DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claim 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.
Claims 1-6, 10-13 and 16-17 are rejected under 35 U.S.C. 103 as being unpatentable over Tulsyan et al., in an article entitled, “A machine-learning approach to calibrate generic Raman models for real-time monitoring of cell culture processes”, Published in Biotechnology and Bioengineering, 2019, as reproduced in Wiley Periodicals 2019 (Pages 2575 – 2586) (hereinafter: Tulsyan et al.), in view of XU et al., U.S. Patent Application Publication No. US 2020/0372225 A1 (hereinafter: ‘225).
As per claim 1, Tulsyan et al. in view of ‘225 teaches or suggests a computer-implemented method for monitoring and/or controlling a biopharmaceutical process, the method comprising:
querying, by one or more processors and based on a first spectral scan vector of the biopharmaceutical process obtained by a spectroscopy system, an observation database comprising a plurality of observation data sets associated with past scans of biopharmaceutical processes, wherein each of the observation data sets includes spectral data and a corresponding actual analytical measurement (e.g., See Tulsyan et al., which discloses these features by disclosing MATLAB software being used to run a current Raman scan through a stored library of Raman scans paired with real analytical measurements; See page 2578, Section 2.2, disclosing the use of “Five RXN2 Raman spectrometers”; Also see page 2579, Sections 2.4 – 2.5, disclosing “temporally match analytical measurements to Raman spectra” and a “library containing 3,800 Raman and analytical measurements”; also see page 2579, Section 2.5, disclosing that “Once a new Raman spectrum was collected, it was run through the library” and that the JITL algorithm was deployed in MATLAB using a proprietary script);
calibrating, by the one or more processors and using the selected training data, a local model specific to the biopharmaceutical process, the local model being trained to predict analytical measurements based on spectral data inputs (e.g., See Tulsyan et al., which discloses using the selected training data to train a local model to predict analytical measurements from Raman scan data; See Tulsyan et al.; page 2579; Section 2.5, which discloses that the “top 100 entries constituted a local calibration set” and “equipped with PLS and GPs for local model calibration.”; also see page 2577; section 1, which states “calibrates local models in real-time upon query” and “using only the relevant data from the library.”);
predicting, by the one or more processors, an analytical measurement of the biopharmaceutical process, wherein predicting the analytical measurement of the biopharmaceutical process includes using the local model to analyze spectral data that the spectroscopy system generated when scanning the biopharmaceutical process (e.g., See Tulsyan et al., which discloses using the local model to analyze the current Raman scan and predict the current analytical measurement; See Tulsyan et al., page 2579, Section 2.5, which discloses “After predicting the critical parameters for the presented query point.”; also see page 2580, Section 3.1, which discloses “JITL-PLS and JITL-GP models were deployed in MATLAB for real-time predictions of critical parameters.”).
However, Tulsyan et al. does not adequately disclose that querying the observation database includes:
determining first parameters defining a set of distributions for the first spectral scan vector, or
selecting as training data, from among the plurality of observation data sets, particular observation data sets based on (i) the first parameters and (ii) other parameters defining respective sets of distributions for the plurality of observation data sets.
Although Tulsyan selects training data by comparing the current Raman scan to stored Raman scans using Euclidian distance (e.g., See Tulsyan et al., page 2579, Section 2.5, disclosing “run through the library”, “selected based on the Euclidean distance,” and “top 100 entries constituted a local calibration set.”), Tulsyan does not disclose selecting training data based on distribution parameters for the current Raman scan and the stored Raman scans.
‘225 discloses using a VAE encoder to make distribution information for an input. Specifically, ‘225 discloses that an input may be processed by a VAE encoder to obtain a mean vector and a covariance matrix, which together define a Gaussian distribution for the input, and ‘225 also discloses doing this for the training of inputs (e.g., See ‘225; [0050], [0052] and [0080]).
It would have been obvious to one of ordinary skill in the art at the time the invention was made to incorporate the teachings of ‘225 into the method of Tulsyan et al. for the purpose of using distribution information to better identify stored Raman scans that are similar to the current scan, thereby improving the selection of the stored Raman scans used to train the local model.
As per claim 2, Tulsyan et al. in view of ‘225 further discloses that determining the first parameters includes processing the query sample using an encoder of a variational autoencoder, and wherein the encoder outputs the first parameters (e.g., See ‘225; [0046] and [0050], which discloses using an encoder to process an input and using a VAE to output distribution information for the input, including a mean vector and a covariance matrix).
As per claim 3, Tulsyan et al. in view of ‘225 further discloses that the encoder includes exactly one hidden layer (e.g., See ‘225; [0094], which discloses a single-layer bi-directional LSTM encoder).
As per claim 4, Tulsyan et al. in view of ‘225 further discloses determining, by the one or more processors, the other parameters using the encoder of the variational autoencoder, wherein the encoder outputs the other parameters (e.g., See ‘225; [0052], which discloses processing a plurality of training inputs with the encoder and outputting distribution information for the plurality of training inputs, including a mean vector and a covariance matrix, wherein when incorporated into Tulsyan et al., the plurality of training inputs correspond to the stored Raman scans).
As per claim 5, Tulsyan et al. in view of ‘225 further discloses selecting the particular observation data sets includes calculating multivariate KL divergence metrics based on the first parameters and the other parameters (e.g., See ‘225; [0052] and [0086], which disclose determining a KL divergence value using distribution information, wherein when incorporated into Tulsyan et al., the KL divergence value is used to compare the distribution information for the current Raman scan with the distribution information for the stored Raman scans).
As per claim 6, Tulsyan et al. in view of ‘225 further discloses that calibrating the local model specific to the biopharmaceutical process includes calibrating a Gaussian process machine learning model specific to the biopharmaceutical process (e.g., Tulsyan et al. discloses calibrating a Gaussian process local model using selected training data; See page 2579; section 2.5, which discloses “equipped with PLS and GPs for local model calibration”).
As per claim 10, Tulsyan et al. in view of ‘225 further discloses that using the local model to analyze the spectral data includes using the local model to analyze the first spectral scan vector (e.g., See Tulsyan et al. disclosing using the current Raman scan as the scan analyzed by the local model; See page 2579; section 2.5, which discloses “Once a new Raman spectrum was collected” and “After predicting the critical parameters for the presented query point.”).
As per claim 11, Tulsyan et al. in view of ‘225 further discloses that the predicted analytical measurement of the biopharmaceutical process is a metabolite concentration (e.g., See Tulsyan et al., which discloses predicting a current analytical measurement, including a metabolite level, such as a glucose level or a lactate level; See page 2575, Abstract, disclosing “glucose, glutamate, glutamine, ammonium, lactate”).
As per claim 12, Tulsyan et al. in view of ‘225 further discloses that the predicted analytical measurement of the biopharmaceutical process is osmolality, viability, viable cell density, or titer (e.g., See Tulsyan et al., which discloses predicting a current analytical measurement, including viability or viable cell density; see page 2575; Abstract, which discloses “viability, and viable cell density.”).
As per claim 13, Tulsyan et al. in view of ‘225 further discloses that the spectroscopy system is a Raman spectroscopy system (e.g., See Tulsyan et al., which discloses using a Raman spectroscopy system to obtain a current Raman scan from the bioprocess; see page 2578, section 2.2, which discloses “Five RXN2 Raman spectrometers” and “spectra were acquired in situ using a stainless steel optical probe”).
As per claim 16, the rationale as set forth with respect to the rejections of claims 1 and 13, from above, are applied herein. Further, Tulsyan et al. discloses a spectroscopy system including Raman spectrometers and a stainless steel optical probe that uses a 785 nm laser source and acquires spectra in situ from the bioprocess (e.g., See Tulsyan et al.; page 2578, section 2.2., which discloses “Five RXN2 Raman spectrometers” and “spectra were acquired in situ using a stainless steel optical probe”; also see page 2579; section 2.5, disclosing that “The JITL was deployed in MATLAB with the library.”).
As per claim 17, the rationale as set forth with respect to the rejection of claim 1, from above, is applied herein. It is noted that claim 17 is directed to a non-transitory computer readable medium storing instructions that, when executed by one or more processors, cause the one or more processors to implement the features of claim 1, and Tulsyan et al. in view of ‘225 teaches or suggests these features (e.g., See ‘225; [0143]).
Claims 7-9 are rejected under 35 U.S.C. 103 as being unpatentable over Tulsyan, in view of XU et al., as applied to claim 1, from above, in further view of Lang et al., U.S. Patent Application Publication No. 2013/0295596 A1 (hereinafter: ‘596).
As per claim 7, Tulsyan et al. in view of ‘225 does not specifically disclose that querying the observation database includes downsampling the first spectral scan vector; and using the local model to analyze the spectral data includes downsampling the spectral data.
‘596 discloses these features (e.g., See ‘596; [0074] and [0160], which disclose downsampling spectral profiles to reduce the number of data points).
It would have been obvious to one of ordinary skill in the art at the time the invention was made to incorporate the teachings of ‘596 into Tulsyan et al. in view of ‘225 for the purpose of downsampling the current Raman scan and the spectral data analyzed by the local model so that the database matching and model analysis run faster using less computer processing and memory.
As per claim 8, Tulsyan et al., in view of ‘225, in further view of ‘596 further discloses that querying the observation database includes baseline-correcting the downsampled first spectral scan vector; and using the local model to analyze the spectral data includes baseline-correcting the downsampled second spectral data (e.g., See ‘596; [0162] and [0163], which disclose that after downsampling, the unwanted baseline is removed so that the comparisons are better).
As per claim 9, Tulsyan et al., in view of ‘225, in further view of ‘596 further discloses that querying the observation database includes normalizing the downsampled and baseline-corrected first spectral scan vector; and using the local model to analyze the spectral data includes normalizing the downsampled and baseline-corrected spectral data (e.g., See ‘596; [0164] and [0165], which disclose that after baseline removal, the scans are normalized to account for variation in signal amplitude).
Claims 14 and 15 are rejected under 35 U.S.C. 103 as being unpatentable over Tulsyan, in view of XU et al., as applied to claim 1, from above, in further view of Czeterko et al., U.S. Patent Application Publication No. 2019/0112569 A1 (hereinafter: ‘569).
As per claim 14, Tulsyan et al. in view of ‘225 does not explicitly disclose controlling, by the one or more processors and based on the predicted analytical measurement of the biopharmaceutical process, at least one parameter of the biopharmaceutical process.
Further, as per claim 15, Tulsyan et al. in view of ‘225 does not explicitly disclose causing, by the one or more processors, a user interface to display the predicted analytical measurement.
‘569 discloses these missing features by disclosing using the predicted analytical measurement to control the bioprocess and providing a user interface for displaying process information to a user (e.g., See ‘569; [0045], [0046], [0050] and [0052], which disclose using the predicted analytical measurement to control the bioprocess and display process information to the user).
It would have been obvious to one of ordinary skill in the art at the time the invention was made to incorporate the teachings of ‘569 into ‘474 in view of ‘225 for the purpose of making the predicted analytical measurement more useful in real time by allowing the bioprocess to be automatically adjusted toward target conditions, thereby improving process stability, product quality and manufacturing consistency.
References Considered but Not Relied Upon
The following references have been considered but were not relied upon with respect to any prior art rejections:
(1) US 2019/0112569 A1, which discloses using in-situ Raman plus chemometric models to measure nutrients in real time and automatically adjust feeds to keep cell culture conditions on target;
(2) US 2019/0137338 A1, which discloses building a broad Raman calibration models for cell culture, predicting glucose, lactate, and cell density so manufacturing runs can be monitored and controlled;
(3) US 10,563,163 B2, which discloses creating Raman multivariate models that work across reactor scales, letting small scale data predict glucose and other culture parameters in large bioreactors;
(4) US 2022/0018782 A1, which discloses using Raman spectra plus lab test results to train models that estimate viral titer during bioreactor runs, with optional baseline correction and normalization; and
(5) US 2015/0247210 A1, which discloses using near infrared spectroscopy to continuously measure chemicals and cells in bioprocess fluids in real time, enabling automated monitoring and feedback control.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to RONALD D HARTMAN JR whose telephone number is (571)272-3684. The examiner can normally be reached M-F 8:30 - 4:30 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, Mohammad Ali can be reached at (571) 272-4105. 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.
/RONALD D HARTMAN JR/Primary Patent Examiner, Art Unit 2119 June 22, 2026
/RDH/
Prosecution Insights