DETAILED ACTION
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Herein, “the previous Office action” refers to the Final Rejection filed 9/17/2025.
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and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114.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 12/9/2025 has been entered.
Priority
As detailed on the Filing Receipt filed 4/3/2025, the instant application claims priority to as early as 4/12/2018. At this point in prosecution, all claims are accorded the earliest claimed priority date.
Information Disclosure Statement
The Information Disclosure Statement filed on 12/18/2025 is in compliance with the provisions of 37 CFR 1.97 and has been considered in full. A signed copy of the IDS is included with this Office Action.
Claim Status
Claim 18 is canceled.
Claims 1-17 and 19-21 are pending, and under examination.
Withdrawn Rejections
The rejections of claims 1-17 and 19-21 under 35 USC § 103, as being unpatentable over combinations of Xiong, Roth, Robison, Jia and/or Patterson, are hereby withdrawn in view of Applicant’s amendment of the claims and persuasive argument that the cited art does not teach analysis of sequencing data and metadata features for a tumor-only sample as required by the amended claims (Remarks filed 12/9/2025, pg. 18, para. 3 – pg. 23, para. 2).
Response to Arguments - Claim Rejections Under 35 USC § 101
In the Remarks filed 12/9/2025 (hereafter “the Remarks”), Applicant traverses the rejection under 35 USC § 101 and presents supporting arguments. Herein, “CardioNet” refers to the court decision of CardioNet, LLC v. Infobionic, Inc., 955 F.3d 1358 (Fed. Cir. 2020).
Applicant highlights amendment of the claims to expressly recite specialized hardware (i.e., a sequencer instrument), similarly to the ‘device’ claims held eligible in CardioNet, while noting recognition by the CardioNet court of certain considered claims (e.g., claim 22) as demonstrating patent-eligible improvement to cardiac monitoring technology despite their not reciting specialized hardware (pg. 12, para. 2 – pg. 13, para. 2).
The claims held eligible in CardioNet were considered by the court to accomplish improved distinction of atrial fibrillation and flutter via employment of techniques that had not previously been employed in the relevant field (“Nothing in the record in this case suggests that the claims merely computerize pre-existing techniques for diagnosing atrial fibrillation and atrial flutter”, 955 F.3d at 1370). Accordingly, the implementation of these techniques by the claimed ‘device’ and ‘article’ was considered by the court as demonstrating improvement in cardiac monitoring technology.
Unlike those considered in CardioNet, the instant claims stand rejected over prior art disclosing employment of recited technology (e.g., the sequencer instrument, processor(s), and non-transitory computer readable media) to perform recited techniques (e.g., sequencing a tumor sample, processing sequence data via a neural network, and classifying somatic variants). See ‘Claim Rejections – 35 USC § 103’ section.
Regarding the newly-incorporated sequencer instrument in particular, use of a machine that contributes only nominally or insignificantly to the execution of a claimed method (e.g., in a data gathering step) does not integrate a recited judicial exception or provide significantly more (MPEP 2106.05(b) § III). Neither does the addition of the sequencer instrument provide an inventive concept by virtue of unconventionality, as employment of sequencer instruments to generate sequencing data is well-understood, routine and conventional in the field of the invention. Thus, the inclusion of the sequencer instrument is insufficient to render the claims as patent eligible.
Applicant alleges that, similarly to how claims held eligible in CardioNet demonstrated improved accuracy in distinguishing between atrial fibrillation, flutter, and other types of arrythmias, the claimed technology exhibits improved accuracy in classifying somatic and germline variants as compared to traditional classifiers (pg. 14, para. 2 – pg. 15, para. 1).
Applicant highlights supporting evidence in the specification, including Figs. 9 and 13 which indicate improved performance of an exemplary embodiment of the invention (Sojourner) in classifying variants as compared to several versions of a non-deep neural network (pg. 15, para. 2 – pg. 16, para. 1).
Predictive accuracy and computation times are model performance metrics. Achievement of improved model performance metrics by the claimed invention, as compared to those achieved by prior classification models, is based on comparative improvements to model parameters embodied by the recited neural network. Mathematical algorithms, including those executed on generic computer hardware, are abstract ideas (DDR Holdings, LLC v. Hotels.com LP, 773 F.3d 1245, 1256 (Fed. Cir. 2014)) and an inventive concept cannot be furnished by an abstract idea itself (Genetic Techs. V. Merial LLC, 818 F.3d 1369, 1376 (Fed. Cir. 2016)). Comparative improvement of a mathematical algorithm is not a sufficient point of eligibility under § 101.
Neither is the implementation of an improved algorithm on generic computer hardware equivalent to improvement in the functioning of the computer hardware itself. The court of Bancorp Serves., LLC v. Sun Life Assur. Co. of Canada (U.S.), 687 F.3d 1266 (Fed. Cir. 2012) found that improved performance of calculations via a computer did not materially alter patent eligibility of claimed subject matter (687 F.3d at 1278). See also FairWarning IP, LLC v. Iatric Sys., 839 F.3d 1089, 1095 (Fed. Cir. 2016). Thus, evidence of superior model performance is found unpersuasive.
Applicant alleges that, similarly to how claims held eligible in CardioNet demonstrated improved detection of heart-activity anomalies, the claimed invention provides improved detection of somatic variants related to cancer (pg. 16, para. 2 – pg. 17, para. 1).
The production of a particular analytical result (e.g., a list of somatic variants) is not, by itself, an indicator of patent eligibility. The court of TLI Communications LLC v. AV Auto. LLC, 823 F.3d 607 (Fed. Cir. 2016) found that considered functions of gathering and analyzing information, and displaying the result, did not improve the functionality of the implementing computer (823 F.3d at 612-13). The production of results bearing particular clinical significance results from application to a particular clinical field of use (detection of somatic variants related to cancer). In Recentive Analytics, Inc. v. Fox Corp., 134 F.4th 1205 (Fed. Cir. 2025), the court held that claimed application of a trained machine learning algorithm to a particular field of use was insufficient to confer patent eligibility.
Applicant alleges that, similarly to how claims held eligible in CardioNet improved the speed of detection of a medical condition, the claimed invention improves efficiency of variant classification and conserves physical and computational resources through operation based on a ‘tumor-only’ biological sample and avoidance of matched ‘normal’ sample sequencing (pg. 17, para. 2).
The Examiner agrees that, the undertaken processes being otherwise equal, sequencing and analysis of matched tumor and normal samples is logically less efficient than sequencing and analysis of only the tumor samples. However, performance of variant classification based on sequencing and analysis of tumor-only samples is known in the art (see, e.g., the Smith reference cited herein).
For the above reasons, the arguments are found unpersuasive. The rejection is maintained, and has been revised to fully address the amended claims (filed 12/9/2025).
Claim Rejections - 35 USC § 101
35 USC § 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 1-17 and 19-21 are rejected under 35 USC § 101 because the claimed invention is directed to an abstract idea without significantly more (i.e., non-statutory subject matter). This rejection is maintained from the previous Office action, and has been revised to address the amended claims (filed 12/9/2025).
"Claims directed to nothing more than abstract ideas, natural phenomena, and laws of
nature are not eligible for patent protection" (MPEP 2106.04 § 1). Abstract ideas include
mathematical concepts (including formulas, equations and calculations), and procedures for
evaluating, analyzing or organizing information, which are a type of mental process (MPEP
2106.04(a)(2)). The claims as a whole, considering all claim elements both individually and in
combination, do not amount to significantly more than an abstract idea.
Step 1: The Four Categories of Statutory Subject Matter (MPEP 2106.03)
The claims are directed to a system (claims 1-15), a method (claims 16 and 19-20), and a non-transitory computer-readable storage medium (claims 17 and 21), which fall under categories of statutory subject matter.
Step 2A, Prong One: Whether the Claims Set Forth or Describe a Judicial Exception
(MPEP 2106.04 § I.A.1)
‘Mathematical concepts’ are relationships between variables and numbers, numerical formulas or equations, or acts of calculation, which need not be expressed in mathematical symbols (MPEP 2106.04(a)(2) § I). The claims recite elements that encompass mathematical concepts, at least under their broadest reasonable interpretation, including:
“feed[ing] the input sequence and the metadata features to a neural network; detect[ing], by the neural network, one or more sequence patterns… and generat[ing], based on the detected sequence patterns and by processing the input… through the neural network, classification scores” (claims 1 and 16-17), i.e., evaluating an algorithm for certain input to calculate values wherein:
“the neural network is trained end-to-end on training examples from a first dataset of cancer causing mutations, followed by training on training examples from a second dataset of cancer-causing mutations” (claim 15) i.e., the algorithm has been sequentially optimized for two training sets,
classification scores indicate “that the variant, when flanked by the flanking bases in the target genomic region for the tumor-only biological sample, is a somatic variant, a germline variant, or noise resulting from one or more errors in the sequencing process of the sequencer instrument” (claims 1 and 16-17) and “the one or more errors in the sequencing process of the sequencer instrument comprises at least one error causing a base call quality score for a set of base calls supporting detection of the variant to have a value below a quality score threshold” (claim 20).
A neural network is a series of mathematical equations, while the recited steps of “feed[ing]… generat[ing]… [and] process[ing]” encompass acts of evaluating the claimed series of equations to calculate values. Hence, the claims encompass mathematical concepts.
The particular representative significance of the classification scores (e.g., as indicating that the variant is a somatic variant) does not alter the mathematical nature of the acts of calculation that produce the scores, thus the above elements that specify the representative significance of classification scores are merely directed to embodiments of the recited acts. See SAP America, Inc. v. InvestPic, LLC, 898 F.3d 1161 (Fed. Cir. 2018); Recentive v. Fox, 134 F.4th 1205, 1213-14 (Fed. Cir. 2025).
‘Mental processes’ are processes that can be performed in the human mind at least with use of a physical aid, e.g., a slide rule or pen and paper (MPEP 2106.04(a)(2) § III). The claims recite elements encompassing mental processes, at least with use of a physical aid and under the broadest reasonable interpretation:
“correlat[ing] the variant with the one or more metadata features” (claim 3) and “correlat[ing] the variant with [particular] feature[s]” (claims 4-14), e.g., associating corresponding feature attributes with variants, wherein features specify:
“whether the variant is a nonsynonymous variant” (claim 4), “whether the variant is a single-nucleotide polymorphism, an insertion, or a deletion” (claim 5), “quality parameters of read mapping that identified the variant” (claim 6), “allele frequencies of the variant in sequenced populations” (claim 7), “allele frequencies of the variant in ethnic sub-populations stratified from sequenced populations” (claims 8), “conservativeness of the target position across multiple species” (claim 9), “a clinical effect of the variant, drug sensitivity of the variant, and histocompatibility of the variant as determined from clinical tests” (claim 10), “the impact of the variant on functionality of a protein resulting from an amino acid substitution caused by the variant” (claim 11), “likelihoods identifying ethnic makeup of an individual that provided a tumor sample associated with the variant” (claim 12), “frequency of the variant in sequenced cancerous tumors” (claim 13) and “at least one base mutated by the variant at the target position” (claim 14); and
“validat[ing], in a sample report, that the variant detected within the tumor-only biological sample comprises a germline variant or a somatic variant using the classification scores” (claims 19 and 21), i.e., generating a sample report including the classification scores, wherein the scores indicate that the variant comprises a germline variant or a somatic variant.
The above processes of organizing associated information are performable in the human mind. For example, the human mind is capable of appending given sets of frequency values to each of a given list of genetic variant positions. Hence, the claims encompass mental processes.
The particular representative significance of the features (e.g., as specifying allele frequencies or clinical effects) does not preclude mental performance of their correlation, and the above elements that specify the representative significance of features are merely directed to embodiments of the recited acts. See, e.g., Digitech Image Techs., LLC v. Elecs. for Imaging, Inc., 758 F.3d 1344, 1350 (Fed. Cir. 2014); Electric Power Group v. Alstom, S.A., 830 F.3d 1350, 1354 (Fed. Cir. 2016).
Mathematical concepts and mental processes constitute enumerated groupings of abstract ideas (MPEP 2106.04(a)(2) §§ I and III). Hence, the claims recite numerous elements that, individually and in combination, constitute an abstract idea. The claims must be examined further to determine whether they integrate the abstract idea into a practical application (MPEP 2106.04(d)).
Step 2A, Prong Two: Whether the Claims Contain Additional Elements that
Integrate the Judicial Exception(s) into a Practical Application (MPEP 2106.04 § 11.A.2)
The claims recite the following additional elements, which are directed to computer hardware that performs claimed functions encompassing the abstract idea:
“A system… comprising: at least one processor; and a non-transitory computer-readable medium comprising instructions that, when executed by the at least one processor, cause the system to” perform claimed functions (claim 1), wherein:
“The system… further compris[es] instructions that, when executed by the at least one processor, cause the system to” perform additional claimed functions (claims 3-14); and
“A non-transitory computer readable storage medium storing instructions that, when executed by at least one processor, cause a computer system to” perform claimed functions (claim 17), wherein:
“The non-transitory computer readable storage medium… further stor[es] instructions that, when executed by at least one processor, cause the computer system to” perform additional claimed functions (claim 21).
The claims do not describe any specific computational steps by which computer hardware performs or carries out the abstract idea, nor do they provide any details of how specific structures of computer hardware are used to implement these functions. The claims state nothing more than that generic computer hardware (e.g., at least one processor) performs the functions that constitute the abstract idea, and are therefore mere instructions to apply the abstract idea using computer hardware. As such, the claims do not integrate that abstract idea into a practical application (MPEP 2106.04(d) § I; MPEP 2106.05(f)).
The claims further recite additional elements that gather data necessary for performance of claimed system functions and method steps, including:
“a sequencer instrument” (claims 1 and 16-17);
“execut[ing] a sequencing process on the sequencer instrument to (a) detect light emitted by nucleic acids amplified on a substrate from a tumor-only biological sample and (b) generate sequencing data for reads of the tumor-only biological sample based on the detected nucleic acids” (claim 1);
“access[ing], for the tumor-only biological sample, an input sequence determined from [sequencing data generated in a] sequencing process [by a sequencer instrument from nucleic acids of the tumor-only biological sample] and comprising a variant at a target position flanked by flanking bases [on each side] in a target genomic region” (claims 1 and 16-17); and
“access[ing] metadata features [correlated with the variant]” (claims 1 and 16-17), wherein:
the metadata feature(s) “represent[] mutation characteristics of the variant, read mapping statistics of the variant within [the] reads of the tumor-only biological sample, and occurrence frequency of the variant” (claims 1 and 16-17), and
“the mutation characteristics of the variant include variant type, amino acid impact, evolutionary conservation, and clinical significance” (claim 2).
Necessary data gathering is considered to be insignificant pre-solution activity, and as such insufficient to integrate an abstract idea into a practical application (MPEP 2106.05(g)). The inclusion of a particular machine (the sequencer instrument) is likewise insufficient, as inclusion of machines which merely perform necessary data gathering is insufficient to integrate an abstract idea into a practical application (MPEP 2106.05(b) § III).
None of the claims recite any further additional elements.
When the claims are considered as a whole: they do not improve the functioning of a computer, other technology, or technical field (MPEP 2106.04(d)(1)); they do not apply the abstract idea to effect a particular treatment or prophylaxis for a disease or medical condition (MPEP 2106.04(d)(2)); they do not implement the abstract idea with, or in conjunction with, a particular machine (MPEP 2106.05(b)); they do not effect a transformation or reduction of a particular article to a different state or thing (MPEP 2106.05(c)); and they do not apply or use the abstract idea in some other meaningful way (MPEP 2106.05(e)) beyond linking the use of the abstract idea to a particular field of use and/or technological environment (i.e., computerized analysis of sequence data; MPEP 2106.05(h)).
Therefore, the claims do not integrate the abstract idea into a practical application. See MPEP 2106.04(d) § I.
Because the claims recite an abstract idea, and do not integrate that abstract idea into a practical application, the claims are directed to that abstract idea. Claims that are directed to abstract ideas must be examined further to determine whether the additional elements besides the abstract idea render the claims significantly more than the abstract idea. Claims that are directed to abstract ideas and that raise a concern of preemption of those abstract ideas must be examined to determine what elements, if any, they recite besides the abstract idea, and whether these additional elements constitute inventive concepts that are sufficient to render the claims significantly more than the abstract idea (MPEP 2106.05).
Step 2B: Whether the Claims Contain Additional Elements that Amount to an
Inventive Concept (MPEP 2106.05)
Mere addition of instructions to implement an abstract idea using generic computer hardware does not amount to an inventive concept that would render the claims significantly more than the abstract idea (MPEP 2106.05(d) and 2106.05(f)).
As noted above, several recited additional elements amount to insignificant extra-solution activity. The conventionality of recited additional elements that amount to insignificant extra-solution activity must be further considered.
The specification indicates that the following additional elements, which amount to data gathering (i.e., insignificant extra-solution) activity, may be performed using commercially-available products:
executing a sequencing process on a sequencer instrument to (a) detect light emitted by nucleic acids amplified on a substrate from a tumor-only biological sample and (b) generate corresponding sequencing data (paras. 0083-84: “In SBS a plurality of fluorescently-labeled nucleotides are used to sequence a plurality of clusters of amplified DNA… present on the surface of an optical substrate… different fluorophores may emit different wavelengths of emission light”; para. 0108: “the variant call application may be… implemented on the MiSeq® sequencer instrument”; see discussion regarding cited reference Matsuguchi in ‘Claim Rejections – 35 USC 103’ section).
The specification also indicates that the following additional elements, which amount to data gathering (i.e., insignificant extra-solution) activity, may be performed using computer hardware:
accessing an input sequence (pg. 14, para. 0059: “a collection of sequence data that describes a fragment of a nucleotide sample or reference… may be stored in a memory device and… may be obtained directly from a sequencing apparatus or indirectly from stored sequence information concerning the sample”); and
accessing metadata features (pg. 33, para. 0106: “variant annotation… included a large number of annotations from External data sources, like ExAC, EVS, 1000 Genomes project, dbSNP, ClinVar, Cosmic, DGV and ClinGen… We mainly extracted the allele frequencies information from 1000 Genome Projects”).
The courts have held that computer-implemented functions of retrieving data, both via a computer network and from system memory, are coextensive with generic computer hardware and/or well-understood, routine, and conventional. See In re Katz, 639 F.3d 1303, 1316 (Fed. Cir. 2011); buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355 (Fed. Cir. 2014); OIP Techs., Inc., v. Amazon.com, Inc., 788 F.3d 1359, 1363 (Fed. Cir. 2015); Versata Dev. Group, Inc. v. SAP America, Inc., 793 F.3d 1306, 1334 (Fed. Cir. 2015).
In this way the specification indicates that recited additional elements amounting to insignificant extra-solution activity encompass well-understood, routine, and conventional activity (see MPEP 2106.07(a) § III). Well-understood, routine and conventional activity is insufficient to constitute an inventive concept that would render the claims significantly more than judicial exceptions (MPEP 2106.05(g)).
When the claims are considered as a whole: they do not improve the functioning of a computer, other technology, or technical field (MPEP 2106.05(a)); they do not apply the abstract idea with, or by use of, a particular machine (MPEP 2106.05(b)); they do not effect a transformation or reduction of a particular article to a different state or thing (MPEP 2106.05(c)); they do not add specific limitations, other than what is well-understood, routine, conventional activity in the field, or unconventional steps that confine the claim to a particular useful application (MPEP 2106.05(d)); and they do not provide other meaningful limitations (MPEP 2106.05(e)) beyond generally linking the use of the abstract idea to a particular field of use and/or technological environment (i.e., computerized analysis of sequence data; MPEP 2106.05(h)).
Conclusion: Claims are Directed to Non-statutory Subject Matter
For these reasons, the claims, when the limitations are considered individually
and as a whole, are directed to an abstract idea and lack an inventive concept. Hence,
the claimed invention does not constitute significantly more than the abstract idea, so
the claims are rejected under 35 USC § 101 as being directed to non-statutory subject
matter.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 USC §§ 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 USC § 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 USC § 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 USC § 102(b)(2)(C) for any potential 35 USC § 102(a)(2) prior art against the later invention.
Claims 1, 3, 6-7, 9, 13-14, 16-17 and 20 are rejected under 35 USC § 103 as being unpatentable over Smith et al (Bioinformatics 32(6): 808-813; published 11/20/2015), in view of Xiong et al (WO 2018/006152; filed 7/4/2016; on IDS filed 3/27/2025; previously cited) and Matsuguchi et al (US 2018/0089373; effectively filed 9/23/2016; on IDS filed 8/15/2025). The new grounds of rejection presented herein were necessitated by Applicant’s amendment of the claims (filed 12/9/2025) to introduce new limitations.
Claim 1 recites a system for classifying variants, comprising: a sequencer instrument; at least one processor; and a non-transitory computer-readable medium comprising instructions that, when executed by the processor(s), cause the system to perform recited functions.
The claim recites the following particular functional capabilities: executing a sequencing process on the sequencer instrument to (a) detect light emitted by nucleic acids amplifies on a substrate from a tumor-only biological sample and (b) generate corresponding sequencing data for reads of the sample; accessing, for the tumor-only biological sample, an input sequence determined from the sequencing data and comprising a variant at a target position flanked by flanking bases in a target genomic region; accessing metadata features representing mutation characteristics, read mapping statistics of the variant within the reads, and occurrence frequency of the variant; feeding the input sequence and metadata features to a neural network; detecting, by the neural network, one or more sequence patterns among the variant and the flanking bases; and generating, based on the detected sequence pattern(s) and by processing the input sequence and metadata features through the neural network, classification scores that the variant, when flanked by the flanking bases in the target genomic region for the sample, is a somatic variant, germline variant, or noise resulting from one or more errors in a sequencing process of the sequencer instrument.
With respect to claim 1, Smith discusses SomVarIUS, a computational method for detecting somatic variants using sequencing data from unpaired tumor tissue samples, and describes functions of: accepting BAM files as input; pre-processing sequencing data to prioritize potential variant sites including exclusion of reads based on thresholds for mapping quality, coverage and base quality (i.e., metadata features representing read mapping statistics as claimed); estimating probabilities of a sequencing error and of a germline SNP; and classifying somatic variants based on sub-threshold error and germline probabilities (pg. 808, Abstract and r. column – pg. 809, r. column and Fig. 1A).
Smith further describes performance of the functions of estimating germline probability (via a beta-binomial model) and estimating error probability (via a Chernoff-bound Bernoulli error model) based on information including: base quality, coverage, allelic abundance and information about known germline heterozygous SNPs in the neighborhood (i.e., metadata features representing read mapping statistics and occurrence frequency of the variant as claimed) (pg. 809, r. column and Fig. 1 – pg. 810, l. column).
Smith also states that SomVarIUS fits the beta-binomial distribution for each entire chromosome by default, but can optionally fit the distribution for particular sub-chromosomal segments according to an input copy number segmentation file (pg. 810, l. column). Estimating germline probability for a given variant in this manner requires information regarding chromosomal segment location and coordinate position of the variant (i.e., metadata features representing mutation characteristics as claimed).
Smith discusses the working principle of SomVarIUS: that non-reference bases are probably either of germline or somatic origin, or indicate sequencing error. Therefore potential somatic mutations can be identified by calculating probabilities that an observed, non-reference bases are unlikely due to sequencing error or germline SNPs (pg. 809, l. column). In other words, since the prevailing alternative possibilities for variant origin are germline/somatic/error, relatively low probabilities of error and germline origin indicate a relatively high probability of somatic origin. The disclosed estimation of probabilities that an observed variant is an error or germline variant and classification as a somatic variant based thereon, particularly in view of the referenced discussion, is considered equivalent to generating classification scores that the variant is a somatic variant, germline variant, or noise resulting from sequencing error(s).
Smith states that SomVarIUS is written in Python 2.7 (a computer programming language), and is available on the internet (pg. 808, Abstract). From this disclosure, a person of ordinary skill in the art would understand that SomVarIUS is implemented as processor-executable instructions and performance requires use of a hardware system including at least one processor. A person of ordinary skill in the art would also find it obvious to store processor-executable instructions on a non-transitory computer-readable medium. However, Smith does not disclose a sequencer instrument and does not describe a system as claimed.
Although Smith discloses estimation of probabilities based on fitted statistical models, Smith furthermore does not disclose feeding input data to a neural network; detecting sequence patterns by the neural network; or generating classification scores by the neural network.
Xiong discusses methods for generating and training neural networks that take as input biological sequences and additional information (‘relevance scores’) and output molecular phenotypes (Abstract). Xiong describes analytical operation of the disclosed neural networks, comprising: applying a set of convolutional layers to an input sequence and inputting relevance scores to an associated weighting unit, i.e., feeding the input sequence and metadata features to a neural network (paras. 0050 and 0067-69); detecting the occurrence of sequence features (i.e., sequence patterns) within a spatial window around a given position (paras. 0051, 0055 and 0095); and generating outputs representing a molecular phenotype, which can be quantified as categorical variables, numerical values, or probability distributions, i.e., classification scores (paras. 0034 and 0040).
Xiong also discusses system implementations of their methods, including implementation of any disclosed method using instructions that may be stored on computer-readable media, such as magnetic disks or flash memory (i.e., non-transitory computer-readable media), and executed by one or more processors (para. 0032). Xiong exemplifies a wide range of computer storage media that can be employed, and states that any described processing function may be carried out by a single processor, arrayed or distributed plurality of processors, or threads in a multi-threaded machine (paras. 0032 and 0069).
Xiong indicates that neural network implementation is an enabled means of computationally generating of classification scores based on analysis of input sequence data and metadata features (e.g., performing steps of the analytical method of Smith), and can be performed with a wide range of general-purpose computer hardware configurations. Although Xiong does not discuss application of neural networks to directly classify sequence variants, Xiong does teach that changes in molecular phenotypes arise from sequence variants (para. 0035), i.e., the neural network-implemented system of Xiong quantifies molecular phenotypes based on analysis of underlying sequence variant patterns.
Xiong also exemplifies input of relevance scores that are mutation-specific and represent population allele frequencies (paras. 0009, 0056 and 0100). In this way, Smith and Xiong both discuss classification methods based on algorithmic analysis of sequencing data and similar metadata features. Smith furthermore particularly discusses analysis of tumor-only sequencing data (i.e., sequencing data for reads of a tumor-only biological sample). However, neither Smith nor Xiong describes a process of generating sequencing data, from a biological sample, in further detail. Accordingly, neither Smith nor Xiong teaches executing a sequencing process on a sequencer instrument to (a) detect light emitted by nucleic acids amplifies on a substrate from a tumor-only biological sample and (b) generate corresponding sequencing data for reads of the sample.
Matsuguchi discusses a method for qualifying a subject for a subset of therapies, based on computer analysis of biologic data generated from one or more biological samples of the subject (Abstract), and describes active steps of generating the data. Matsuguchi describes sequencing a tumor tissue sample of a subject to provide sequencing reads, and generating a classifier for identifying genomic aberrations in the sample based at least in part on the sequencing reads (paras. 0004 and 0029). Matsuguchi specifically teaches identifying one or more somatic mutations (para. 0044), and also exemplifies classifiers including neural networks (para. 0135). Matsuguchi also teaches an automated system implementation comprising: a sequencer that subjects the one or more biological samples to sequencing to generate the biologic data, and one or more computer processors that receive the biologic data over a network (para. 0018), which performs sequencing without any user involvement (para. 0040). Matsuguchi exemplifies automated (user-free) performance of particular sequencing techniques including sequencing-by-synthesis (para. 0108). The instant specification states: “In SBS”, sequencing-by-synthesis, “a plurality of fluorescently-labeled nucleotides are used to sequence a plurality of clusters of amplified DNA… present on the surface of an optical substrate” (pg. 18, para. 0087). The specification provides evidence for the fact that the sequencing technique of SBS, taught by Matsuguchi, encompasses generation of sequencing data through detection of the light emitted by labeled nucleic acids amplified on an optical substrate.
Smith and Matsuguchi both discuss variant calling methods based on algorithmic analysis of sequencing data derived from tumor samples, while Xiong and Matsuguchi both discuss system-implemented classification methods based on analysis of sequencing data via neural networks. In this way, the automated system of Matsuguchi presents an advantageous, enabled means of providing neural network-compatible sequencing data as is required for implementation of the analytical method disclosed by Smith via neural networks as taught by Xiong.
In this way, the combined teachings of Smith, Xiong and Matsuguchi make obvious utilizing a system, including a sequencer instrument and one or more processors, to execute a sequencing process on the sequencer instrument to (a) detect light emitted by nucleic acids amplifies on a substrate from a tumor-only biological sample and (b) generate corresponding sequencing data for reads of the sample.
With respect to claim 3, Smith states that SomVarIUS is written in Python 2.7 (a computer programming language), and is available on the internet (pg. 808, Abstract). From this disclosure, a person of ordinary skill in the art would understand that SomVarIUS is implemented as processor-executable instructions.
Smith further describes performance of the functions of estimating germline probability and error probability via statistical modelling, based on information including: base quality, coverage, allelic abundance and information about known germline heterozygous SNPs in the neighborhood (pg. 809, r. column and Fig. 1 – pg. 810, l. column). This involves correlating the variant with the recited features.
Additionally, Xiong teaches implementation of any disclosed method using instructions executed by one or more processors (para. 0032) and discusses evaluating input position-dependent and mutation-specific additional information (paras. 0012, 0056, 0091-92, 0094 and 0100), i.e., variant-correlated metadata features.
With respect to claim 6, Smith discloses performing functions of pre-processing sequencing data and estimating variant probabilities based on mapping quality, coverage and base quality, i.e., quality parameters of read mapping read mapping statistics as claimed (pg. 809, r. column and Fig. 1 – pg. 810, l. column).
With respect to claim 7, Smith discloses user provision of known SNP information as a SNP file, e.g., including SNPs from the 1000 Genomes database filtered by population allele frequency (pg. 810, l. column).
Additionally, Xiong exemplifies utilization of relevance scores obtained from population allele frequency sequences (para. 0009), and position-dependent tracks including tracks of common and rare mutations in human populations (para. 0093).
With respect to claim 9, Xiong teaches that relevance scores may be based on evolutionary conservation, and that position-dependent features may include evolutionary conservation scores (paras. 0009 and 0093).
With respect to claim 13, Smith discloses annotation of known oncogenic/disease associated variants, analysis based on allelic abundance and information about known germline heterozygous SNPs in the neighborhood, and analysis of sequencing data from The Cancer Genome Atlas, i.e., an online database of sequenced cancerous tumor information (pg. 810, l-r. columns pg. 810, r. column). Smith also exemplifies analysis of frequently mutated cancer genes (pg. 812, l. column). In this way, Smith is considered to at least suggest correlation of the variant with a feature that specifies its frequency in sequenced cancerous tumors.
With respect to claim 14, Smith discloses analysis of allelic abundance of reference and alternate alleles at known germ line heterozygous SNP positions (pg. 811, l. column).
Additionally, Xiong discloses that variants may be derived by sequencing samples from patients and aligning them to a reference sequence (para. 0037), and analyzed sequences may contain mutations (para. 0100).
Claim 16 recites a method comprising process steps of substantial similarity to the functional limitations of the system of claim 1.
With respect to claim 16, Smith discloses SomVarIUS as a computational method (pg. 808, Abstract) while Xiong teaches methods for generating and training neural networks using biological sequences and relevance scores (Abstract). The teachings of Smith and Xiong are considered to apply to the process limitations of claim 16 in the same ways as outlined above with respect to the functional limitations of claim 1.
Claim 17 recites a non-transitory computer readable storage medium storing instructions that, when executed by at least one processor, cause a computer system to perform functions of substantial similarity to the functional limitations of the system of claim 1.
With respect to claim 17, Smith states that SomVarIUS is written in Python 2.7, a computer programming language (pg. 808, Abstract). From this disclosure, a person of ordinary skill in the art would understand that SomVarIUS is implemented as processor-executable instructions. A person of ordinary skill in the art would also find it obvious to store processor-executable instructions on a non-transitory computer-readable medium.
Xiong also teaches implementation of computational methods using instructions stored on computer-readable media, such as magnetic disks or flash memory (i.e., non-transitory computer-readable media), and executed by one or more processors (para. 0032). The teachings of Smith and Xiong are considered to apply to the functional limitations of claim 17 in the same ways as outlined above with respect to the functional limitations of claim 1.
With respect to claim 20, Smith discloses pre-processing including exclusion of reads based on position-specific base quality being below a threshold, and estimation of error probability based on base quality (pg. 809, l. column and Fig. 1A – pg. 810, l. column).
An invention would have been obvious to one of ordinary skill in the art if simple substitution of one known element for another, to yield predictable results, would have led one of ordinary skill in the art to arrive at the claimed invention. Before the effective filing date of the claimed invention, said practitioner would have modified the analytical method of Smith to substitute the disclosed models (e.g., beta-binomial distribution) for a neural network, because Xiong presents neural networks as an enabled algorithmic means of generating of classification scores based on input sequence data and metadata features, that can be performed with a wide range of general-purpose computer hardware configurations (Abstract; paras. 0032, 0034, 0040, 0050 and 0067-69). Said practitioner would have had a reasonable expectation of success because Smith and Xiong both discuss classification methods based on algorithmic analysis of sequencing data.
An invention would have been obvious to one of ordinary skill in the art if some teaching in the prior art would have led that person to combine prior art reference teachings to arrive at the claimed invention. Before the effective filing date of the claimed invention, said practitioner would have combined a computational system including a sequencer, taught by Matsuguchi, with the analytical method of Smith, because Matsuguchi presents this system as an enabled means of automatically generating and providing tumor sequencing data (i.e., data that is required for performance of the analytical method of Smith). Said practitioner would have had a reasonable expectation of success because Smith and Matsuguchi both discuss classification methods based on algorithmic analysis of sequencing data.
In this way the disclosure of Smith, in view of Xiong and Matsuguchi, makes obvious the limitations of claims 1, 3, 6-7, 9, 13-14, 16-17 and 20. Thus, the claimed invention is prima facie obvious.
Claims 2, 4-5, 11, 19 and 21 are rejected under 35 USC § 103 as being unpatentable over Smith, in view of Xiong and Matsuguchi, as applied to claims 1, 3 and 16-17 above, and further in view of Robison et al (US 2015/0066378; published 3/5/2015; previously cited). The new grounds of rejection presented herein were necessitated by Applicant’s amendment of the claims (filed 12/9/2025) to introduce new limitations.
With respect to claim 2, Smith discloses annotation of detected variants according to a user-provided catalog of known oncogenic/disease-associated variants, e.g., mutations conferring drug-resistance (pg. 810, r. column; pg. 812, r. column). In this way Smith discloses input of metadata features representing clinical significance of the variant. Smith does not disclose metadata features representing variant type, amino acid impact, or evolutionary conservation.
Xiong teaches application of outputs to identify drug targets and patients having similar drug response (para. 0034), which necessarily involves correlation of variants with clinical significance features. Xiong further teaches that that position-dependent features may include evolutionary conservation scores (para. 0093). Xiong does not teach consideration of variant type; or amino acid impact.
Matsuguchi teaches identifying one or more somatic mutations (para. 0044). Matsuguchi does not teach consideration of variant type; or amino acid impact.
Robison discusses identification of disease-causing genetic variants by machine learning classification (Abstract) and teaches annotation, filtering and probabilistic modeling including annotating single nucleotide variants (SNVs) and insertion/deletions (para. 0047), i.e., specifying variant type. Robison further teaches that features may include a likelihood value indicating that an amino acid substitution is associated with a disruption of the protein of the observed variant (para. 0033), i.e., amino acid impact. Robison teaches that their methods may be used to help discover the prevalence of genetic diseases and decipher which genes are actually contributing to genetic change (para. 0054).
With respect to claim 4, Robison teaches that features may include a likelihood value indicating that an amino acid substitution is associated with a disruption of the protein of the observed variant (para. 0033), i.e., is a nonsynonymous variant that changes a codon so as to produce a new codon which codes for a different amino acid.
With respect to claim 5, Robison teaches annotation, filtering and probabilistic modeling including annotating single nucleotide variants (SNVs) and insertion/deletions (para. 0047), i.e., specifying whether a variant is a single-nucleotide polymorphism, an insertion, or a deletion. Robison teaches that their methods may be used to help discover the prevalence of genetic diseases and decipher which genes are actually contributing to genetic change (para. 0054).
With respect to claim 11, Robison teaches that features may include a likelihood value indicating that an amino acid substitution is associated with a disruption of the protein of the observed variant (para. 0033).
With respect to claims 19 and 21, Robison teaches determining scores for a variant, classifying the variant, and rendering a human-readable annotation with links to external supporting evidence (paras. 0046-47), i.e., a report that validates that the variant comprises a germline or somatic variant using the classification scores.
An invention would have been obvious to one of ordinary skill in the art if some teaching in the prior art would have led that person to combine prior art reference teachings to arrive at the claimed invention. Before the effective filing date of the claimed invention, said practitioner would have implemented features specifying variant type (i.e., whether a variant is a single-nucleotide polymorphism, insertion or deletion) and amino acid impact, as taught by Robison, to enhance the neural network-implemented system of Smith, in view of Xiong and Matsuguchi, because Robison teaches that annotation of these features allows for discovery of genes that contribute to genetic change and disease (para. 0054). Said practitioner would have had a reasonable expectation of success because Smith, Xiong and Robison all discuss methods of classifying genetic variants by applying algorithmic techniques to sequencing data.
In this way the disclosure of Smith, in view of Xiong, Matsuguchi and Robison, makes obvious the limitations of claims 2, 4-5, 11, 19 and 21. Thus, the claimed invention is prima facie obvious.
Claims 8, 10 and 12 are rejected under 35 USC § 103 as being unpatentable over Smith, in view of Xiong and Matsuguchi, as applied to claims 1 and 3 above, and further in view of Jia et al (PLoS ONE 8(6): Article e64683; published 6/3/2013; previously cited). The new grounds of rejection presented herein were necessitated by Applicant’s amendment of the claims (filed 12/9/2025) to introduce new limitations.
With respect to claims 8 and 12, Smith discloses analysis of allelic abundance and information about known germline heterozygous SNPs in the neighborhood, i.e., frequencies of the variant (pg. 810, l. column). Smith does not disclose consideration of frequencies of the variant in ethnic sub-populations stratified from sequenced populations; or ethnicity prediction.
Xiong teaches that relevance score sequences may be obtained from population allele frequency sequences (para. 0009), and that position-dependent tracks may include tracks of common and rare mutations in human populations (para. 0093). Xiong does not teach implementation of relevance scores representing frequencies of the variant in ethnic sub-populations stratified from sequenced populations; or ethnicity prediction.
Matsuguchi teaches identifying somatic mutations (para. 0044), and discusses consideration of aggregated population data (para. 0182). Matsuguchi does not teach consideration of frequencies of the variant in ethnic sub-populations stratified from sequenced populations; or ethnicity prediction.
Jia discusses analysis of imputed DNA sequence variation, within human leukocyte antigen (HLA) genes, and disease susceptibility (pg. 1, Abstract). Jia teaches imputation of alleles using three haplotype panels, representing loci in Europeans (CEU/CEPH), African (YRI) and East-Asian (CHB+JPT) populations, using population-stratified public genotype data (pg. 3, r. column; pg. 7, Table 3 and caption). In this way, Jia teaches determination of variant frequencies in ethnic sub-populations stratified from sequenced populations and ethnicity predication.
Jia further teaches that imputation quality is inconsistent when imputing HLA variants in non-European populations using a predominantly European reference panel (pg. 8, l. column).
With respect to claim 10, Jia teaches that HLA genes are among the most polymorphic in the human genome, and HLA alleles have large effect sizes in autoimmune diseases, infectious diseases, severe drug reactions, and transplant medicine that dwarf those of many other disease-associated variants (pg. 1, l. column).
An invention would have been obvious to one of ordinary skill in the art if some teaching in the prior art would have led that person to combine prior art reference teachings to arrive at the claimed invention. Before the effective filing date of the claimed invention, said practitioner would have implemented prediction of ethnicity of a tested sample, as taught by Jia, to improve the analytical method of Smith, in view of Xiong and Matsuguchi, because Jia teaches that variant frequencies differ between ethnic sub-populations (pg. 3, r. column; pg. 8, l. column). Thus, the teachings of Jia indicate that the validity of statistical inferences based on variant frequency depends on the concordance of ethnicity between the source of a tested sample and the source of reference resources.
Additionally, said practitioner would have implemented analysis of histocompatibility of a variant, because Jia teaches that HLA genes have numerous variants and relatively large effect sizes in many clinical conditions including transplant medicine (pg. 1, l. column). The teachings of Jia thus indicate that the histocompatibility of a variant can be an important aspect of its clinical significance.
Said practitioner would have had a reasonable expectation of success because Smith, Xiong, Matsuguchi and Jia all discuss classification methods based on algorithmic analysis of sequencing data.
In this way the disclosure of Smith, in view of Xiong, Matsuguchi and Jia, makes obvious the limitations of claims 8 and 12. Thus, the claimed invention is prima facie obvious.
Claim 15 is rejected under 35 USC § 103 as being unpatentable over Smith, in view of Xiong and Matsuguchi, as applied to claim 1 above, and further in view of Patterson et al (Deep learning: A practitioner’s approach. O’Reilly Media, Inc.; published July 2017; previously cited). The new grounds of rejection presented herein were necessitated by Applicant’s amendment of the claims (filed 12/9/2025) to introduce new limitations.
With respect to claim 15, Smith discloses estimation of probabilities based on fitted statistical models (pg. 809, r. column – pg. 810, l. column). Smith also discloses analysis and annotation of oncogenic mutations (pg. 809, r. column). Smith does not disclose training a neural network end-to-end on a first dataset and on a second dataset of examples of cancer-causing mutations.
Xiong teaches a neural network-implemented system wherein all neural network components are trained end-to-end (para. 0098, Figure 4). As one of ordinary skill in the art would have been aware, applying a neural network to perform the algorithmic function, disclosed by Smith, of annotating oncogenic mutations would necessarily involve training the neural network on examples of oncogenic mutations. Thus, the combined teachings of Smith and Xiong are considered to make obvious training a neural network end-to-end on training examples from a dataset of cancer-causing mutations.
However, Xiong does not specifically teach training a neural network on training examples from a first dataset, followed by training on training examples from a second dataset.
Matsuguchi teaches performance of analysis via a trained machine learning algorithm, and exemplifies neural networks (paras. 0134-35). Matsuguchi does not teach training a neural network on training examples from a first dataset, followed by training on training examples from a second dataset.
Patterson reviews machine-learning techniques and teaches pre-training on a base dataset and re-training on a narrower dataset, to save time and required processing power (pp. 551, 553-54). Patterson further teaches that training parameters (e.g., number of training examples, epochs and batch size) are routinely optimized and task-specific (pp. 493 and 495).
An invention would have been obvious to one of ordinary skill in the art if it simply applies known techniques to a known method. Before the effective filing date of the claimed invention, said practitioner would have optimized the training regimen, as Patterson indicates to be standard in the art (pp. 493 and 495), that is utilized by a neural network implementation of the analytical method of Smith, in view of Xiong and Matsuguchi. Said practitioner would have had a reasonable expectation of success because Xiong, Matsuguchi and Patterson all discuss training of neural networks.
In this way the disclosure of Smith, in view of Xiong, Matsuguchi and Patterson, makes obvious the limitations of claim 15. Thus, the claimed invention is prima facie obvious.
Conclusion
At this point in prosecution, no claim is allowed.
The following prior art, made of record and not relied upon, is considered pertinent to applicant's disclosure:
Pisces 5.1.3 Design Document (published 5/5/2016; see PTO-892 for full citation) is online documentation for version 5.1.3 of a publicly-available algorithmic variant calling tool, called Pisces, that utilizes input tumor-only sequencing data to perform functions of: extracting candidate alleles; calculating allele frequency; calculating a quality score that represents error probability; calling each allele as ‘variant’ or ‘reference’; and outputting annotated variant and reference calls.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Theodore C. Striegel whose telephone number is (571)272-1860. The examiner can normally be reached Mon-Fri 12pm-8pm ET.
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 M. Wise can be reached at (571)272-2249. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/T.C.S./Examiner, Art Unit 1685
/JESSE P FRUMKIN/Primary Examiner, Art Unit 1685 May 1, 2026
Prosecution Insights