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
Original claims 1-20 filed 5/2/2022 are pending
Election/Restriction
Applicant’s election without traverse of Group I, claims 1-11, in the reply filed on 12/15/2025 is acknowledged. Upon initial search and consideration of the claim requirements, it appears that it would not be an undue burden to examine both the method and system together. Accordingly, the restriction is withdrawn.
Claims 1-20, drawn to a method of identifying a variant among sequence reads suing a machine learning model and a system.
Priority
This application filed 5/3/2022 is a continuation of PCT/USA2020/058955 filed 11/4/2020 which claims benefit to US provisional application 62/931100 filed 11/5/2019.
Information Disclosure Statement
The 16 information disclosure statements (IDS) submitted on 1/20/2023 through 2/27/2026 are in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
It is noted that several of the IDS provide office actions for other applications and from other countries. See for example, IDSs filed 8/27/2025 and 7/15/2025 both for ref #1. These have been reviewed, however only for what they provide as the applications and claims that are subject to the action are not provided for any context of the examination outlined in the actions. Additionally, some of the websites provided were not attributed to a specific person or source, and they were designated ‘Anonymous’, see IDS filed 9/7/2023 for example.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1-20 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Specifically, claim 1 is unclear as to what reference biological sample is relative to the sequence data or how it corresponds, and the requirements of ‘unavailable’ with respect to the data. It is unclear if this intended to be directed to define the data, the inability to get more sample, and more generally what ‘corresponding’ excludes. With respect to includes non-tumor cells only, it is unclear that using includes how this is to limit the sample source and corresponding data. It is unclear if no cells are present, for example cell free DNA, but which might contain tumor DNA is included or if the starting sample is cells, the cells must be free of tumor cells completely. With respect to using cfDNA, it appears that no tumor cells are present and yet it provides for the analysis of tumor/cancer DNA. For the identification step, it is unclear how a somatic variant and germline variants are to be determined, is this relative some undefined reference or some other reference that defines a sequence to be a variant; and since it is sequence data, it is unclear how variants introduced during the isolation and sequencing process are identified as non-somatic or non-germline variants. The step appears vague and incomplete in how the step is performed. Finally, it is unclear where the machine learning model is derived or more generally how it is trained. It is not clear if the model exists in the art, or if somehow is created and trained, if trained, with what data and how it would provide identification of a variant without any reference point or context for the sequence it is analyzing.
Claim 2 is unclear and appears to be in contrast to the requirements of claim 1 in requiring that the sample is a tumor, when the sample ‘includes non-tumor cells only’.
Claims 9 and 10 appear incomplete because even if a variant is identified, it is unclear how a correlation between that data provides the basis that it is a biomarker or a prognostic marker. The breadth is large for any sample source and then to any type of biomarker for anything and any possible prognostic application. It is unclear if a variant is found it is inherently a marker, or if additional steps are required once a potential variant is identified.
Remaining dependent claims are included because they fail to clarify the issue and only provide known machine learning model names and types without any indication of how they are applied.
For claim 12, the same issues outlined for claims 1-11 apply as the system provides the same method steps. Further, the claim is unclear with respect to a system which appears to be physically a processor and medium, but somehow provides the ability to perform and obtain sequence data from a biological sample.
More clearly providing steps which clearly set forth steps that define the metes and bounds of what is practiced, how models are obtained or trained for the claimed application, and what the sequence data represents would address the basis of the issues.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (i.e., a law of nature, a natural phenomenon, or an abstract idea) without significantly more.
Claim analysis
Claim 1 is generally directed to a method of identifying variants in sequence read data. More specifically, in view of the specification and claim 12 for the system, the claims are directed to a computer implemented method where a sequence is provided, aligned with a reference, differences are observed and termed variants, then using machine learning to identify the differences as candidate variants. Dependent claims provide the source of the sample is a tumor and that known machine learning models are applied to the analysis. In view of the specification and the art of record, the machine learning algorithms appear to provide for a statistical analysis to classify the read data. No specific data to train the models, nor any specific rules or training requirements are defined in the specification or required of the claims.
For step 1 of the 101 analysis, the claims are found to be directed to a statutory category of a method and product
For step 2A of the 101 analysis, the judicial exception of the claims are the steps of analyzing sequence data for possible variants that might exist relative to a reference within the two sequences and identifying the differences as variants. The analysis is performed by machine learning algorythmns. The step of aligning and comparing sequence to arrive at the identification of differences between sequences are instructional steps. The claim may require computing similarity scores with the broad steps of receiving data, comparing and thus determine potential variants sequence. The judicial exception is a set of instructions for analysis of sequence data and appears to fall into the category of Mathematical Concepts to the extent the use of machine learning algorithmns are statistical analysis in nature and in the category of Mental Processes, that is concepts performed in the human mind (including an observation, evaluation, judgment, opinion) where here aligned sequences can be observed to be the same or different, and easily identified variants can be observed. The breadth of “obtaining”, “aligning”, and “outputting” encompasses non-transformative visual assessment of possible differences between sequences. This breadth does not impose a meaningful limit on the claim scope, such that all others are not precluded from using the natural principle of observing differences. To the extent the system is a computer and would implement machine learning algorithmns, the courts have also identified limitations that did not integrate a judicial exception into a practical application; for example, merely including instructions to implement an abstract idea on a computer, or merely using a computer as a tool to perform an abstract idea, as discussed in MPEP § 2106.05(f). The recited breeding crosses are interpreted as naturally occurring plant crosses. Computing, constructing datasets and using statistical models was well understood, conventional, and routinely performed in the art at the time the application was filed. Furthermore, the limitation of genotyping for potential types of markers or a plurality of markers or alleles does not change the steps to be performed. See MPEP § 2106.05(g) for a discussion on adding insignificant extra-solution (both pre-solution and post-solution) activity to the judicial exception. See also MPEP § 2106.05(h) for a discussion on generally linking the use of a judicial exception to a particular technological environment or field of use. The claims appear to fall into the category of Mathematical Concepts, as it applies the use of statistics and mathematical relationships in analyzing probabilities, and also into the category of mental processes, as concepts performed in the human mind (including an observation, evaluation, judgment, opinion) because there is no apparent complexity to or amount of data that is collected and analyzed as presently claimed.
Recent guidance from the office requires that the judicial exception be evaluated under a second prong to determine whether the judicial exception is practically applied. In the instant case, the claims do not have an additional element to which the analysis is applied, just outputting what is observed. This judicial exception requires steps recited at high level of generality and are only stored on a non-transitory, and is not found to be a practical application of the judicial exception as broadly set forth.
For step 2B of the 101 analysis, each of the independent claims recites additional elements and are found to be the steps of obtaining sequence data. As such, the claims do not provide for any additional element to consider under step 2B. It is noted that in explaining the Alice framework, the Court wrote that "[i]n cases involving software innovations, [the step one] inquiry often turns on whether the claims focus on the specific asserted improvement in computer capabilities or, instead, on a process that qualifies as an abstract idea for which computers are invoked merely as a tool." The Court further noted that "[s]ince Alice, we have found software inventions to be patent-eligible where they have made non-abstract improvements to existing technological processes and computer technology." Moreover, these improvements must be specific -- "[a]n improved result, without more stated in the claim, is not enough to confer eligibility to an otherwise abstract idea . . . [t]o be patent-eligible, the claims must recite a specific means or method that solves a problem in an existing technological process." As indicated in the summary of the judicial exception above and in view of the teachings of the specification, the steps are drawn to analysis of sequence data. While the instruction are stored on a medium and could be implemented on a computer, together the steps do not appear to result in significantly more than a means to compare sequences. The judicial exception of the method as claimed can be performed by hand and in light of the previous claims to a computer medium and in light of the teaching of the specification on a computer. In review of the instant specification the methods do not appear to require a special type of processor and can be performed on a general purpose computer. Dependent claims set forth additional steps which are more specifically define the considerations and steps of calculating, and comparing, and do not add additional elements which result in significantly more to the claimed method for the analysis. It is noted that while the claims set forth or imply information about the sequences being analyzed (that they are ‘markers’), this is only description of the data being analyzed and context and user defined. As such, the instant claims set forth an inventive concept that are drawn only to an abstract process that only manipulates data and, therefore, are not directed to statutory subject matter. No additional steps are recited in the instantly claimed invention that would amount to significantly more than the judicial exception. Without additional limitations, a process that employs mathematical algorithms (aligning sequences) to manipulate existing information (identify a probe that starts at a deletion or insertion) to generate additional information is not patent eligible. Furthermore, if a claim is directed essentially to a method of calculating, using a mathematical formula, even if the solution is for a specific purpose, the claimed method is non-statutory. In other words, patenting abstract idea (designing probes to a target sequence) cannot be circumvented by attempting to limit the use to a particular technological environment or purpose and desired result.
One way to overcome a rejection for non-patent-eligible subject matter is to persuasively argue that the claimed subject matter is not directed to a judicial exception. Another way for the applicants to overcome the rejection is to persuasively argue that the claims contain elements in addition to the judicial exception that either individually or as an ordered combination are not well understood, routine, or conventional. Another way for the applicants to overcome the rejection is to persuasively argue that the claims as a whole result in an improvement to a technology. Persuasive evidence for an improvement to a technology could be a comparison of results of the claimed subject matter with results of the prior art, or arguments based on scientific reasoning that the claimed subject matter inherently results an improvement over the prior art. The applicants should show why the claims require the improvement in all embodiments.
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.
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 1-20 are rejected under 35 U.S.C. 103 as being unpatentable over Den Akker et al. (A machine learning model to determine the accuracy of variant calls in capture based next generation sequencing BMC Genomics (2018) 19:263 doi.org/10.1186/s12864-018-4659-0), Olson et al. (Best practices for evaluating single nucleotide variant calling methods for microbial genomics. Front. Genet. 6:235. doi: 10.3389/fgene.2015.00235), Spinella et al. (SNooPer: a machine learning-based method for somatic variant identification from low-pass next-generation sequencing BMC Genomics (2016) 17:912 DOI 10.1186/s12864-016-3281-2) and Sun et al (bioRxiv, 17 October 2019, https://doi.org/10.1101/805135).
Claim 1 provides a method of identifying variants in sequence read data, specifically the claims of a method and system comprise and are directed to a computer implemented method where a sequence is provided, aligned with a reference, differences are observed and termed variants, then using machine learning to identify the differences as candidate variants. Dependent claims provide the source of the sample is a tumor and that known machine learning models are applied to the analysis. The claims provide no specific data to train the models, nor any specific rules or training requirements are defined in the specification or required of the claims. Encompassed by the claims are the steps of obtaining sequence data representing tumor and non-tumor sources, aligning the plurality of reads to a reference, identifying differences between the read and reference as potential variants, using machine learning model to identify somatic variants and reporting what is observed.
Den Akker et al is provided as evidence that read data can be obtained, aligned with known references for identification and that machine learning models can be applied to the data to determine the accuracy of variant calls. Den Akker et al. provide analysis and fine tuning to their model to improve the accuracy, in particular for low pass data from NGS data. Den Akker et al conclude that the work shows that NGS data contains sufficient characteristics for a machine-learning-based model to differentiate low from high confidence variants, and that it reveals the importance of incorporating site-specific features as well as variant call features in such a model. Olsen et al. provide for a review of methods used to evaluate variants in read data. It is noted that the Olsen et al. analyze microbial DNA, however the methods and potential complexity of the DNA from microbial sources is provided to demonstrate that analysis methods could successfully be applied to complex read data. Olsen et al. highlight the importance of validation in the analysis for correct variant calling and need for standards. Spinella et al. is provided as additional evidence for the application of machine learning and the ability to identify somatic variants in NGS, in particular low pass read data. Spinella et al. specifically discuss the application of variant detection for cancer with the goal of applying the analysis to any cancer sequencing project to identify a relevant, and limited, set of somatic variants for further sequence/functional validation. They note that the inherently complex nature of cancer genomes combined with technical issues directly related to sequencing and alignment can affect either the specificity and/or sensitivity of most callers, and detail the work in creating the machine learning model SNooPer’s which is based in part on random forest in order to protect against technical bias and systematic errors because it demonstrated an ability in that it does not rely on user-defined parameters. Finally, Sun et al. if provided for methods using machine learning to detect variants in complex read data representing tumor sources. Sun discloses A method of determining tumor purity comprising: Abstract: “Here we present a machine learning model, Deep Purity (DePuty) that leverages convolutional neural networks to accurately predict tumor purity from next-generation sequencing data from clinical samples without matched normals.”; Page 3, first paragraph: “Using genomic data from ovarian cancer clinical FFPE tumor samples sequenced with the Foundation Medicine FoundationOne CDx panel [13], we first created a training ‘ground truth’ dataset by manually annotating 102 patient samples for purity. This was done by visualizing scatterplots of the data (Fig. 1)…”. Page 3, third paragraph: “With scatterplots, individual datapoints tend to cluster in space at certain allele frequencies or log copy number ratios, despite the fact that nearest neighbors in chromosome order are not necessarily similarly valued (Fig. 1). This representation therefore naturally yields a 2D histogram (Fig. 1B)…”. As seen in Fig. 1B, minor allele frequency is plotted for loci throughout the genome. Therefore, sequencing data must have been obtained for “a plurality of nucleic acid molecules”. As seen in Fig. 1B, the sequencing information is plotted according to chromosomal location. This represents an aligning of the nucleic acid sequence data to a reference genome. For identifying, based on the aligned nucleic acid sequence data, a set of genomic regions, wherein each genomic region of the set of genomic regions includes one or more nucleotide-sequence variants relative to a corresponding genomic region of the reference genome; As seen in Fig. 1B, minor allele frequency is measured for a plurality of genomic loci (regions). Fig. 1B, being a 2D histogram of minor-allele frequency, is a frequency distribution. See page 3, third paragraph: “With scatterplots, individual datapoints tend to cluster in space at certain allele frequencies or log copy number ratios, despite the fact that nearest neighbors in chromosome order are not necessarily similarly valued (Fig. 1). This representation therefore naturally yields a 2D histogram (Fig. 1B)…”. More specifically, for frequency distribution using a trained machine-learning model to estimate a metric identifying tumor purity in the biological sample; Abstract: “As input, our model utilizes SNP-based copy number and minor allele frequency data formulated as a scatterplot image.” Page 2, second paragraph: “We therefore reasoned that for unstructured data of copy number variation and minor allele frequency, a human-suitable representation coupled with a CNN may yield performance that exceeds current statistical models. We built and trained a CNN regression model to output tumor purity based upon the very scatterplot images of input data that an expert human annotator would use for prediction.” Page 4, first full paragraph: “We next optimized the scatterplot representation for our model. We systematically varied the amount of spatial binning (scatterplot image size) and the clipping threshold (scatterplot dot size) to minimize the MAE of an OLS linear regression. After optimizing our representation, we trained a simple convolutional neural network to predict tumor purity.” and outputting the data as evidenced in the publication, for example page 2, last paragraph: “We built and trained a CNN regression model to output tumor purity based upon the very scatterplot images of input data that an expert human annotator would use for prediction.”
For the instant system claims, the teachings of Sun have been discussed. While Sun did not specifically disclose that the method was performed on a “system” comprising a processor and computer-readable medium comprising instructions for carrying out the computational steps, it would have been prima facie obvious to one of ordinary skill in the art prior to the effective filing date of the application to do so, and one of ordinary skill would have assumed this was how Sun implemented the method, since “machine learning” implies a computer processor receiving input data. It would also have been obvious to report the identified variants, since these would be tumor-associated (i.e. somatic mutations), and Sun disclosed that “it is critical to find novel disease-related genetic variants from patient genomic next-generation sequencing data to identify disease biomarkers” (page 2, first paragraph). Sun also disclosed that the method for determining tumor purity “increased the number of correct germline versus somatic variant determinations” (page 5, lines 1-2).
Given the state of the art and specific detailed guidance for the application of machine learning to analysis of read data and the use of machine learning models to identify variants it would have been prima facie obvious to one having ordinary skill in the art at the time the invention was made to use machine learning and develop models which were accurate and which could be applied to read data as provided in each of the citations. One having ordinary skill in the art would have been motivated to approach complex read data, for example DNA reads from heterogeneous tumor cell sources or ones that contain both normal and tumor DNA, such as cfDNA/ctDNA, in the manner each citation provides and provide steps of obtaining read data, identifying relevant reads by alignment and applying machine learning to identify variants that might exist. Given the detailed guidance and success of each, there would have been a reasonable expectation of success to apply machine learning models to the analysis of read data to identify potential variants.
Thus, the claimed invention as a whole was clearly prima facie obvious.
Conclusion
No claim is allowed.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Joseph T Woitach whose telephone number is (571)272-0739. The examiner can normally be reached Mon-Fri; 8:00-4:00.
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, Karlheinz R Skowronek can be reached at 571 272-9047. 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.
/Joseph Woitach/Primary Examiner, Art Unit 1687
Prosecution Insights