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 .
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and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114.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 10/02/2025 has been entered.
Status of Claims
Claim 7 is cancelled.
Claims 1-6 and 8-20 are pending and are examined on the merits.
Priority.
Applicant’s claim for the benefit of a prior-filed application under 35 U.S.C. 119(e) or under 35 U.S.C. 120, 121, or 365(c) is acknowledged. Priority of REPUBLIC OF KOREA application 10-2024-0008609 filed 1/19/2024 is acknowledged.
Withdrawn Rejections/Objections
The rejection to claims 1-20 under 35 U.S.C. 112(b) in the Office action posted 23 July 2025 are withdrawn in view of claim amendments filed 2 October 2025.
The rejection to claims 1-6 and 14-20 under 35 U.S.C. 102(a)(1) in the Office action posted 23 July 2025 are withdrawn in view of claim amendments filed 2 October 2025.
Claim Rejections - 35 USC § 112 –First Paragraph
This is a newly established rejection.
The following is a quotation of the first paragraph of 35 U.S.C. 112(a):
(a) IN GENERAL.—The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.
The following is a quotation of the first paragraph of pre-AIA 35 U.S.C. 112:
The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.
Claims 1-6 and 8-20 are rejected under 35 U.S.C. 112(a) or 35 U.S.C. 112 (pre-AIA ), first paragraph, as failing to comply with the written description requirement. The claim(s) contains subject matter which was not described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor or a joint inventor, or for applications subject to pre-AIA 35 U.S.C. 112, the inventor(s), at the time the application was filed, had possession of the claimed invention.
Amended claims 1, 17 and 20 recite (emphasis added):
“identifying, based on the predicting how the individual will respond to treatment using the drug, a quantity of the drug ; and
causing the quantity of the drug to be administered to the individual”
at the last two steps. The disclosure never described drug quantity, drug amount, or drug dosage.
Further, an active treatment to the individual using drug is not supported by the disclosure. Although the disclosure mentioned “drug” ten times, they are all about information related to drug or drug treatment.
Further, if “causing the quantity of the drug to be administered to the individual” is interpreted as an active step, in the framework of a processor executed method as recited in claims 1 and 17, or in the framework of an apparatus as recited in claim 20, how such an active step is performed on an individual (human being) is not described in the disclosure.
Claims hence recite new matter.
Claim Rejections - 35 USC § 112 –Second Paragraph
This is a newly established rejection.
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 1-6 and 8-20 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.
Amended claims 1, 17 and 20 recite (emphasis added):
“causing the quantity of the drug to be administered to the individual.”
This step is indefinite because it is not clear whether an active administration of drug is applied to the individual. Further, it is not clear what causes the drug to be administered to the individual.
To advance compact prosecution, the last step in claims 1, 17 and 20 are interpreted as an intended use of the claimed invention or a field of use limitation.
Claim Rejections - 35 USC § 101
This rejection is maintained from a previous Office action. Modifications are necessitated by claim amendments.
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 1-6 and 8-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to non-statutory subject matter.
Step 1: Process, Machine, Manufacture or Composition
Claims 1-6 and 8-16 are directed to a process, here a "method," for training a machine learning model, with process steps "receiving”, “detecting", “generating”, and “training”.
Claims 17-18 are directed to another process, here another “method”, with process steps “receiving”, “determining”, and “performing”.
Claim 19 is directed to a 101 machine, here “a non-transitory computer-readable storage medium”, with known structure.
Claim 20 is directed to another 101 machine, here “an apparatus”, with structural components like “at least one processor”, and “a memory”.
Step 2A Prong One: Identification of Abstract Ideas
Claims 1, 17 and 20 recite:
Detecting, based on the non-tumorous sequencing data and the tumorous sequencing data, a reference variant candidate in a reference sample comprising the non-tumor-derived sample and the tumor-derived sample;
----This step recites a data analysis process that leads to a reference variant candidate. Under a broadest reasonable interpretation (BRI), this step requires sequence comparison and decision-making. Hence this step equates to an abstract idea of mental processes.
Generating, based on the POF and the reference variant candidate, annotation information comprising one or more of:
A number of samples, of the plurality of FFPE samples, associated with a variant allele frequency (VAF), at a position in a base sequence in the samples, less than a predetermined threshold; or
A number of samples, among the plurality of FFPE samples, having a predetermined number of variant reads at a predetermined position;
----This step recites data manipulation activities of adding annotation information (to the reference variant candidate) through cross-referencing, which equates to an abstract idea of mental processes.
Generating training data based on:
the reference variant candidate;
the annotation information;
second non-tumorous sequencing data of the individual; and
second tumorous sequencing data, of the individual, that corresponds to the second sample type, wherein labels of the training data comprise classification information corresponding to the reference variant candidate;
----This step recites data manipulation activities that extract and re-organize data in the way required by the machine learning model. Therefore, this step equates to an abstract idea of mental processes.
Training, based on the training data, the machine learning model by adjusting one or more weights of one or more nodes of an artificial neural network.
----This step recites training a machine learning model, which has weighted adjustment of nodes. Under a BRI, the machine learning model is a regression model, which will require mathematical operations. Therefore, this step equates to an abstract idea of mathematical concepts.
Predicting, based on the classification result, how the individual will respond to treatment using a drug associated with one or more tumors of the individual;
----This step recites an decision-making activity based on the data observation on the predicted output. Therefore, this step equates to an abstract idea of mental processes.
Identifying, based on the predicting how the individual will respond to treatment using the drug, a quantity of the drug ; and
----This step recites a decision-making activity based on the data observation on the predicted output. Therefore, this step equates to an abstract idea of mental processes.
Causing the quantity of the drug to be administered to the individual.
----This step is interpreted as an intended use of the claimed invention or a field of use limitation. An intended use of the claimed invention reads on mental activities. Therefore, this step equates to an abstract idea of mental processes.
Step 2A Prong Two: Consideration of Practical Application
The claims result in a process of:
“identifying, based on the predicting how the individual will respond to treatment using the drug, a quantity of the drug”; and
“causing the quantity of the drug to be administered to the individual.”
The first step reads on a judgement/decision-making process based on data (ML prediction output, more specifically validated variants) observation, which equates to an abstract idea of mental processes. The second step is interpreted as merely an intended application of the claimed invention or a field of use limitation, it cannot integrate a judicial exception under the "treatment or prophylaxis" consideration (MPEP §2106.04(d)(2)), because it is not a particular treatment or prophylaxis (no disease, no drug is specified).
The claims do not recite any additional elements that integrate the abstract idea/judicial exception into a practical application.
This judicial exception is not integrated into a practical application because the claims do not meet any of the following criteria:
An additional element reflects an improvement in the functioning of a computer, or an improvement to other technology or technical field;
an additional element that applies or uses a judicial exception to effect a particular treatment or prophylaxis for a disease or medical condition;
an additional element implements a judicial exception with, or uses a judicial exception in conjunction with, a particular machine or manufacture that is integral to the claim;
an additional element effects a transformation or reduction of a particular article to a different state or thing; and
an additional element applies or uses the judicial exception in some other meaningful way beyond generally linking the use of the judicial exception to a particular technological environment, such that the claim as a whole is more than
a drafting effort designed to monopolize the exception.
Step 2B: Consideration of Additional Elements and Significantly More
The claimed method also recites "additional elements" that are not limitations drawn to an abstract idea. The recited additional elements are drawn to:
Claim 1 recite:
receiving: non-tumorous sequencing data based on a non-tumorous sample of an individual; and tumorous sequencing data based on a tumorous sample, of the individual, that corresponds to a first sample type processed differently from a second sample type, wherein a plurality of artifacts are associated with the first sample type;
----This step recites receiving two datasets, which equates to an additional element.
receiving a Panel of FFPEs (POF) generated based on sequencing data associated with a plurality of Formalin-Fixed, Paraffin-Embedded (FFPE) samples, wherein the plurality of FFPE samples corresponds to the first sample type;
----This step recites receiving one dataset, which equates to an additional element.
Executing the trained machine learning model by:
providing, as input to an input node of the artificial neural network of the trained machine learning model, input comprising:
information corresponding to a first variant candidate of a first sample, and a feature of the first variant candidate in the first sample; and
receiving, as output from an output node of the artificial neural network of the trained machine learning model and based on the input, a classification result indicating whether the first variant candidate is a true positive or an artifact;
----This step recites prediction input/output, which equates to an additional element.
Claim 19 recites:
A non-transitory computer-readable storage medium.
----this step recites a storage medium, which equates to an additional element.
Claim 20 recites:
An apparatus, comprising: at least one processor;
----this step recites one processor, which equates to an additional element.
and a memory storing instructions that, when executed, configure the at least one processor to:
----this step recites a memory, which equates to an additional element.
The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception because it is routine and conventional to perform the acts of acquiring sequencing data for further analysis. Other elements of the method include processor, storage memory and an ANN model, which are recitations of generic computer components that serve to perform generic computing functions that are well-understood, routine, and conventional activities previously known to the pertinent industry.
Particularly, the courts have recognized the following laboratory techniques as well-understood, routine, conventional activity in the life science arts when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity (MPEP 2106.05(d).II):
v. Analyzing DNA to provide sequence information or detect allelic variants, Genetic Techs. Ltd., 818 F.3d at 1377; 118 USPQ2d at 1546;
Similarly, the current claims are about analyzing DNA to provide sequence information and detect allelic variants.
Viewed as a whole, applying machine learning technology to variant calling is taught by at least the following listed art, as discussed in the previously Office action:
Heo, Dong-hyuk, et al. "Reducing artifactual somatic variant calls from formalin-fixed paraffin-embedded specimens by using DEEPOMICSⓇ FFPE, a bioinformatic approach based on deep neural networks." (2023). Previously cited.
Wu, Chao, et al. "Using machine learning to identify true somatic variants from next-generation sequencing." Clinical chemistry 66.1 (2020): 239-246. Previously cited.
Dodani, et al. ("Combinatorial and machine learning approaches for improved somatic variant calling from formalin-fixed paraffin-embedded genome sequence data." Frontiers in Genetics 13 (2022): 834764. Previously cited).
Reference A, B and C applied different machine learning models to call variants from next generation sequencing data. Reference A specifically tried to improve variant calls from FFPE samples and reference B specifically tried to avoid FFPE samples (Wu: last para, col 1, pg. 240) as it is known that FFPE samples harbor artifacts of variants. Reference C performed an in-depth comparison of somatic SNVs called on matching FF and FFPE Whole Genome Sequence (WGS) samples extracted from the same tumor, and illustrates that when using the correct variant calling strategy, the majority of clonal SNVs can be recovered in an FFPE sample with high precision and sensitivity (De Schaetzen van Brienen: section Abstract, pg. 1).
Viewed as a whole, these additional claim element(s) do not provide meaningful limitation(s) to transform the abstract idea recited in the instantly presented claims into a patent eligible application of the abstract idea such that the claim(s) amounts to significantly more than the abstract idea itself. Therefore, the claim(s) are rejected under 35 U.S.C. 101 as being directed to non-statutory subject matter.
Response to Applicant’s Argument
In the Remarks filed 2 October 2025, Applicant argued (page 13, 2nd para) that “there is a key distinction between claims directed to an exception (which require further eligibility analysis) and claims that merely involve an exception (which are eligible and do not require further eligibility analysis). Id. The Action confuses this concept: it mistakes the hypothetical presence of an exception (e.g., that the claims might in some attenuated way be performable by a human and/or mathematically-related) as indicating that the claims are, as a whole, directed to the exception. This is simply not the case. The present claims involve steps performed by computing devices”.
To response, Applicant’s argument refers to Step 2A/Prong one in the 101 analysis, relating to whether claims recite abstract ideas or not. Applicant’s argument is not persuasive. Examination relies on MPEP and BRI in the 101 analysis, as discussed above in the 101 rejection.
Further, the instant claims recite the use of a generic computer in the form of processors and storage medium. The configured computer defined only by the expected functions that it is to perform. This merely "applies” the abstract data analysis using generic computer elements at a high level of generality. The courts have been clear that even though the analysis may be for a specific purpose, such processes are not directed to a statutory invention and properly is identified as being directed to the judicial exception in the form of an abstract idea.
In the Remarks, Applicant argued (page 14, 2nd para) that amended claims now recite:
"administration of a "particular treatment' when considered in the context of a claim as a whole." For instance, the claims here now recite use of unique and specifically-processed machine learning techniques to predict "how the individual will respond to treatment using a drug associated with one or more tumors of the individual' and then "identify[], based on the predicting how the individual will respond to the treatment using the drug, a quantity of the drug; and caus[e] the quantity of the drug to be administered to the individual." This is far beyond the routine application of such treatments, and thus imposes a meaningful limitation on any alleged abstract idea.
To response, Applicant’s argument refers to Step 2A/Prong two in the 101 analysis, relating to whether claims are integrated into a practical application or not. Applicant’s argument is not persuasive. The claimed treatment is merely an intended use of the claimed invention or a field of use limitation. There is no particular treatment recited; neither the disease nor the drug are claimed. Therefore, the limitation of “causing” cannot integrate a judicial exception under the "treatment or prophylaxis" consideration (MPEP §2106.04(d)(2)). Also see the 112(b) rejection above.
In the Remarks, Applicant argued (page 14, last para through page 15, 1st para) that at Step 2B, the claims add significantly more than any alleged abstract ideas as:
“generic computer components are able in combination to perform functions that are not merely generic." May 2016 Subject Matter Eligibility Update at 4 (emphasis added). The present independent claims do precisely this by reciting a process involving, for example, training machine learning models to process samples "how the individual will respond to treatment using a drug associated with one or more tumors of the individual' using machine learning information and then "identifying, based on the predicting how the individual will respond to the treatment using the drug, a quantity of the drug; and causing the quantity of the drug to be administered to the individual."
In response, Applicant’s argument is not persuasive. The claim(s) does/do not include additional elements that are sufficient to amount to significantly more than the judicial exception because the claims do not include additional elements that are sufficient to amount of significantly more than the judicial exception. It is routine and conventional to perform the acts of acquiring sequencing data for further analysis. Other elements of the method include processor and storage memory, which are recitation of generic computer components that serve to perform generic computer functions that are well-understood, routine, and conventional activities previously known to the pertinent industry. The argued recitation:
"how the individual will respond to treatment using a drug associated with one or more tumors of the individual' using machine learning information and then "identifying, based on the predicting how the individual will respond to the treatment using the drug, a quantity of the drug; and causing the quantity of the drug to be administered to the individua,."
is all interpreted as additional information that can be deduced from the model prediction output, which are directed to abstract idea of mental processes. “Causing the quantity of the drug to be administered to the individual" is interpreted as an intended application, which reads on an intention and is directed to an abstract idea of mental processes. Under another BRI and at an alternative embodiment, “causing the quantity of the drug to be administered to the individual" is interpreted as a physical thing of administering drug to a patient. Then it is still routine and conventional. Without being specific with the disease, and the drug, generic treatments using drugs (at a certain quantity) are happening every day. So it is routine and conventional
Hence, the 101 rejection is maintained.
Claim Rejections - 35 USC § 103
This rejection is a new rejection necessitated by claim amendments.
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.
Claim(s) 1-6 and 8-20 are rejected under 35 U.S.C. 103 as being unpatentable over Dodani, et al. ("Combinatorial and machine learning approaches for improved somatic variant calling from formalin-fixed paraffin-embedded genome sequence data." Frontiers in Genetics 13 (2022): 834764. Previously cited), in view of Chan, et al. ("The roles of common variation and somatic mutation in cancer pharmacogenomics." Oncology and therapy 7.1 (2019): 1-32. Newly cited).
Claim 1 is interpreted as a method for training a machine learning model for calling variants out of next generation sequence.
Regarding claim 1, Dodani disclosed methods for somatic variant calling using machine learning approaches (Dodani: Title and Abstract, page 1) that is implicitly implemented in computers. More specifically,
Dodani provides “Feature extraction is performed using a modified version of DeepSVR (Ainscough et al., 2018) and bam-readcounts3 that require a tumor BAM and a VCF file (either tumor-only or tumor-normal paired) as input” (3rd para lines 5-9, col 1, pg. 4), which teaches acquiring sequence data from paired normal and abnormal (tumor) specimen of cancer patients and generating annotation information.
Dodani further provides “we analyzed the overlap between somatic variants using the FFPE and FF tumors to evaluate and optimize the recallest and precisionest of somatic SNV calls from FFPE tissue samples” (3rd para lines 1-3, col 2, pg. 4), which teaches the FFPE (formalin-fixed, paraffin-embedded) and FF (fresh frozen) tissues originated tumor sequence.
Dodani further provides “the fixation process used to produce FFPE samples creates nucleic acid damage that presents unique challenges to accurate and comprehensive whole-genome sequencing (WGS) analyses (Robbe et al., 2018a; Haile et al., 2019). Methods to improve the extraction of nucleic acids from FFPE tissues have emerged (Haile et al., 2017; Haile et al., 2019). Still, FFPE-induced artifacts, such as formaldehyde crosslinks, DNA fragmentation, abasic sites, and deamination of cytosine bases (Do and Dobrovic, 2015; Haile et al., 2017), remain problematic and can confound the identification of somatic single nucleotide variants (SNVs)” (last para last 4 lines, page 1 through 1st para lines 1-4, col 1, pg. 2), which teaches the FFPE samples are associated with artifacts.
Dodani further provides “each patient sample in our study had an FFPE tumor, FF tumor, and FF normal genome available for analysis. For our feasibility testing, we attempted to call somatic FFPE tumor and FF tumor variants in a single sample using 10 variant callers: LoFreq (Wilm et al., 2012), Pisces (Dunn et al., 2019), SomVarIUS (Smith et al., 2016), Platypus (Rimmer et al., 2014), Shimmer (Hansen et al., 2013), outLyzer (Muller et al., 2016), Strelka2 (Saunders et al., 2012), Virmid (Kim et al., 2013),
Octopus (Cooke et al., 2018) and Mutect2 (McKenna et al., 2010)” (last para, col 2, pg. 2 through 1st para, col 1, pg. 3), which teaches detecting variant candidates based on comparison paired tumor-normal samples.
“POF” is interpreted as sample and data related descriptions and annotations. Dodani further provides (page 2, col 2, 4th para) “metrics including error rate, coverage, insert size, mapping quality, percentage of reads with insertions or deletions and GC bias were estimated and extracted using Qualimap (2.2.1) (García-Alcalde et al., 2012) and Picard (2.4.1)1. IGV (Robinson et al., 2011) was used for manual read and alignment inspection”, which teaches acquiring sample and data related descriptions and annotations.
Dodani further provides “feature extraction is performed using a modified version of DeepSVR (Ainscough et al., 2018) and bam-readcounts3 that require a tumor BAM
and a VCF file (either tumor-only or tumor-normal paired) as input. A total of 31 features (Supplementary Table S4) are extracted from the tumor bam are used to classify variants in FFPolish and can be divided into three categories:
• Summary metrics (e.g., tumor variant allele fraction (VAF), tumor depth)
• Read-count metrics (e.g., the number of reads on the negative strand) for both the variant and reference allele
• Read-averaging metrics (e.g., the average base quality of reads) for both the variant and reference allele”
(3rd para, col 1, pg. 3), which teaches generating annotation information (here “feature”). The “tumor_VAF” (which stands for “variant allele fraction of tumor reads”) in the Supplementary Table S4 and the VCF file suggest annotation information come from a genetic database.
Additionally, Dodani teaches extract features associated with the sequences. The BAM file has sample information. Hence Dodani also anticipate the following newly added limitation:
receiving a Panel of FFPEs (POF) generated based on sequencing data associated with a plurality of Formalin-Fixed, Paraffin-Embedded (FFPE) samples, wherein the plurality of FFPE samples corresponds to the first sample type; and
generating, based on the POF, the annotation information;
Dodani further provides ““FFPolish is an FFPE variant filtering approach based on a logistic regression model from scikit-learn (Pedrosa, 2011). The model was trained on somatic Strelka2 FFPE calls from both HTMCP cervical and BLGSP cohorts, with a total of 8,698,388 training SNVs. As above, variants were evaluated by comparing FFPE variant calls and the ground truth variants (FF variants from Strelka2 and Mutect2)” (2nd para lines 1-7, col 1, pg. 4), which teaches getting training data based on both the curated normal and abnormal sequencing data (here the second normal/abnormal sequencing data), because the somatic Strelka2 FFPE calls requires normal sequencing data, and then train the FFPolish model.
Dodani teaches execute the machine learning model to acquire the filtered FFPE VCFs output (Fig. 2, page 5).
Dodani provides “As FFPE samples remain central to clinical diagnostics, methods for confidently calling variants in genomic data derived from such samples are required for enhanced utility in clinical and translational research settings” (last para, col 2, pg. 6 through 1st para line 1, col 1, pg. 7), which suggests the genomic profiling might be helpful in disease diagnosis.
Dodani does not teach using the identified variant to prognosis or to come up with a precise treatment plan.
Chan provides (page 10, col 1, last para) “in metastatic colorectal patients treated with FOLFIRI regimen, patients with the ABCG2 rs7699188-GG genotype show decreased tumor response [28]. In a separate study, metastatic colorectal patients carrying AREG rs11942466 C[A and rs9996584 C>T were associated with OS while those carrying EGFR rs712829 G>T were associated with progression-free survival
(PFS) and OS [29]”, and (page 12, col 1 last para through col 2 1st para) “Wang and colleagues performed targeted sequencing to evaluate the link between EGFR
and EGFR-linked pathway gene SNPs with EGFR-TKI response and ADRs in patients with advanced NSCLC [55]. They identified rs1042640 in UGT1A10, rs1060463 and
rs1064796 in CYP4F11 and rs2074900 in CYP4F2 as being associated with erlotinib
treatment response, with improved an median PFS of 12.57 months compared to that of nonresponders (median PFS 3.55 months) [55]. SNP rs1064796 in CYP4F11 and SNP rs10045685 in UGT3A1 were also linked to ADRs, with carriers of CYP4F11-rs1064796-C and UGT3A1-G showing an increased risk for skin rash or digestive track injury [55]”, which teaches cancer variant can be used to predict response to treatment, and how the patient will response to the drug treatment.
Chan provides (page 11, col 1, 2nd para) “one specific variant is rs116855232
(c.415C>T) in NUDT15; this variant shows a distinct population distribution, with a particularly higher occurrence of the rare allele rs116855232-T in East Asians (10%) compared to Hispanics, Europeans and Africans (http://www.internationalgenome. org/1000-genomesbrowsers). A GWAS study on children with ALL reported an association of rs116855232 in NUDT15 (P = 8.8 9 10-9) to mercaptopurine sensitivity in only East Asian patients, with a lower tolerance to mercaptopurine resulting in hematologic toxicities [40]. Patients carrying the rs116855232-TT genotype were less tolerant, with an average dose intensity of 8.3%, compared with those with TC and CC genotypes, who tolerated 63 and 83.5% of the planned dose of 75 mg/m2 per day [40]”, which suggests to come up with a precise treatment plan with drug dosage based on the identified cancer variants.
Regarding claim 20, claim 20 is the “apparatus” version of the claim 1 method. Dodani provides “After initial testing, we eliminated Pisces, outLyzer, and Octopus from further analysis due to computational requirements beyond what we allocated to this project (Supplementary Table S1)” (2nd para lines 1-3, col 1, pg. 3), and “For each patient, variant calls from five tools (LoFreq, Mutect2, Strelka2, Virmid, Shimmer) were each evaluated in isolation and then combined to test for improved FFPE variant calling recallest, precisionest, and F1est. To make the test computationally feasible, ten data sets, each containing results from 3 tools, were used for the analysis” (last para, col 2, pg. 3 through 1st para, col 1, pg. 4), which suggests the computational implementation and the computer processors, memories used in the studies. Therefore, the art taught claim 1 also teaches claim 20.
Regarding Claim 2, Dodani provides “each patient sample in our study had an FFPE tumor, FF tumor, and FF normal genome available for analysis. For our feasibility testing, we attempted to call somatic FFPE tumor and FF tumor variants in a single sample using 10 variant callers: LoFreq (Wilm et al., 2012), Pisces (Dunn et al., 2019), SomVarIUS (Smith et al., 2016), Platypus (Rimmer et al., 2014), Shimmer (Hansen et al., 2013), outLyzer (Muller et al., 2016), Strelka2 (Saunders et al., 2012), Virmid (Kim et al., 2013), Octopus (Cooke et al., 2018) and Mutect2 (McKenna et al., 2010)” (last para, col 2, pg. 2 through 1st para, col 1, pg. 3), which teaches detecting variant candidates based on comparison paired tumor-normal samples.
Regarding Claim 3, Dodani teaches claim 2. Dodani further provides “The comparison of results from the matched FF and FFPE tissues is not without uncertainty, as the cells in each partition may express non-identical biological signals. This uncertainty motivated a rigorous selection of our performance metrics. Using the intersection of FF variant sets from Strelka2 and Mutect2 (ground truth callers) as the control for precision and recall estimates is the standard approach. This would lead to misclassification of FFPE-derived variants if they were called by one of the two ground truth callers. In this case, a variant called in the FFPE and any one ground truth caller would be classified as a false positive while being present in two of three datasets being compared. To account for this scenario, we introduce recallest, precisionest, and F1est as defined in Equations 1–3, where Mutect2 and Strelka2 are FF variant sets and ∪ , ∩ represent an intersection and union of the two sets involved” (2nd para, col 2, pg. 3), which teaches acquiring variant candidates and using union of detection modules.
Regarding Claim 4, Dodani teaches claim 1. Dodani further provides “feature extraction is performed using a modified version of DeepSVR (Ainscough et al., 2018) and bam-readcounts3 that require a tumor BAM and a VCF file (either tumor-only or tumor-normal paired) as input. A total of 31 features (Supplementary Table S4) are extracted from the tumor bam are used to classify variants in FFPolish and can be divided into three categories:
• Summary metrics (e.g., tumor variant allele fraction (VAF), tumor depth)”
(3rd para, col 1, pg. 3), which teaches generating annotation information wherein a portion of mapped positions of each of the plurality of reads overlaps with a position of the reference variant candidate. Because the tumor variant allele fraction (VAF) equal to the plurality of reads overlaps with a position of the reference variant candidate.
Regarding Claim 5, Dodani teaches claim 4. Dodani further provides “Metrics including error rate, coverage, insert size, mapping quality, percentage of reads with insertions or deletions and GC bias were estimated and extracted using Qualimap (2.2.1) (García-Alcalde et al., 2012) and Picard (2.4.1)1. IGV (Robinson et al., 2011) was used for manual read and alignment inspection” (4th para, col 2, pg. 2), which suggests extract features of insert size of a threshold range.
Regarding Claim 6, Dodani teaches claim 1. Dodani further provides “feature extraction is performed using a modified version of DeepSVR (Ainscough et al., 2018) and bam-readcounts3 that require a tumor BAM and a VCF file (either tumor-only or tumor-normal paired) as input. A total of 31 features (Supplementary Table S4) are extracted from the tumor bam are used to classify variants in FFPolish and can be divided into three categories” (3rd para, col 1, pg. 3), which teaches receiving and generating information of a Panel of Normals (PON) based on sequencing data associated with a plurality of normal samples.
Regarding claim 8, Dodani further provides “feature extraction is performed using a modified version of DeepSVR (Ainscough et al., 2018) and bam-readcounts3 that require a tumor BAM and a VCF file (either tumor-only or tumor-normal paired) as input. A total of 31 features (Supplementary Table S4) are extracted from the tumor bam are used to classify variants in FFPolish and can be divided into three categories:
• Summary metrics (e.g., tumor variant allele fraction (VAF), tumor depth)
• Read-count metrics (e.g., the number of reads on the negative strand) for both the variant and reference allele
• Read-averaging metrics (e.g., the average base quality of reads) for both the variant and reference allele”
(3rd para, col 1, pg. 3), which teaches generating annotation information associated with variant type and sequence context information.
Regarding Claim 9, Dodani provides “feature extraction is performed using a modified version of DeepSVR (Ainscough et al., 2018) and bam-readcounts3 that require a tumor BAM and a VCF file (either tumor-only or tumor-normal paired) as input. A total of 31 features (Supplementary Table S4) are extracted from the tumor bam are used to classify variants in FFPolish and can be divided into three categories:
• Summary metrics (e.g., tumor variant allele fraction (VAF), tumor depth)
• Read-count metrics (e.g., the number of reads on the negative strand) for both the variant and reference allele
• Read-averaging metrics (e.g., the average base quality of reads) for both the variant and reference allele”
(3rd para, col 1, pg. 3), which teaches generating labeling and classification information (here “feature”) to the training data.
Regarding Claim 10, Dodani provides “FFPolish is an FFPE variant filtering approach based on a logistic regression model from scikit-learn (Pedregosa, 2011). The
model was trained on somatic Strelka2 FFPE calls from both HTMCP cervical and BLGSP cohorts, with a total of 8,698,388 training SNVs. As above, variants were evaluated by comparing FFPE variant calls and the ground truth variants (FF variants
from Strelka2 and Mutect2). We trained FFPolish with the most sensitive (Strelka2) and precise (Lofreq) variant caller (See Results and Section 6 of the Supplementary Material); however, we saw a decrease in the median F1est as well as a decrease in flexibility in the LoFreq model compared to using Strelka2 calls (Supplementary Tables S2, S3)” (2nd para, col 1, pg. 4), which teaches using FFPE as reference samples and FF samples for true positive (ground truth) labeling.
Regarding Claim 11, Dodani provides Fig. 1 (page 3), which teaches acquiring false positive variant labels based on reference variant candidate not present in control or FF samples.
Regarding Claim 12, Dodani provides Fig. 2 (page 5), which teaches extracting features associated with reference variants using DeepSVR, and True/False classification labels for FFPE variants.
Regarding Claim 13, Dodani provides Fig. 2 (page 5), which teaches:
Inputting reference variants to the machine learning model;
Determining a classification result as true or false positive variant;
Adjusting a hyperparameter optimization of logistic regression using grid search and 10-fold cross-validation.
Regarding Claim 14, Dodani provides “FFPolish is an FFPE variant filtering approach based on a logistic regression model from scikit-learn (Pedregosa, 2011). The
model was trained on somatic Strelka2 FFPE calls from both HTMCP cervical and BLGSP cohorts, with a total of 8,698,388 training SNVs. As above, variants were evaluated by comparing FFPE variant calls and the ground truth variants (FF variants
from Strelka2 and Mutect2). We trained FFPolish with the most sensitive (Strelka2) and precise (Lofreq) variant caller (See Results and Section 6 of the Supplementary Material); however, we saw a decrease in the median F1est as well as a decrease in flexibility in the LoFreq model compared to using Strelka2 calls (Supplementary Tables S2, S3)” (2nd para, col 1, pg. 4), which teaches inputting, to the trained machine learning model: a target variant candidate detected based on: normal target sequencing data from a normal target sample of a target individual and abnormal target sequencing data from an abnormal target sample of the target individual; and a feature of the target variant candidate. Because both the HTMCP and the BLGSP datasets have paired normal-abnormal samples from cancer patients and the abnormal samples have FF and FFPE samples (2nd – 3rd paras, col 2, pg. 2).
Dodani provides Figure 3 (pg. 6), which teaches outputting the classification results indicating whether the target variant candidate is a true positive variant.
Regarding Claim 15, Dodani teaches claim 14. Dodani teaches the target variant candidate is detected based on a comparison, via the Strelka2 and Mutect2 variant detection modules, of the normal target sequencing data and the abnormal target sequencing data, as discussed above regarding claim 14.
Regarding Claim 16, Dodani teaches claim 14. Dodani teaches the abnormal target sample is a Formalin-Fixed, Paraffin-Embedded (FFPE) sample, as discussed above regarding claim 14.
Regarding Claim 17, Dodani provides “FFPolish is an FFPE variant filtering approach based on a logistic regression model from scikit-learn (Pedregosa, 2011). The
model was trained on somatic Strelka2 FFPE calls from both HTMCP cervical and BLGSP cohorts, with a total of 8,698,388 training SNVs. As above, variants were evaluated by comparing FFPE variant calls and the ground truth variants (FF variants
from Strelka2 and Mutect2). We trained FFPolish with the most sensitive (Strelka2) and precise (Lofreq) variant caller (See Results and Section 6 of the Supplementary Material); however, we saw a decrease in the median F1est as well as a decrease in flexibility in the LoFreq model compared to using Strelka2 calls (Supplementary Tables S2, S3)” (2nd para, col 1, pg. 4), which teaches:
receiving information indicating a target variant candidate in a target abnormal sample;
determining, using the trained machine learning model, target variant candidates. detected based on: normal target sequencing data from a normal target sample of a target individual and abnormal target sequencing data from an abnormal target sample of the target individual;
wherein: the machine learning model is trained on the HTMCP and the BLGSP datasets that have paired normal-abnormal samples from cancer patients and the abnormal samples have FF and FFPE samples (2nd – 3rd paras, col 2, pg. 2).
Dodani provides supplementary Tables S1-S16, which teaches genomic profiling on target samples with determined classification results.
Dodani provides Figure 3 (pg. 6), which teaches outputting the classification results indicating whether the target variant candidate is a true positive variant.
Dodani teaches the target variant candidate is detected based on a comparison, via the Strelka2 and Mutect2 variant detection modules, of the normal target sequencing data and the abnormal target sequencing data, as discussed above regarding claim 1.
Dodani teaches the abnormal target sample is a Formalin-Fixed, Paraffin-Embedded (FFPE) sample, as discussed above regarding claim 1.
Dodani does not teach using the identified variant to prognosis or to come up with a precise treatment plan.
Chan provides (page 10, col 1, last para) “in metastatic colorectal patients treated with FOLFIRI regimen, patients with the ABCG2 rs7699188-GG genotype show decreased tumor response [28]. In a separate study, metastatic colorectal patients carrying AREG rs11942466 C[A and rs9996584 C>T were associated with OS while those carrying EGFR rs712829 G>T were associated with progression-free survival
(PFS) and OS [29]”, and (page 12, col 1 last para through col 2 1st para) “Wang and colleagues performed targeted sequencing to evaluate the link between EGFR
and EGFR-linked pathway gene SNPs with EGFR-TKI response and ADRs in patients with advanced NSCLC [55]. They identified rs1042640 in UGT1A10, rs1060463 and
rs1064796 in CYP4F11 and rs2074900 in CYP4F2 as being associated with erlotinib
treatment response, with improved an median PFS of 12.57 months compared to that of nonresponders (median PFS 3.55 months) [55]. SNP rs1064796 in CYP4F11 and SNP rs10045685 in UGT3A1 were also linked to ADRs, with carriers of CYP4F11-rs1064796-C and UGT3A1-G showing an increased risk for skin rash or digestive track injury [55]”, which teaches cancer variant can be used to predict response to treatment, and how the patient will response to the drug treatment.
Chan provides (page 11, col 1, 2nd para) “one specific variant is rs116855232
(c.415C>T) in NUDT15; this variant shows a distinct population distribution, with a particularly higher occurrence of the rare allele rs116855232-T in East Asians (10%) compared to Hispanics, Europeans and Africans (http://www.internationalgenome. org/1000-genomesbrowsers). A GWAS study on children with ALL reported an association of rs116855232 in NUDT15 (P = 8.8 9 10-9) to mercaptopurine sensitivity in only East Asian patients, with a lower tolerance to mercaptopurine resulting in hematologic toxicities [40]. Patients carrying the rs116855232-TT genotype were less tolerant, with an average dose intensity of 8.3%, compared with those with TC and CC genotypes, who tolerated 63 and 83.5% of the planned dose of 75 mg/m2 per day [40]”, which suggests to come up with a precise treatment plan with drug dosage based on the identified cancer variants.
Regarding claim 18, Dodani provides “As FFPE samples remain central to clinical diagnostics, methods for confidently calling variants in genomic data derived from such samples are required for enhanced utility in clinical and translational research settings” (last para, col 2, pg. 6 through 1st para line 1, col 1, pg. 7), which suggests the genomic profiling might be helpful in disease diagnosis.
Regarding claim 19, Dodani provides “After initial testing, we eliminated Pisces, outLyzer, and Octopus from further analysis due to computational requirements beyond what we allocated to this project (Supplementary Table S1)” (2nd para lines 1-3, col 1, pg. 3), and “For each patient, variant calls from five tools (LoFreq, Mutect2, Strelka2, Virmid, Shimmer) were each evaluated in isolation and then combined to test for improved FFPE variant calling recallest, precisionest, and F1est. To make the test computationally feasible, ten data sets, each containing results from 3 tools, were used for the analysis” (last para, col 2, pg. 3 through 1st para, col 1, pg. 4), which suggests the computational implementation of the method in claim 1.
It would have been prima facie obvious to combine Dodani’s pipeline of identifying credible cancer variant from FFPE samples with Chan’s method of correlating the variant identified with cancer treatment response as well as the appropriate treatment drug quantity, because there is sufficient evidence to
match cancer patients to therapies by utilizing the information of clinical-relevant alterations. Variants that act as predictive biomarkers for drug-induced toxicity and drug response as well as somatic mutations in cancer cells that function as drug targets are important. (Chan: page 1, Section “ABSTRACT”).
One would reasonably expect success as Dodani’s pipeline identifying cancer variant from FFPE is a common and credible way to provide cancer variant information, and the incorporation of Chan’s genetic information could improve treatment precision, leading to a better quality of life for cancer patients (Chan: page 23, col 1, 1st para).
Response to Applicant’s argument
In the Remarks filed 2 October 2025, Applicant argued (page 15, last two paras) that Dodani alone does not teach the amended claims. The argument is unpersuasive. As set forth above, the combination of Dodani and Chan teach the amended claims.
Conclusion
No claims are allowed.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to GUOZHEN LIU whose telephone number is (571)272-0224. The examiner can normally be reached Monday-Friday 8-5.
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.
/GL/
Patent Examiner
Art Unit 1686
/Anna Skibinsky/
Primary Examiner, AU 1635