer sequence data. Fischer and Bulik-Sullivan only disclosed estimating, a dextramer binding background noise; estimating a dextramer sorting gate efficiency, and data preprocessing, dataset assembly, and training a model. In the same field of research, Tungatt provided an improved staining technique in which they subtracted background noise from tetramer/dextramer signals. Combining the background noise estimation of Fischer and Bulik-Sullivan with background noise subtraction of Tungatt would have allowed removal of non-specific binding and unwanted signals, resulting in improved signal to noise ratio. One ordinary skilled in the art before he effective filing data of the claimed invention would have had a reasonable expectation of success at combining the method of Fischer, Bulik-Sullivan, and Tungatt. This combination would have been expected to have provided an improved noise to signal ratio ensuring that data reflects the actual antigen-specific binding event. Therefore, the invention would have been prima facie obvious to one of skill in the art before the effective filing date of the claimed invention, absent evidence to the contrary.
Claim 20 is rejected under 35 U.S.C. 103 as being unpatentable over Fischer et al. (Predicting antigen specificity of single T cells based on TCR CDR3 regions, bioRxiv, August 14, 2019, pages 1-26), as evidence by 10X Genomics (A new way of exploring immunity: linking highly multiplexed antigen recognition to immune repertoire and phenotype, 10xgenomics, 2019, pages 1-16) in view of Bulik-Sullivan et al. (US 11,264,117 B2), as applied to claim 10, 14-15, 19 above, in view of Redmond et al. (Single-cell TCRseq: paired recovery of entire T-cell alpha and beta chain transcripts in T-cell receptors from single-cell RNAseq, Genome Medicine (2016) 8:80).
Regarding claim 20, Fischer teaches designing a model to predict TCR-antigen binding based on ɑ- and β-chain sequences and cell-specific covariates accounting for the variability found in data including chains (Fig. 1b) (Results) indicating that only sequences containing both alpha and beta chains were used. Fischer further discloses removing all cellular barcodes that contain more than one ɑ- or β-chain as mature CD8+ T cells are expected to only have a single functional ɑ- and β-chain allele (pg. 22, L. 455-460). Fischer does not expressly teach removing, from the normalized dextramer sequence data, based on the presence or the absence of the at least one a-chain and the at least one b-chain, data associated with cells having only an a-chain, only a b-chain, or multiple a- or b-chains.
However, Redmond teaches a method of characterizing of the repertoire of the T-cell receptor (TCR) alpha and beta chains. Redmond further teaches that after normalization, potential alpha and beta chains were concatenated and gap-filled assembly was performed and that only samples with expression of both TRAV and TRBV was used in the analysis (pg. 8) implying that cells with only an alpha chain or only a beta chain might not be considered as part of the core output of the scTCRseq analysis.
Applying the KSR standard to Fischer, Bulik-Sullivan, and Redmond, the examiner concludes that the combination of Fischer, Bulik-Sullivan, and Redmond represents the use of known techniques to improve similar methods. Fischer, Bulik-Sullivan, and Redmond are directed to predicting binding affinity of antigen to T cells by using Single-cell RNAseq reads. Fischer and Bulik-Sullivan only disclosed designing a model to predict TCR-antigen binding based on a- and B-chain sequences and removing barcodes containing multiple alpha and beta chains. In the same field of research, Redmond provided method of excluding data that includes either of the chains after the normalization step. Combining the predictive model of Fischer and Bulik-Sullivan with technique of chain exclusion after normalization for a higher accuracy in the prediction. One ordinary skilled in the art before he effective filing data of the claimed invention would have had a reasonable expectation of success at combining the method of Fischer, Bulik-Sullivan, and Redmond. This combination would have been expected to have provided improved targeted therapeutics. Therefore, the invention would have been prima facie obvious to one of skill in the art before the effective filing date of the claimed invention, absent evidence to the contrary.
Claim 21 is rejected under 35 U.S.C. 103 as being unpatentable over Fischer et al. (Predicting antigen specificity of single T cells based on TCR CDR3 regions, bioRxiv, August 14, 2019, pages 1-26), as evidence by 10X Genomics (A new way of exploring immunity: linking highly multiplexed antigen recognition to immune repertoire and phenotype, 10xgenomics, 2019, pages 1-16) in view of Bulik-Sullivan et al. (US 11,264,117 B2), as applied to claim 10, 14-15, 19 above, in view of Yelenski et al. (US 20220265812 A1).
Regarding claim 21, Fischer discloses that in their predictive algorithm, a python package was built (TcellMatch) that hosts a pre-trained model zoo for analysts to impute pMHC-derived antigen specificities and allows transfer and re-training of models on new data sets (introduction). Further Fischer teaches predicting TCR-antigen binding based on ɑ- and β-chain sequences and cell-specific covariates and modeling binding events within a panel of antigens as a single- or multi-task prediction model through a vector of output nodes corresponding to antigens. (results); reading on limitations of training a predictive model based on the data remaining in the normalized filtered dextramer sequence data; predicting a binding status of a newly presented receptor sequence according to the trained predictive model; presenting, to the predictive model, subject TCR sequence data.
Further regarding claim 21, Fischer and Bulik-Sullivan do not expressly disclose determining, based on a repository of antigen locations and the subject TCR binding pattern, a likelihood that a subject associated with the TCR sequence data has traveled to one or more locations. However, Yelenski teaches determining whether the subject expresses one or more HLA alleles involves a population-based analysis. More specifically, determining whether the subject expresses one or more HLA alleles includes determining the origin of the subject and further identifying one or more HLA alleles that are known to be commonly expressed by the population of individuals of that origin. Examples of an origin can be geographic location [0507].
Applying the KSR standard to Fischer, Bulik-Sullivan and Yelenski , the examiner concludes that the combination of Fischer, Bulik-Sullivan, and Yelenski represents the use of known techniques to improve similar methods. Fischer, Bulik-Sullivan and Roman are directed to predicting binding affinity of antigen to T cells using dextramer sequence data. Fischer and Bulik-Sullivan only disclosed training a predictive model based on the data remaining in the normalized filtered dextramer sequence data; predicting a binding status of a newly presented receptor sequence. In the same field of research, Yelenski provided a population-based analysis where the geographic location of the subject can be determined. Combining the predictive model of Fischer and Bulik-Sullivan with population-based analysis of Yelenski would have provided insight into targeted therapy for the subject. One ordinary skilled in the art before he effective filing data of the claimed invention would have had a reasonable expectation of success at combining the method of Fischer, Bulik-Sullivan and Yelenski. This combination would have been expected to have provided improved targeted therapeutics. Therefore, the invention would have been prima facie obvious to one of skill in the art before the effective filing date of the claimed invention, absent evidence to the contrary.
Claim 25 is rejected under 35 U.S.C. 103 as being unpatentable over Fischer et al. (Predicting antigen specificity of single T cells based on TCR CDR3 regions, bioRxiv, August 14, 2019, pages 1-26), as evidence by 10X Genomics (A new way of exploring immunity: linking highly multiplexed antigen recognition to immune repertoire and phenotype, 10xgenomics, 2019, pages 1-16) in view of Bulik-Sullivan et al. (US 11,264,117 B2), as applied to claim 10, 14-15, 19 above, in view of Peng (US11,513,113B2).
Regarding claim 25, Fischer teaches steps as disclosed above with respect to claim 10. Fischer further teaches predicting TCR-antigen binding based on ɑ- and β-chain sequences and cell-specific covariates and modeling binding events within a panel of antigens as a single- or multi-task prediction model through a vector of output nodes corresponding to antigens. (Results, pg. 2, para. 2); reading on limitations of generating, based on the data remaining in the normalized dextramer sequence data associated with reliable TCR-pMHC binding events, a TCR binding pattern for a subject.
Further regarding claim 25, Fischer and Bulik-Sullivan do not expressly disclose receiving, at a subsequent point in time, second single cell sequence data, second dextramer sequence data, and second single cell T Cell Receptor (TCR) sequence data for the subject; determining, based on the second single cell sequence data, second dextramer sequence data, and second single cell T Cell Receptor (TCR) sequence data for the subject, a second TCR binding pattern; and identifying, based on a comparison of the TCR binding pattern for the subject and the second TCR binding pattern, the subject.
However, Peng teaches a method of monitoring an immune repertoire in a subject, comprising: providing two or more distinct particle sets, each distinct particle set comprising a unique antigen peptide and at least one defined barcode operably associated with the identity of the antigen peptide, and each set comprises a first particle comprising a first identifying label and a second particle comprising a second identifying label distinct from the first identifying label; providing a sample known or suspected to comprise one or more T cells, wherein the sample is obtained from a subject over time; contacting the sample with the two or more particle sets, wherein the contacting comprises providing conditions sufficient for a single T cell to bind to the unique antigen of at least one particle set; isolating one or more T cells associated with the first and second identifying label; performing an assay to identify one or more barcodes bound to the isolated T cell; determining a ratio of the barcodes bound to the isolated T cell wherein the ratio is calculated by identifying a first copy number of a predominant barcode and a second copy number of a distinct barcode from step (e) and dividing the first copy number by the second copy number; identifying the antigen specificity of the T cell based on the ratio; and monitoring changes in the antigen specific T cells identified by the method in the subject (col. 6, last two para.); reading on limitations of receiving, at a subsequent point in time, second single cell sequence data, second dextramer sequence data, and second single cell T Cell Receptor (TCR) sequence data for the subject; determining, based on the second single cell sequence data, second dextramer sequence data, and second single cell T Cell Receptor (TCR) sequence data for the subject, a second TCR binding pattern; and identifying, based on a comparison of the TCR binding pattern for the subject and the second TCR binding pattern, the subject.
Applying the KSR standard to Fischer, Bulik-Sullivan, and Peng, the examiner concludes that the combination of Fischer, Sid Bulik-Sullivan, and Peng represents the use of known techniques to improve similar methods. Fischer, Bulik-Sullivan, and Peng are directed to characterization of antigen specify of T cells using dextramer sequence data. Fischer and Bulik-Sullivan disclosed generating, based on the data remaining in the normalized dextramer sequence data associated with reliable TCR-pMHC binding events, a TCR binding pattern for a subject. In the same field of research, Peng provided a second dextramer sequence data, and second single cell T Cell Receptor (TCR) sequence data for the subject by monitoring of the immune repertoire of the subject that allows gaining insight about the health status of the subject. Combining the predictive model of Fischer and Bulik-Sullivan with monitoring system of Peng would have provided insight into the health of the subject, which in turn is crucial for disease diagnosis and treatment response prediction. One ordinary skilled in the art before he effective filing data of the claimed invention would have had a reasonable expectation of success at combining the method of Fischer, Bulik-Sullivan, and Peng. This combination would have been expected to have provided an improved prognosis. Therefore, the invention would have been prima facie obvious to one of skill in the art before the effective filing date of the claimed invention, absent evidence to the contrary.
Response to Applicant’s Arguments
Applicant's arguments filed 09/30/2025 have been fully considered but they are not persuasive.
Applicant states:
a. Sidhom does not describe one-dimensional input vectors with paired ap
CDR3 Sequences, V Gene, and J Gene
b. Sidhom does not generate training data with one-dimensional input vectors
c. Fischer in view of Sidhom does not present a TCR sequence to a trained
d. The combination of Fischer and Sidhom is not supported because Fischer is
directed to antigen classification within fixed antigen panels, while Sidhom is
directed to structural repertoire modeling and biomarker development
It is respectfully submitted that this is not persuasive. A new round of art rejections are applied to the instant claims. As stated above, the combination of Fischer and Bulik-Sullivan disclose all the limitations of claim 10.
Conclusion
No claims are allowed.
Claim 17 appears to be free from the prior art because the closest prior art to Fischer, Bulik-Sullivan, Dash, Tungatt, Redmond, Yelenski, and Peng does not appear to teach or fairly suggest the steps directed to estimating, based on the maximum sorted dextramer signals and the maximum unsorted dextramer signals, the dextramer sorting gate efficiency comprises determining a maximum difference between the maximum sorted dextramer signals and the maximum unsorted dextramer signals.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to GHAZAL SABOUR whose telephone number is (703)756-1289. The examiner can normally be reached M-F 7:30-5:00.
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, Larry D. Riggs can be reached at (571) 270-3062. 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.
/G.S./Examiner, Art Unit 1686
/LARRY D RIGGS II/Supervisory Patent Examiner, Art Unit 1686