DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
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and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114.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 May 11, 2026 has been entered.
Response to Arguments
This action is responsive to the Request for Continued Examination and amendment filed May 11, 2026. Claims 1-20 are pending. Claims 1, 19, and 20 are amended.
Applicant's Argument:
Applicant argues that amended claim 1 requires generating two separate variable-to-trait models (a first model using observed values and a second model using substitute values) and comparing a coefficient of the variable of interest from each model, whereas Candes uses a single algorithm acting on both observed and substitute values together. Applicant concludes that neither Lightner nor Candes teaches the amended limitations.
Examiner's Response:
The argument is not persuasive. As an initial matter, the rejection is now made over Lightner in view of Candes and further in view of Watson, and Applicant's argument is directed only to Candes. The two-model comparison Applicant relies upon to distinguish the claims is expressly taught by Watson. Watson evaluates a fitted supervised model on the observed data and again on a dataset in which the variable of interest is replaced by its substitute (knockoff) values, and compares the two to determine whether the variable is causal (Watson: "replacing the submatrix with the corresponding knockoff variables, X_S, rendering a new dataset" (§3); "we test whether the model performs better using the original or the knockoff data" (§3); the measure "may also be applied in causal discovery" (Abstract).). That is precisely the claimed first (observed-value) and second (substitute-value) variable-to-trait model and the comparison of their influence metrics. Applicant's single-versus-two-model distinction over Candes therefore does not distinguish over the combination.
Further, the amended limitations were known techniques as of the effective filing date, as Applicant's own specification admits. Specification ¶ 88 states that the influence metric "can be a measure of conditional model reliance," and ¶ 94 describes the metric as "a comparison ... between a coefficient/or the observed variable ... and a coefficient/or the corresponding test variable." Conditional model reliance is the established method of gauging a variable by comparing a model's reliance on the variable's observed values against its reliance on altered values, which is the very two-model comparison now claimed. This is applicant-admitted prior art.
Lightner supplies the first, observed-value model and its coefficient (Lightner: "a regression coefficient as found by PLS-R" (¶¶272-273)), Candes supplies both the substitute (knockoff)values and a coefficient-level comparison of the original against the substitute variable (Candes: W j = Z_j - Z_j (eq. 3.6)), and Watson ties these into the observed-versus-substitute model comparison and operates on any supervised learner, including Lightner's regression (Watson: "works in conjunction with any valid knockoff sampler, supervised learning algorithm, and loss function" (Abstract)). The combination therefore teaches every amended limitation. To the extent Applicant contends the two models must be trained by two separate fitting procedures, that is at most an obvious implementation choice. Watson expressly relates its procedure to leave-one-covariate-out refitting (Watson: "With knockoffs, we can directly import LOCO's statistical guarantees" (§2.2)), so a person of ordinary skill would recognize the substitute-value comparison and a separate refit as art-recognized equivalents yielding the same result. Such a predictable variation is obvious. KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007). The rejection is maintained.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Lightner, et al. (US 2023/0367272 A1 – hereinafter “Lightner”) in view of Candes, et al. (Panning for gold: ‘model-X’ knockoffs for high dimensional controlled variable selection – hereinafter “Candes”) in view of Watson & Wright (Testing Conditional Independence in Supervised Learning Algorithms, Machine Learning – hereinafter "Watson").
Claims 1, 19, and 20.
Lightner discloses a method for generating organisms having a target trait (Lightner ¶79: “desired trait.” Lightner further teaches: "Progeny are selfed and selected so that the newly developed inbred has many of the attributes of the recurrent parent and yet several of the desired attributes of the non-recurrent parent “ (¶301).), comprising:
(Claim 19 preamble) A system for generating a new organism having a target trait, comprising :one or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including instructions for (Lightner ¶50: “(a) a computer
with a memory … software”) …
(Claim 20 preamble) A non-transitory computer-readable storage medium storing one or more programs, the one or more programs comprising instructions, which when executed by one or more processors of an electronic device cause the electronic device to (Lightner ¶50: “(a) a computer with a memory … software”) …
receiving one or more spectrograms corresponding to each organism of a set of organisms (Lightner ¶280: “Hyper spectral data was collected for each of the samples by remote sensing.” Lightner discloses: "Multi- or hyper-spectral data was collected for the plots by remote sensing imaging from which X-block calibration data can be extracted" (¶344). Lightner further teaches: "Spectra from the plant canopies were collected in digitized form and stored" (¶281 ).);
generating, from the one or more spectrograms, a plurality of spectrogram attributes characterizing wavelengths for each organism of the set of organisms (Lightner discloses: " An example of remote sensing data is a remote sensed image of the plant or plants, which can be converted by well-known methods into a spectrogram. The spectrogram can be sampled at a number of wavelengths to provide a matrix of values correlated to wavelength" (¶165). Lightner further teaches: “data suitably include spectra of a plurality of wavelengths from one of more of visible (VIS), infrared (IR), near infrared (NIR), or ultraviolet (UV) light … The spectra may be reflectance or absorbance; absorbance spectra are considered especially suitable. In some examples, the spectra are hyper spectral data” (¶135).);
reducing a dimensionality of the plurality of spectrogram attributes to obtain a set of spectral variables, wherein a number of spectral variables in the set of spectral variables is smaller than a number of spectrogram attributes in the plurality of spectrogram attributes (Lightner discloses: "A PCA method can be used to reduce the dimensionality of a large number of interrelated variables (absorption intensities at different wavelengths) while retaining information that distinguishes one component from another … The new set of variables is ordered such that the first few retain most of the variation present in the original set” (¶181).);
selecting a spectral variable of interest from the set of spectral variables (Lightner discloses: "Predictions can also be made from a regression of the scores against the traits of interest. Calibration models based on scores is called a principal component regression model." (¶186));
receiving observed trait values for the target trait from the set of organisms (Lightner discloses: "analyzing the spectroscopic data to quantitatively measure one or more characteristics (e.g., phenotypic variables) of the one or more progeny plants " (¶133); list of desired traits
including disease resistance, herbicide resistance, and yield enhancements (¶306).);
generating a first variable-to-trait model using observed variable values for the spectral variable
of interest and the observed trait values for the target trait (Lightner: "Predictions can also be
made from a regression of the scores against the traits of interest ... a principal component
regression model" (¶ 186); regression analysis was performed to determine the relationship
between the principal components and the phenotypic traits (¶¶ 287-290). This is a model
relating spectral variables to a trait, trained on observed variable values.);
determining a first influence metric for the spectral variable of interest based on a first
coefficient for the spectral variable of interest in the first variable-to-trait model (Lightner further teaches: "The regression equation for chlorophyll consists of a set of terms comprising a regression coefficient as found by PLS-R, and corresponding absorbance value at each of the spectral points … chlorophyll content may be indicative of other things, such as phenotype, health, vitality, yield, etc." (¶¶272-273). Lightner: “Results show very strong correlation between the actual and predicted as shown by the regression coefficients” (¶273) shows correlation coefficients were calculated between the principal component and the disease resistance trait (¶306).);
based on the determination, generating a new organism with the target trait (Lightner discloses: "Backcrossing can be used to transfer a specific desirable trait from one
line… for the desired trait to be transferred from the non-recurrent parent" (¶299) and "Typically after about four or more backcross generations with selection for the desired trait, the progeny will contain essentially all genes of the recurrent parent except for the genes controlling the desired trait " (¶300). Lightner further teaches: "Progeny are selfed and selected so that the newly developed inbred has many of the attributes of the recurrent parent and yet several of the desired attributes of the non-recurrent parent “ (¶301).).
Lightner discloses all of the subject matter as described above except for specifically teaching “determining a variable-to-variable model configured to predict values for the spectral variable of interest based on one or more other spectral variables of the set of spectral variables; determining substitute variable values for the spectral variable of interest using the variable-to-variable model,” “generating a second variable to trait model using the substitute variable values for the spectral variable of interest and the observed trait values for the target trait,”
and “determining a second influence metric for the spectral variable of interest based on a second coefficient for the spectral variable of interest in the second variable-to-trait model; determining that the spectral variable of interest is a causal variable by comparing between the first influence metric and the second influence metric.”
However, Candes and Watson in the same field of endeavor teach these limitations. Candes teaches determining a variable-to-variable model configured to predict values for the spectral variable of interest based on one or more other spectral variables of the set of spectral variables (Candes “
s
a
m
p
l
e
X
~
j
f
r
o
m
L
(
X
j
|
X
-
j
,
X
~
1
:
j
-
1
)
” In the Model-X procedure, the construction of the substitute value
X
~
j
requires knowledge of the conditional distribution (the V2V model),
L
(
X
j
|
X
-
j
,
X
~
1
:
j
-
1
)
. This distribution predicts the spatial variable of interest
(
X
j
)
based on one or more other spectral variables (
X
-
j
).);
determining substitute variable values for the spectral variable of interest using the variable-to-variable model (Candes: “Correct inference in such a broad setting is achieved by constructing knockoff variables probabilistically instead of geometrically” (Summary, p. 551). Candes: The knockoffs
X
~
1
…
X
~
p
sampled from
L
(
X
j
|
X
-
j
,
X
~
1
:
j
-
1
)
, is the substitute value for the spectral variable of interest, and the pair (Xj, ˜Xj) is exchangeable conditionally on all the other variables (§3.4.1, p. 563).)
generating a second variable to trait model using the substitute variable values for the spectral variable of interest and the observed trait values for the target trait (Watson: “replacing the submatrix with the corresponding knockoff variables,
X
~
S
, rendering a new dataset” (§ 3); the procedure “works in conjunction with any valid knockoff sampler, supervised learning algorithm, and loss function” (Abstract), including a regression of the traits as in Lightner)
and
determining a second influence metric for the spectral variable of interest based on a second coefficient for the spectral variable of interest in the second variable-to-trait model (Candes:
Z
~
j
is the fitted coefficient of the substitute/knockoff variable (eq. 3.6); Watson: the conditional predictive impact “
∆
i
=
L
f
,
z
¯
i
-
L
(
f
,
z
i
)
)” is “the difference in sample-wise loss between predictions made using knockoff data and original data” (§ 3.1, eqs. 1-2). Candes teaches the second influence metric as the coefficient of the substitute (knockoff) variable, and Watson teaches the corresponding metric obtained from the substitute-value model.).
determining that the spectral variable of interest is a causal variable by comparing between the first influence metric and the second influence metric (Candes: the knockoff statistic W=Zj –
Z
~
j
compares the coefficient of the original variable (Z, for the original variable X) against the substitute variable (
Z
~
, for the substitute variable
X
~
)
and is thresholded to select causal varibles (eq. 3.6); Watson: the test “
H
0
:
R
f
,
Z
≥
R
(
f
,
Z
~
)
” determines “whether the model performs better using the original or the knock-off data” (§ 3) and “may also be applied in causal discovery” (Abstract).). Watson expressly builds on Candes' knockoff framework, so the references are readily combinable (Watson: Building on the knockoff framework of Candès et al. ...(2018), ...we develop a novel testing procedure... (Abstract).).
Therefore, it would have been obvious to one of ordinary skill in the art to combine Lightner, Candes, and Watson before the effective filing date of the claimed invention. The motivation would have been to overcome the inherent risk of false correlation and overfitting caused by high collinearity of Lightner's spectral variables. Candes supplies substitute (knockoff) values that preserve the covariance structure while carrying no true signal, and Watson supplies the observed versus substitute model comparison that isolates whether a spectral variable is genuinely causal. Applying that causality to filter Lightner's spectral trait model predictably yields an accurate and stable plant trait model.
Claim 2.
The combination of Lightner and Candes discloses the method of Claim 1, wherein the set of organisms comprises organisms at a seed stage (Lightner " The estimation or prediction may be used to select for or against a plant, seed, or condition. " (¶132)).
Claim 3.
The combination of Lightner and Candes discloses the method of Claim 1, wherein the set of organisms comprises organisms at a seedling stage (Lightner " The predictions generated by the methods may be used for selection of a plant or its seed.” (¶149)).
Claim 4.
The combination of Lightner and Candes discloses the method of Claim 1, wherein the one or more spectrograms are received from a spectrometer (Lightner ¶¶192, 324 spectroscopy).
Claim 5.
The combination of Lightner and Candes discloses the method of Claim 1, wherein the one or more spectrograms are received from a camera (Lightner ¶104 “imaging spectroscopy is similar to an image produced by a digital camera”).
Claim 6.
The combination of Lightner and Candes discloses the method of Claim 5, wherein the camera is a drone camera (Lightner ¶241 “to several tens of meters above the ground on a cherry picker or crane, to one-hundred or more meters via a plane, helicopter, or even a satellite, using a hyper-spectral digital imaging system or other specta-gathering devices.”).
Claim 7.
The combination of Lightner and Candes discloses the method of Claim 1, wherein the wavelengths comprise near-infrared wavelengths (Lightner ¶242 “near infrared (NIR)”).
Claim 8.
The combination of Lightner and Candes discloses the method of Claim 1, wherein the plurality of spectrogram attributes comprise at least one of: an absorption, a reflectance, a transmission, or an emission at a wavelength (Lightner “include spectra from one or more wavelengths from the visible light spectrum, from the infrared spectrum, the near-infrared spectrum" (¶112)).
Claim 9.
The combination of Lightner and Candes discloses the method of Claim 1, wherein reducing the dimensionality comprises generating one or more wavelets from the plurality of spectrogram attributes (Lightner "Martens and Naes 1989 describes various pretreatments of data useful in model building. The identified software allows a variety of pretreatments ... mathematical treatments such as smoothing or derivatives are available in the software packages for pretreatment of the data" (¶215). Wavelet transformation is a known mathematical pretreatment technique for spectral data.).
Claim 10.
The combination of Lightner and Candes discloses the method of Claim 1, wherein reducing the dimensionality comprises removing one or more principal components from the plurality of spectrogram attributes (Lightner "PCA method can be used to reduce the dimensionality of a large number of interrelated variables ( absorption intensities at different wavelengths)" (¶181). Lightner further teaches: "PCA replaces the many variables (the individual wavelengths of multi-spectral or hyper spectral spectrum range) with new variables" (¶184)).
Claim 11.
The combination of Lightner and Candes discloses the method of Claim 1, wherein reducing the dimensionality comprises fitting a regression to one or more spectrogram attributes (Lightner teaches regression to one or more spectrogram attributes as a dimensionality reduction technique (¶¶287-290).).
Claim 12.
The combination of Lightner and Candes discloses the method of Claim 1, wherein reducing the dimensionality comprises determining a summary statistic of one or more spectrogram attributes (Lightner discloses: "These reflectance readings at 1.2 nm increments collected over a visible and near infrared wavelength range for each image of each sample canopy can be each stored. Optionally, the reflectance readings for each wavelength for all sample images can be averaged into one hyper-spectral plot" (¶243). Averaging is a summary statistic.).
Claim 13.
The combination of Lightner and Candes discloses the method of Claim 1, wherein determining the variable-to-variable model comprises fitting a regression to one or more spectrogram attributes (Candes teaches the conditional distribution (the V2V model),
L
(
X
j
|
X
-
j
,
X
~
1
:
j
-
1
)
. (p. 563). Candes’s summary on p. 1 discloses “although we provide preliminary experimental evidence that our procedure is robust to unknown or estimated distributions.” Estimating a conditional distribution in statistics often involves a regression model. Lightner discloses regression analysis was performed to determine the relationship between the principal components and the phenotypic traits (¶287) ).
Claim 14.
The combination of Lightner and Candes discloses the method of Claim 1, wherein the variable-to-variable model is configured to generate a distribution of substitute variable values based on the one or more other spectral variables (Candes “Correct inference in such a broad setting is achieved by constructing knockoff variables probabilistically instead of geometrically” (p. 551, Summary). Knockoff variables (sub values) are created via probabilistic process (sampling ), meaning they are generated from a distribution.
“
s
a
m
p
l
e
X
~
j
f
r
o
m
L
(
X
j
|
X
-
j
,
X
~
1
:
j
-
1
)
”
, where,
L
is a conditional distribution.).
Claim 15.
The combination of Lightner and Candes discloses the method of Claim 1, wherein the first influence metric and the second influence metric are regression coefficients (Lightner discloses regression analysis was performed to determine the relationship between the principal components and the phenotypic traits (¶287) and "correlation coefficients" (¶268) where, the regression coefficient quantifies the strength of association.).
Claim 16.
The combination of Lightner and Candes discloses the method of Claim 1, further comprising determining a target value for the causal spectral variable based on the target trait and generating the new organism based on the target value (Lightner discloses: Plants identified as having favorable spectral signatures associated with disease resistance were selected for breeding (¶432) and "Selection of progeny containing the trait of interest is accomplished by direct selection for a trait associated with a dominant allele" (¶307). This involves determining target spectral values associated with desired trait values.).
Claim 17.
The combination of Lightner and Candes discloses the method of Claim 1, wherein generating the new organism comprises breeding one or more organisms from the set of organisms based on the determination (Lightner teaches: "A plant can be selected for further use in a breeding program based on the predicted constituent from one or more of the models" (¶432) and "These plants may then be grown and self-pollinated, backcrossed, and/or outcrossed" (¶119)).
Claim 18.
The combination of Lightner and Candes discloses the method of Claim 1, wherein generating the new organism comprises altering one or more organisms from the set of organisms based on the determination (Lightner discloses: "The plant cells and/or tissue that have been transformed may be grown into plants" (¶119) and discusses "transgenic trait" applications (¶120). Lightner further teaches: "Such traits include – but are not limited to – insect resistance, com rootworm resistance, herbicide resistance, drought tolerance …" (¶120), indicating that organisms are altered (transformed) to generate new organisms with target traits. Additionally, Lightner discusses: “transformation" as a plant breeding technique (¶440).).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Ross Varndell whose telephone number is (571)270-1922. The examiner can normally be reached M-F, 9-5 EST.
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, O’Neal Mistry can be reached at (313)446-4912. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see https://ppair-my.uspto.gov/pair/PrivatePair. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/Ross Varndell/Primary Examiner, Art Unit 2674
Prosecution Insights