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 .
Status of the Claims
Claims 1-20 are pending and examined herein.
No claims are canceled.
Priority
As detailed on the 02 March 2023 filing receipt, the application claims priority as early as 07 February 2022 to provisional application 63/267,660. At this point in examination, all claims have been interpreted as being accorded this priority date as the effective filing date.
Information Disclosure Statement
Information disclosure statements (IDS) were filed on 03 February 2023 and 12 July 2023. The submissions are in compliance with the provisions of 37 CFR 1.97. Accordingly, the references are being considered by the examiner.
Claim Objections
Claim 5 is objected to because the list elements in the second list do not end with “ing” as the other actively recited steps, and the last item in the second list, which begins with “repeat”, should begin as the others with a letter, here “(e)”.
Claim 6 recites “molecules with high objection function value” rather than “values.”
Claim 6 recites “molecules with calculated objection function value” instead of “values”.
Claim 9 is objected to because “target” is misspelled as “targe” in the “selecting” element.
Claim 9 is also objected to because “closest” is misspelled as “closest.”
Claim 10 recite “by protocol” and should recite “by a protocol.”
Claim 10 recites “and” between the third to last and second to last list elements as well as between the second to last and last list elements, and only is required for the former.
Claim 17 recites “calculating similarity metric between molecules” and should read “calculating a similarity metric.” Additionally, claim 17 recites “selecting generated molecules closest to similarity metric” This should likely read “selecting generated molecules with the closest to similarity metric” or a similar amendment.
Appropriate correction is required.
Claim Rejections - 35 USC § 112(b)
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1-20 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim 1 recites selecting scored molecules from the database with relatively larger objective
function values over other scored molecules in the database. It is unclear how much larger the values must be over which other values. That is, it is not clear if there is a threshold or comparison to an average. Therefore, it is unclear which scored molecules are selected. Claims dependent on claim 1 are rejected on similar grounds.
Claim 2 recites selecting molecules, of the generated molecules, which are closest to the selected scored molecules. It is unclear what the metric for closeness is, and therefore it is unclear which molecules should be selected.
Claims 4 and 14 recite selecting a top number of molecules in each cluster. It is unclear what constitutes a top number of molecules per cluster as there is no threshold for what is considered a top molecule, and so it is unclear which are to be selected.
Claim 5 recites what appears to be two lists: a first list comprising determining steps and a second list contingent on when the number of points does not reach a threshold. The first list recites determining step and levels but it is unclear how this relates to sampled points as found in parent claim 1 and the second list. The third and fourth list elements should be joined by a conjunction like “and” to clarify they are all required. The second list recites when a number of sampled points is less than a threshold, but it is unclear what relationship this has to the first list.
Claim 6 recites selecting molecules with high objective functions that are diverse. It is unclear what is meant by a high value in this context, and it is unclear what diverse entails in this context, and thus unclear which molecules are to be selected.
Claims 10 and 17 also recite diverse molecules and it unclear what diverse means.
Claims 13 and 15 recite molecules with the highest score, but is unclear what cut off constitutes a high score.
Claim 16 recites “using encoder”, “using decoder”, and “calculating objective function,” and it is unclear if it is the same encoder, decoder, and objective function introduced in parent claim 1. Therefore, the terms have unclear antecedence. This rejection may be overcome by amending to recite “using the encoder”, “using the decoder”, and “calculating the objective function.”
Claim 17 recites selecting molecules closest to the similarity metric. It is unclear what metric for closeness would need to be satisfied to select the molecules.
35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-20 are rejected under 35 USC § 101 because the claimed inventions are directed to an abstract idea without significantly more. "Claims directed to nothing more than abstract ideas (such as a mathematical formula or equation), natural phenomena, and laws of nature are not eligible for patent protection" (MPEP 2106.04 § I). Abstract ideas include mathematical concepts, and procedures for evaluating, analyzing or organizing information, which are a type of mental process (MPEP 2106.04(a)(2)). The claims as a whole, considering all claim elements individually and in combination, are directed to a judicial exception at Step 2A, Prong 2, and the additional elements of the claims, considered individually and in combination, do not provide significantly more at Step 2B than the abstract idea of generating molecular structures.
MPEP 2106 organizes JE analysis into Steps 1, 2A (Prong One & Prong Two), and 2B as analyzed below.
Step 1: Are the claims directed to a process, machine, manufacture, or composition of matter (MPEP 2106.03)?
Step 2A, Prong One: Do the claims recite a judicially recognized exception, i.e., a law of
nature, a natural phenomenon, or an abstract idea (MPEP 2106.04(a-c))?
Step 2A, Prong Two: If the claims recite a judicial exception under Prong One, then is the judicial exception integrated into a practical application by an additional element (MPEP 2106.04(d))?
Step 2B: Do the claims recite a non-conventional arrangement of elements in addition to any identified judicial exception(s) (MPEP 2106.05)?
Step 1: Are the claims directed to a 101 process, machine, manufacture, or composition of matter (MPEP 2106.03)?
The claims are directed to a method (claims 1-18), a non-transitory computer-readable medium (claim 19), and a computer system (claim 20), each of which falls within one of the categories of statutory subject matter. [Step 1: Yes]
Step 2A, Prong One: Do the claims recite a judicially recognized exception, i.e., a law of nature, a natural phenomenon, or an abstract idea (MPEP 2106.04(a-c))?
With respect to Step 2A, Prong One, the claims recite judicial exceptions in the form of abstract ideas. MPEP § 2106.04(a)(2) further explains that abstract ideas are defined as:
• mathematical concepts (mathematical formulas or equations, mathematical relationships
and mathematical calculations) (MPEP 2106.04(a)(2)(I));
• certain methods of organizing human activity (fundamental economic principles or practices, managing personal behavior or relationships or interactions between people) (MPEP 2106.04(a)(2)(II)); and/or
• mental processes (concepts practically performed in the human mind, including observations, evaluations, judgments, and opinions) (MPEP 2106.04(a)(2)(III)).
Claim 1 recites selecting molecules from the database based on objective function values and selecting a point in latent space, where making a selection is a step practically performed by the human mind.
Claim 1 recites sampling points, which is interpreted as selecting points based on a mathematical distance metric, and thus is a combination of mental and mathematical steps.
Claim 1 recites providing a report, which could be interpreted as an abstract step of relaying information.
Claim 2 recites comparing molecules, selecting molecules, and providing molecules, which are all interpreted as abstract steps in the form of mental processes where the human mind is practically equipped to make comparisons, making selections, and provide information.
Claim 3 recites additional information related to making a selection.
Claims 4 and 10 recite selecting molecules, which is a mental process of data selection; calculating fingerprints, which is interpreted as coding information as a vector and thus a mathematical process; clustering, which is a mathematical process or a mental process of sorting data; selecting top molecules, which is a data selection step which can be performed by the human mind; sorting by numerical value of the objection function, which can be interpreted as a mental process or mathematical one; randomly sampling, which is considered to a data selecting step; and providing molecules from the clusters, which is relaying information and thus an abstract idea.
Claim 5 recites several steps making determinations, which include selecting points, a step length, number of levels, and number of steps, all of which are mental steps of deciding parameters. Claim 5 recites sampling points and adding them to a list, which is a data manipulation step and thus a mental process, and increasing a distance for data generation, which is a mental step. Claim 5 recites repeating steps already determined to be abstract.
Claim 6 recites selecting molecules, which is a mental step.
Claim 6 recites calculating an objective function, which is a verbal description of a mathematical function.
Claim 7 recites filtering molecules for validity and selecting molecules. Checking for validity and making selections are steps practically performed in the human mind.
Claim 8 recites determining a molecule property, which is interpreted under a broadest reasonable interpretation as interpreting the molecule and thus a mental process.
Claims 8 and 12 recite determining a similarity metric. Determining a similarity metric is interpreted as either a mathematical calculation producing a value for comparison or a mental step of determining whether molecules are alike or not. Claims 8 and 12 recite selecting molecules, where selecting is a step practically performed by the human mind.
Claim 9 recites calculating descriptors, where calculating is a mathematical step.
Claim 9 recites comparison descriptors, where comparing is a mental step.
Claim 9 recites selecting molecules, which a mental step.
Claim 11 recites calculating molecular descriptors, which is a mathematical concept including counting hydrogen bond acceptors or donors.
Claim 13 recites selecting models, which is a mental process; calculating a fingerprint, interpreted as coding information as a vector and thus a mathematical process; applying a clustering method, where a clustering method may be a mathematical or mental process of data interpretation and evaluation; data selection for clusters, which is a mental step, and randomly selecting a molecular per cluster, where selecting is a mental step.
Claim 14 recites mental steps in the forms of selecting molecules, sorting molecules, and clustering molecules, as well as mathematical steps in the form of calculating fingerprints and clustering.
Claim 15 recites generating molecules, which is a mental process.
Claim 15 recites providing a base of molecules, which interpreted as data and thus abstract.
Claim 15 recites selecting and identifying steps, where are practically performed by the human mind.
Claim 16 recites selecting molecules, creating points, filtering molecules, and adding them, which are all mental steps, whereas calculating an objective function is a mathematical step.
Claim 17 recites calculating and selecting steps, which are mathematical and mental elements as explained above.
Claim 18 recites additional information about the fingerprint, which is a mathematically produced value.
Thus, the claims recite abstract ideas and thus must be examined further to determine whether elements in addition to the abstract ideas integrate the judicial exceptions into a practical application (MPEP 2106.04(d)). [Step 2A Prong One: Yes]
Step 2A, Prong Two: If the claims recite a judicial exception under Prong One, then is the judicial exception integrated into a practical application by an additional element (MPEP 2106.04(d))?
Because the claims recite judicial exceptions, direction under Step 2A Prong Two provides that the claims must be examined further to determine whether they recite elements in addition to the abstract ideas which integrate the judicial exceptions into a practical application (MPEP 2106.04(d)). A claim can be said to integrate a judicial exception into a practical application when it applies, relies on, or uses the judicial exception in a manner that imposes a meaningful limit on the judicial exception. This is performed by analyzing the additional elements of the claim to determine if the judicial exceptions are integrated into a practical application (MPEP 2106.04(d)(I); MPEP 2106.05(a-h)). If the claim contains no additional elements beyond the judicial exceptions, the claim is said to fail to integrate the judicial exceptions into a practical application (MPEP 2106.04(d)(III)).
Claim 1 recites providing a model, where providing a model is considered data gathering to use the model and thus an insignificant extra-solution activity (MPEP 2106.05(g)).
Claim 1 recites inputting a database into the model, where inputting the data into the model is required to perform the abstract processing steps and thus is data gathering (MPEP 2106.05(g)).
Claim 1 recites processing using the encoder and decoder, which are considered to be instructions to implement an abstract idea on a generic computer. See MPEP 2106.05(f). MPEP 2106.05(f) provides the following considerations for determining whether a claim simply recites a judicial exception with the words “apply it” (or an equivalent), such as mere instructions to implement an abstract idea on a computer: (1) whether the claim recites only the idea of a solution or outcome i.e., the claim fails to recite details of how a solution to a problem is accomplished; (2) whether the claim invokes computers or other machinery merely as a tool to perform an existing process; and (3) the particularity or generality of the application of the judicial exception. Therefore, the limitations merely serve to link the judicial exception of molecule generation to the technological environment of neural networks in the form of an encoder-decoder.
Claims 6 and 16 recite training the model, which is interpreted as directed to preparing the model to generate the molecular structures and thus a data gathering step. While the claims do not explicitly recite training as a mathematical step, under a broadest reasonable interpretation in light of the specification, it could be interpreted as using a loss function (pg. 2, line 12).
Claims 6 and 16 recite encoding and decoding using the encoder and decoder, which are linking to the technological environment of neural networks for the reasons outlined above.
Claim 6 recites obtaining new data points and updating a database, which are data generation steps which do not integrate the abstract ideas into a practical application.
Claims 8 and 17 recite obtaining molecules, which is data gathering.
Claims 19-20 recite a non-transitory computer readable medium and claim 20 recites processors. These are interpreted as general purpose computer elements. The claims state nothing more than that a generic computer performs the functions that constitute the abstract idea. Hence, these are mere instructions to apply the abstract idea using a computer, and therefore the claim does not integrate that abstract idea into a practical application (see MPEP 2106.04(d) § I; and MPEP 2106.05(f)).
Thus, the claims recite elements in addition to the abstract ideas which do not integrate the abstract ideas into a practical application, and must be examined further to determine whether elements in addition to the abstract ideas provide significantly more (MPEP 2106.05). [Step 2A Prong One: Yes]
Step 2B: Do the claims recite a non-conventional arrangement of elements in addition to any identified judicial exception(s) (MPEP 2106.05)?
Claims found to be directed to a judicial exception are then further evaluated to determine if the claims recite an inventive concept that provides significantly more than the judicial exception itself. Step 2B of 101 analysis determines whether the claims contain additional elements that amount to an inventive concept, and an inventive concept cannot be furnished by an abstract idea itself (MPEP 2106.05). The claims recite a computer, interpreted as instructions to apply the abstract idea using a computer, where the computer does not impose meaningful limitations on the judicial exceptions, which can be performed without the use of a computer (MPEP 2106.04(d) § I; and MPEP 2106.05(f)).
Elements in addition to the abstract ideas include: providing a model (claim 1), inputting a database into the model (claim 1), an encoder and decoder as part of an autoencoder-based generative model (claims 1, 6, and 16), training the model (claims 6 and 16), obtaining new data and updating the database (claims 6, 8, and 17), and a non-transitory computer readable medium and processors (claims 19 and 20).
The courts have found that receiving and outputting data are well-understood, routine, and conventional functions of a computer when claimed in a merely generic manner or as insignificant extra-solution activity (see Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information), buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014) (computer receives and sends information over a network), Versata Dev. Group, Inc. v. SAP Am., Inc., 793 F.3d 1306, 1334, 115 USPQ2d 1681, 1701 (Fed. Cir. 2015), and OIP Techs., 788 F.3d at 1363, 115 USPQ2d at 1092-93 (storing and retrieving information in memory), as discussed in MPEP 2106.05(d)(II)(i)).
The additional elements reciting use of neural networks, including autoencoder models, are interpreted as mere instruction to “apply” the abstract ideas, which cannot provide an inventive concept. See MPEP 2106.05(f). Furthermore, Elton (Molecular Systems Design and Engineering 4(4): 828-849, 2019; newly cited) teaches, in a view of deep learning for molecular design, training and use of autoencoder and generative neural networks for artificial creativity to generate molecules (abstract).
Therefore, the recited additional elements, alone or in combination, do not appear to provide an inventive concept. [Step 2B: No]
Conclusion: Claims are Directed to Non-statutory Subject Matter
For these reasons, the claims, when the limitations are considered individually and as a whole,
are directed to an abstract idea and lack an inventive concept. Hence, the claimed invention does not
constitute significantly more than the abstract idea, so the claims are rejected under 35 USC § 101 as
being directed to non-statutory subject matter.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 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.
Claims 1-3, 6-9, 11-12, 15-17, and 19-20
Claim(s) 1-3, 6-9, 11-12, 15-17, and 19-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Kang (Journal of Chemical Information and Modeling 59: 43-52, 2019; newly cited) in view of Sousa (Journal of Chemical Information and Modeling 61: 5343-5361, 2021; newly cited).
Claim 1 recites providing an autoencoder-based generative model for generation of molecular structures.
Kang teaches an autoencoder for generating molecules with desired properties (abstract).
Claim 1 recites inputting into the autoencoder-based generative model a database of scored molecules, each scored molecule having an objective function value calculated from an objective function.
Kang teaches a dataset of molecules (pg. 46, col. 1, second paragraph) where training sets are labeled using an objective function (pg. 46, col. 2, last two paragraphs).
Claim 1 recites selecting scored molecules from the database with relatively larger objective
function values over other scored molecules in the database.
Kang teaches an objective function for scoring generating sequences (pg. 46, col. 2, first paragraph), which is selecting the highest.
Claim 1 recites processing the selected scored molecules through an encoder of the autoencoder-based generative model to obtain latent points in a latent space.
Kang teaches finding latent representation that closely reflects the target condition (pg. 43, col. 2, second paragraph).
Claim 1 recites selecting a latent point in the latent space and sampling neighbor latent points that are within a distance from the selected latent point.
Sousa teaches distance between hidden representations of molecules (pg. 5349, col. 2, first paragraph) and generation of novel latent points converted into molecules (pg. 5353, col. 2, fourth paragraph) where the model makes small deviations which are interpreted as neighbors in latent space, such as keeping transformed points while modifying the number of aromatic rings (pg. 5353, col. 2, fourth paragraph).
Claim 1 recites processing the sampled neighbor latent points with a decoder to generate at least
one generated molecule.
Kang uses a decoder to generate the molecule (pg. 45, col. 2, first paragraph).
Claim 1 recites providing a report having the at least one generated molecule.
Kang teaches determining an improved property based on the generated molecules (pg. 48, col. 1, second paragraph), where information about the molecule is interpreted as a report.
Claim 19 recites a non-transitory computer-readable medium performing the steps of claim 1 on a computer system.
Kang does not explicitly recite a non-transitory computer-readable medium. Kang recites computer simulation (pg. 43, col. 1, first paragraph), suggesting at least a computational environment.
Claim 20 recites a computer system comprising processors and memory performing the steps of claim 1.
Kang does not explicitly recite a processor or non-transitory computer-readable medium. Kang recites computer simulation (pg. 43, col. 1, first paragraph), suggesting at least a computational environment.
Claim 2 recites comparing the generated molecules with selected scored molecules, selecting molecules from the generated molecules that are closest to the selected scored molecules, and providing the selected molecules as candidates for having the at least one property.
Kang teaches comparing target values between training data and new molecules to see if the conditions were satisfied (pg. 48, col. 2, third paragraph), where smaller standard deviations suggests closeness, and the target value relates to a property.
Claim 3 recites the selecting is based on at least one of a fingerprint molecule clustering and sampling protocol or an acceptance function having an acceptance function value equal to 1.
Kang teaches fingerprinting (pg. 46, col. 2, second paragraph). Kang teaches selecting molecules whose properties are close (pg. 46, col. 1, first paragraph) and having smaller standard deviations in target values (pg. 48, col. 2, third paragraph), which is interpreted as clustering.
Claim 6 recites training the autoencoder-based generative model with the scored molecules.
Kang teaches a generative autoencoder (abstract) and a dataset of molecules (pg. 46, col. 1, second paragraph) where training sets are labeled using an objective function (pg. 46, col. 2, last two paragraphs).
Claim 6 recites selecting scored molecules with high objective function value that are diverse to
obtain encodable molecules.
Kang teaches an objective function for scoring generating sequences (pg. 46, col. 2, first paragraph), which is selecting the highest.
Claim 6 recites encoding the encodable molecules to latent points in the latent space using the encoder, obtaining new latent points in the latent space that are neighboring latent points to selected latent point, and decoding the new latent points into newly generated molecules using the decoder.
Kang teaches encoding into latent space and decoding out of latent space conditionally generated structures (Fig. 3).
Claim 6 recites calculating an objective function value for the newly generated molecules.
Kang teaches generating new molecules with properties close to the target condition (pg. 43, col. 2, last paragraph).
Claim 6 recites updating the database of molecules with calculated objective function value with
the newly generated molecules.
Sousa teaches iteratively updating the model (pg. 5353, col. 1, fourth paragraph).
Claim 7 recites filtering the newly generated molecules for valid molecules and selecting newly generated molecules that are closest in latent space to each other.
Kang teaches checking the validity of generating molecules (pg. 47, col. 1, last paragraph).
Claim 8 recites the newly generated molecules are selected by determining a property for a target molecule, obtaining a potential set of molecules, determining a similarity metric for the molecules in the potential set, and selecting molecules in potential set with a similarity metric that is closest to the target molecule having the property.
Kang teaches conditional molecular design wherein molecules were generated with high similarity and efficiently with respect to a target condition (pg. 48, col. 1, last paragraph).
Claim 9 recites calculating molecular descriptors of the generated molecules, calculating molecular descriptors of the selected molecules, comparing molecular descriptors of the generated molecules to molecular descriptors of the selected molecules, selecting generated molecules with molecular descriptors closest to target molecules, and providing the selected generated molecules that are closes to target molecules.
Sousa teaches measuring the divergence of distribution of physicochemical descriptors (pg. 5350, col. 1, first paragraph), where measuring and determining divergence requires comparison, and iterative fine tuning (pg. 5352, col. 1, fourth paragraph).
Claim 11 recites calculating molecular descriptors as one or more of the following: number of hydrogen bond acceptors, number of hydrogen bond donors, partition coefficient of a molecule between aqueous and lipophilic phases, a topological polar surface area, a zagreb index of molecule, or an electro topological index.
Sousa teaches conditional generation with controlling for number of hydrogen ion donors and acceptors (pg. 5355, col. 1, second paragraph).
Claim 12 recites calculating similarity metric between molecules based on the molecular descriptors and selecting generated molecules closest to the similarity metric.
Sousa teaches whether generated molecules are matching properties with the training set and a variety of physicochemical descriptors (pg. 5350, col. 1, first paragraph).
Claim 15 recites generating generated molecules with the generative model, providing a base of scored molecules, performing a selection of molecules to obtain different molecules with high scores, selecting generated molecules closest to a high score of the selected molecules from the generated molecules and the selected molecules, and identifying the selected generated molecules as candidates to have at least one defined property.
Kang uses a decoder to generate the molecule (pg. 45, col. 2, first paragraph) based on finding latent representation that closely reflects the target condition (pg. 43, col. 2, second paragraph).
Kang teaches a dataset of molecules (pg. 46, col. 1, second paragraph) where training sets are labeled using an objective function (pg. 46, col. 2, last two paragraphs).
Kang teaches an objective function for scoring generating sequences (pg. 46, col. 2, first paragraph), which is selecting the highest and thus have a desired target property.
Claim 16 recites training the autoencoder-based generative model with the selected molecules from the database.
Kang teaches an autoencoder for generating molecules with desired properties (abstract) and a dataset of molecules (pg. 46, col. 1, second paragraph) where training sets are labeled using an objective function (pg. 46, col. 2, last two paragraphs).
Claim 16 recites selecting molecules using a selection protocol, encoding molecules to latent points using encoder, creating new points in latent space using a latent space making step protocol, and decoding the new latent points to molecules using decoder.
Kang teaches finding latent representation that closely reflects the target condition (pg. 43, col. 2, second paragraph) and a decoder to generate the molecule (pg. 45, col. 2, first paragraph).
Claim 16 recites filtering new and valid molecules.
Kang teaches checking the validity of generating molecules (pg. 47, col. 1, last paragraph).
Claim 16 recites selecting molecules that are closest in latent space molecules.
Kang teaches finding latent representation that closely reflects the target condition (pg. 43, col. 2, second paragraph).
Claim 16 recites calculating objective function and adding generated molecules to the database.
Kang teaches calculating objective functions (pg. 44, third paragraph) and molecules with close properties to the target (pg. 46, col. 1, first paragraph).
Sousa teaches distance between hidden representations of molecules (pg. 5349, col. 2, first paragraph) and generation of novel latent points converted into molecules (pg. 5353, col. 2, fourth paragraph) where the model makes small deviations which are interpreted as neighbors in latent space, such as keeping transformed points while modifying the number of aromatic rings (pg. 5353, col. 2, fourth paragraph).
Claim 17 recites obtaining a batch of candidate molecules from the at least one generated molecule, calculating a descriptor vector for each candidate molecule, selecting diverse molecules from a cluster of molecules sorted by objective function value, calculating the descriptor vectors for selected diverse molecules, calculating similarity metric between molecules based on the molecular descriptors, and selecting generated molecules closest to similarity metric.
Sousa teaches descriptors generated from generated new molecules (pg. 5350, col. 1, first paragraph), ranked compounds by a descriptor, here affinity (pg. 5357, col. 1, first paragraph), building the condition vector with molecular properties such as Topological Polar Surface Area (TPSA), molecular weight, and bioactivity (pg. 5354, col. 2, last paragraph), and generating molecules satisfying the desired properties (pg. 5354, col. 2, last paragraph), interpreted as having a similar metric.
Combining Kang and Sousa
An invention would have been obvious to one of ordinary skill in the art if some motivation in the prior art would have led that person to modify prior art reference teachings to arrive at the claimed invention prior to the effective filing date of the invention. One would have to combine the teachings of Kang and Sousa because Sousa teaches distance between hidden representations of molecules (pg. 5349, col. 2, first paragraph) and generation of novel latent points converted into molecules (pg. 5353, col. 2, fourth paragraph), where determination of novel points and generation of new molecules based on the latent points would be desirable in combination with the conditional molecular design of Kang, which recites property prediction close to a predetermined target condition, which means finding latent representation close to the target in latent space (pg. 43, col. 2, last paragraph). Both Kang and Sousa are both directed to the shared field of endeavor of using machine learning like autoencoders for novel molecule generation and their combination is prima facie obvious.
Claims 4, 10, 13-14, and 18
Claim(s) 4, 10, 13-14, and 18 is/are rejected under 35 U.S.C. 103 as being unpatentable over Kang in view of Sousa as applied to claims 1-3, 6-9, 11-12, 15-17, and 19-20 above, and further in view of White (Journal of Chemical Information and Modeling 50: 1257-1274, 2010; newly cited).
Claim 4 recites the fingerprint molecule clustering and sampling protocol includes selecting scored molecules from the database that have the acceptance function value equal to l, calculating fingerprints for the selected scored molecules, clustering the selected scored molecules by a fingerprint vector, selecting a top number of molecules in each cluster, sorting the selected top number of molecules by objective function value, randomly sampling one molecules from each cluster, and providing the randomly sampled molecule from each cluster in the report.
White teaches selecting acceptable samples (pg. 1258, col. 2, second paragraph), computing fingerprints for ranking molecules (pg. 1262, col. 2, fifth paragraph), molecule clustering in space (pg. 1269, col. 2, first paragraph), randomly sampling from the projection set (pg. 158, col. 2, fourth paragraph), and expectations of generated molecules’ properties (pg. 1262, col. 2, fourth paragraph).
Claim 10 recites selecting the target molecules by a protocol that selects diverse molecules, wherein the protocol that selects diverse molecules comprises selecting scored molecules from the database that have an acceptance function value equal to l, calculating fingerprints for the selected scored molecules, clustering the selected scored molecules by a fingerprint vector, selecting a top number of molecules in each cluster, sorting the selected top number of molecules by objective function value, and
randomly sampling one molecules from each cluster, providing the randomly sampled molecule from each cluster in the report.
White teaches selecting acceptable samples (pg. 1258, col. 2, second paragraph), computing fingerprints for ranking (or sorting) molecules (pg. 1262, col. 2, fifth paragraph), molecule clustering in space (pg. 1269, col. 2, first paragraph), randomly sampling from the projection set (pg. 158, col. 2, fourth paragraph), and expectations of generated molecules’ properties (pg. 1262, col. 2, fourth paragraph).
Claim 13 recites selecting acceptable molecules with AF(x) = 1, calculating a chemical fingerprint for selected molecules, applying a clustering method on the calculated fingerprints, selecting in every cluster N molecules with highest values of objective function, and randomly choosing one molecule in every cluster from the selected molecules.
White teaches selecting acceptable samples (pg. 1258, col. 2, second paragraph), computing fingerprints for ranking (or sorting) molecules (pg. 1262, col. 2, fifth paragraph), molecule clustering in space (pg. 1269, col. 2, first paragraph), and randomly sampling from the projection set (pg. 158, col. 2, fourth paragraph).
Claim 14 recites selecting molecules with an acceptance function of 1, calculating chemical fingerprints for each selected molecule, clustering molecules by fingerprint vector, selecting top molecules in each cluster, sorting molecules by objective function, selecting molecules with relatively higher objective function in each cluster or randomly sample one molecule in each cluster.
White teaches selecting acceptable samples (pg. 1258, col. 2, second paragraph), computing fingerprints for ranking (or sorting) molecules (pg. 1262, col. 2, fifth paragraph), molecule clustering in space (pg. 1269, col. 2, first paragraph), randomly sampling from the projection set (pg. 158, col. 2, fourth paragraph), and expectations of generated molecules’ properties (pg. 1262, col. 2, fourth paragraph).
Claim 18 recites the fingerprint is a Morgan fingerprint, extended connectivity fingerprint (ECFP), or other molecular fingerprint.
Kang teaches using the extended-connectivity fingerprint (ECFP) (pg. 36, col. 2, second paragraph).
Combining Kang, Sousa, and White
An invention would have been obvious to one of ordinary skill in the art if some motivation in the prior art would have led that person to modify prior art reference teachings to arrive at the claimed invention prior to the effective filing date of the invention. One would have to combine the previously combined teachings with those of White because White teaches using fingerprints to rank molecules with respect to input molecules (1262, col. 2, sixth paragraph), which together with the molecule generation of the previously combined art, and would be desirable as a metric for comparison for molecules derived from latent space because White teaches “the fingerprint method is effective at selecting the hidden input molecules” (pg. 1268, col. 1, first paragraph). The combined art is directed to the shared field of endeavor of novel molecule generation and their combination is prima facie obvious.
Claims Free of the Prior Art
Claim 5 recites a local steps in latent space protocol includes: determining a latent point as a starting point, determining a step length, determining a number of levels, and determining a number of steps in each level. Claim 5 recites when a number of latent points in a sampled points list is less than a threshold, to perform the following: sample a number of random points in the latent space, sample neighboring points within a defined distance from the sampled random points, add the sampled neighboring points to the sample points list, increase the defined distance, and repeat the previous steps until the number of latent points in the sampled points list is equal to the threshold, and then provide the sample points list having the threshold number of latent points.
While referenced prior art teaches steps for sampling latent space for novel molecular structures, the recited steps of determining step lengths and levels for determining a specified number of structures for sampling is not taught, and thus is considered free of the prior art.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Robert J Kallal whose telephone number is (571)272-6252. The examiner can normally be reached Monday through Friday 8 AM - 4 PM 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, Olivia M. Wise can be reached at (571) 272-2249. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/Robert J. Kallal/Examiner, Art Unit 1685
Prosecution Insights