DETAILED ACTION
Applicant’s response filed 11/12/2025 has been fully considered. The following rejections and/or objections are either reiterated or newly applied.
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, 3-11, 13-16, and 18-20 are pending and under consideration in this action. Claims 2, 12, and 17 were canceled in the amendment filed 11/12/2025.
Priority
This application claims foreign priority from People's Republic of China Application No. 202110543978.2, filed 5/18/2021, as reflected in the filing receipt mailed 1/14/2022. Acknowledgment is made of applicant's claim for foreign priority under 35 U.S.C. 119 (a)-(d). Receipt is acknowledged of certified copies of papers required by 37 CFR 1.55. The claims to the benefit of priority are acknowledged and the effective filing date of claims 1, 3-11, 13-16, and 18-20 is 5/18/2021.
Drawings
The objections to the drawings are withdrawn in view of Applicant’s amendments to the specification filed 11/12/2025 (Applicant’s Remarks, Pg. 10).
Specification
The objection to the title is withdrawn in view of Applicant’s amendment to the title filed 11/12/2025 (Applicant’s Remarks, Pg. 10).
Claim Objections
The objection to claims 4, 10, 14, and 19 is withdrawn in view of Applicant’s amendments to the claims filed 11/12/2025 (Applicant’s Remarks, Pg. 10-11).
Claim Rejections - 35 USC § 112(b)
The rejection of claims 2-4, 7, and 11-20 under 35 U.S.C. 112(b) as being indefinite is withdrawn in view of Applicant’s amendments to the claims filed 11/12/2025 (Applicant’s Remarks, Pg. 11).
Claim Rejections - 35 USC § 101
Maintained Rejections
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. The claims recite both (1) mathematical concepts (mathematical relationships, formulas or equations, or mathematical calculations) and (2) mental processes, i.e., concepts performed in the human mind (including observations, evaluations, judgements or opinions) (see MPEP § 2106.04(a)).
Any newly recited portion of this rejection is necessitated by claim amendment.
Step 1:
In the instant application, claims 1 and 3-10 are directed towards a method, claims 11 and 13-15 are directed towards a machine, and claims 16 and 18-20 are directed towards a manufacture, which falls into one of the categories of statutory subject matter (Step 1: YES).
Step 2A, Prong One:
In accordance with MPEP § 2106, claims found to recite statutory subject matter (Step 1: YES) are then analyzed to determine if the claims recite any concepts that equate to an abstract idea, law of nature or natural phenomenon (Step 2A, Prong One). The following instant claims recite limitations that equate to one or more categories of judicial exceptions:
Claims 1, 11, and 16 recite a mathematical concept (i.e., using an algorithm to determine a representation) in “determining initial representation of edges connected between a plurality of atoms in a molecule based on three-dimensional structure information of the molecule”; a mathematical concept (i.e., using a formula or algorithm to determine the representation) in “determining first representation of a neighbor edge of each of the atoms based on the initial representation of the edges, the neighbor edge of each of the atoms indicating at least one edge connected with each of the atoms”; a mathematical calculation (i.e., calculating an angle) in “determining an included angle between each of the other edges except a first edge in the neighbor edge of a first atom in the plurality of atoms and the first edge to obtain a plurality of included angles”; a mathematical concept (i.e., using an algorithm to calculate a representation) in “determining the first representation of the first edge based on the plurality of included angles and initial representation of the other edges”; a mathematical concept (i.e., using a formula or algorithm to determine a representation) in “determining first representation of each of the atoms based on the first representation of the neighbor edge of each of the atoms”; and a mathematical concept (i.e., using a formula or algorithm to determine a feature representation) in “determining feature representation for characterizing the molecule based on the first representation of each of the atoms”.
Claims 3, 13, and 18 recite a mathematical concept (i.e., determining angles) in “dividing the other edges into different angle domains based on plurality of included angles”; a mathematical calculation (i.e., determining a weighted sum) in “determining weighted initial representation for each of the angle domains through weighted summation of the initial representation of the other edges in each of the angle domains based on attention weight of the other edges in each of the angle domains to the first edge”; and a mathematical concept (i.e., combining values to create a representation) in “concatenating the weighted initial representation for each of the angle domains as the first representation of the first edge”.
Claims 4, 14, and 19 recite a mathematical calculation (i.e., calculating a distance) in “determining a distance between the neighbor edge of the first atom in the plurality of atoms and the first atom, the distance between the neighbor edge and the first atom indicating a distance between a second atom connected with the neighbor edge and the first atom”; a mathematical calculation (i.e., calculating an attention weight) in “determining an attention weight of the neighbor edge of the first atom to the first atom based on the distance”; and a mathematical calculation (i.e., determining a weighted average) in “determining a weighted average of the first representation of the neighbor edge of the first atom based on the attention weight as first representation of the first atom”.
Claims 5, 15, and 20 recite a mathematical concept (i.e., using an algorithm to determine a second representation) in “determining second representation of the edges based on the first representation of each of the atoms”; a mathematical concept (i.e., using an algorithm to determine a third representation) in “determining third representation of the neighbor edge of each of the atoms based on the second representation of the edges”; a mathematical concept (i.e., using an algorithm to determine a second representation) in “determining second representation of each of the atoms based on the third representation of the neighbor edge of each of the atoms”; and a mathematical concept (i.e., using a formula or algorithm to determine a feature representation) in “determining the feature representation for characterizing the molecule based on the second representation of each of the atoms”.
Claim 6 recites a mental process (i.e., an observation of how the 3D structure information represented) in “wherein the three-dimensional structure information of the molecule is represented by a polar coordinate system”.
Claim 7 recites a mathematical concept (i.e., using an algorithm to determine an initial representation) in “determining initial representation of the plurality of atoms in the molecule based on the three-dimensional structure information of the molecule”; a mathematical concept (i.e., using an algorithm to characterize the distance) in “determining characterization of a distance between the plurality of atoms connected by the edge based on the three-dimensional structure information of the molecule”; and a mathematical concept (i.e., using an algorithm to determine the representation) in “determining the initial representation of edges connected between the plurality of atoms based on the initial representation of the plurality of atoms and the characterization of the distance between plurality of the atoms”.
Claim 8 recites a mental process (i.e., an evaluation of whether a value is less than a threshold) in “determining whether a distance between any two atoms of the plurality of atoms is less than a threshold distance”; and a mental process (i.e., an observation that since the distance is less than a threshold, the atoms are connected) in “in response to determining that the distance between these two atoms is less than the threshold distance, constructing an edge for connecting these two atoms”.
Claim 9 recites a mental process (i.e., an observation of how the 3D structure information represented) in “wherein the three-dimensional structure information of the molecule at least comprises types and space distribution of the atoms forming the molecule”.
Claim 10 recites a mental process (i.e., an observation of how the initial and feature representations are represented) and a mathematical concept (i.e., vector representation) in “wherein the initial representation of edges is a one-dimensional vector, and the feature representation for characterizing the molecule is a one-dimensional vector”.
These recitations are similar to the concepts of collecting information, and displaying certain results of the collection and analysis is Electric Power Group, LLC, v. Alstom (830 F.3d 1350, 119 USPQ2d 1739 (Fed. Cir. 2016)), comparing information regarding a sample or test to a control or target data in Univ. of Utah Research Found. v. Ambry Genetics Corp. (774 F.3d 755, 113 U.S.P.Q.2d 1241 (Fed. Cir. 2014)) and Association for Molecular Pathology v. USPTO (689 F.3d 1303, 103 U.S.P.Q.2d 1681 (Fed. Cir. 2012)), and organizing and manipulating information through mathematical correlations in Digitech Image Techs., LLC v Electronics for Imaging, Inc. (758 F.3d 1344, 111 U.S.P.Q.2d 1717 (Fed. Cir. 2014)) that the courts have identified as concepts that can be practically performed in the human mind or mathematical relationships.
The abstract ideas recited in the claims are evaluated under the broadest reasonable interpretation (BRI) of the claim limitations when read in light of and consistent with the specification, and are determined to be directed to mental processes that in the simplest embodiments are not too complex to practically perform in the human mind. Additionally, the recited limitations that are identified as judicial exceptions from the mathematical concepts grouping of abstract ideas are abstract ideas irrespective of whether or not the limitations are practical to perform in the human mind. The instant claims must therefore be examined further to determine whether they integrate the abstract idea into a practical application (Step 2A, Prong One: YES).
Step 2A, Prong Two:
In determining whether a claim is directed to a judicial exception, further examination is performed that analyzes if the claim recites additional elements that when examined as a whole integrates the judicial exception(s) into a practical application (MPEP § 2106.04(d)). A claim that integrates a judicial exception into a practical application will apply, rely on, or use the judicial exception in a manner that imposes a meaningful limit on the judicial exception. The claimed additional elements are analyzed to determine if the abstract idea is integrated into a practical application (MPEP § 2106.04(d)(I)). If the claim contains no additional elements beyond the abstract idea, the claim fails to integrate the abstract idea into a practical application (MPEP § 2106.04(d)(III)). The following claims recite limitations that equate to additional elements:
Claims 1 recites “wherein the feature representation for characterizing the molecule is used for prediction of various biochemical properties of the molecule”.
Claim 11 recites “at least one processor”, “a memorizer, in communication connection with the at least one processor”, and “the memorizer stores an instruction capable of being performed by the at least one processor, and the instruction is performed by the at least one processor”.
Claim 16 recites “a non-transitory storage medium, storing a computer instruction”.
Regarding the above cited limitations in claims 11 and 16 of (i) at least one processor, (ii) a memorizer, in communication connection with the at least one processor, (iii) the memorizer stores an instruction capable of being performed by the at least one processor, and the instruction is performed by the at least one processor, and (iv) a non-transitory storage medium, storing a computer instruction. These limitations require only a generic computer component, which does not improve computer technology. Therefore, these limitations equate to mere instructions to implement an abstract idea on a generic computer, which the courts have established does not render an abstract idea eligible in Alice Corp. 573 U.S. at 223, 110 USPQ2d at 1983.
Regarding the above cited limitations in claim 1 of (v) wherein the feature representation for characterizing the molecule is used for prediction of various biochemical properties of the molecule. This limitation equates to insignificant, extra-solution activity of mere data gathering because these limitations gather data after the recited judicial exceptions of determining the feature representation of characterizing the molecule (see MPEP § 2106.04(d)). As such, claims 1, 3-11, 13-16, and 18-20 are directed to an abstract idea (Step 2A, Prong Two: NO).
Step 2B:
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). The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception because the claims recite additional elements that equate to well-understood, routine and conventional (WURC) limitations (MPEP § 2106.05(d)). The instant claims recite same additional elements described in Step 2A, Prong Two above.
Regarding the above cited limitations in claims 11 and 16 of (i) at least one processor, (ii) a memorizer, in communication connection with the at least one processor, (iii) the memorizer stores an instruction capable of being performed by the at least one processor, and the instruction is performed by the at least one processor, and (iv) a non-transitory storage medium, storing a computer instruction. These limitations equate to instructions to implement an abstract idea on a generic computing environment, which the courts have established does not provide an inventive concept (see MPEP § 2106.05(d) and MPEP § 2106.05(f)).
Regarding the above cited limitations in claim 1 of (v) wherein the feature representation for characterizing the molecule is used for prediction of various biochemical properties of the molecule. This limitation when viewed individually and in combination, is a WURC limitation as taught by Shui et al. (Heterogenous Molecular Graph Neural Networks for Predicting Molecule Properties. arXiv preprint. aiXiv:2009.12710v1. https://arxiv.org/abs/2009.1271 0; previously cited). Shui et al. discloses a novel graph representation of molecules, heterogeneous molecular graph (HMG), which allows graph learning methods to explicitly model many-body representation, interaction, and prediction (Pg. 492, Col. 2, Para. 3 and Pg. 493, Col. 1, Para. 2). Shui et al. further discloses that their model is effective at explicitly modeling and computing many-body predictions, beneficial in chemical prediction tasks (limitation (v)) (Pg. 498, Col. 1, Para. 3).
These additional elements do not comprise an inventive concept when considered individually or as an ordered combination that transforms the claimed judicial exception into a patent-eligible application of the judicial exception. Therefore, the instant claims do not amount to significantly more than the judicial exception itself (Step 2B: NO). As such, claims 1, 3-11, 13-16, and 18-20 are not patent eligible.
Response to Arguments under 35 U.S.C. 101
Applicant’s arguments filed 11/12/2025 have been fully considered but they are not persuasive.
Applicant argues that the claims are not directed to a mental process because the claims recite activity that falls outside the enumerated sub-groups of methods that may be performed mentally, a person forming judgment, performing a mental process that requires a generic computer, performing a mental process in a computer environment, or using a computer as a tool to perform a mental process. Based on the specification, and by using the information processing method according to claims 1-20, feature representation for characterizing the molecule is determined not only based on topological structure information, but also based on spatial structure information, for example, an angle and distance between atoms forming the molecule. (Applicant’s Remarks, Pg. 13-14). Applicant’s arguments are not persuasive for the following reasons:
As recited in Step 2A, Prong One above, the following limitations in independent claims 1, 11, and 16 recite mathematical concepts: “determining initial representation of edges connected between a plurality of atoms in a molecule based on three-dimensional structure information of the molecule”, “determining first representation of a neighbor edge of each of the atoms based on the initial representation of the edges, the neighbor edge of each of the atoms indicating at least one edge connected with each of the atoms”, “determining an included angle between each of the other edges except a first edge in the neighbor edge of a first atom in the plurality of atoms and the first edge to obtain a plurality of included angles”, “determining the first representation of the first edge based on the plurality of included angles and initial representation of the other edges”, “determining first representation of each of the atoms based on the first representation of the neighbor edge of each of the atoms”, and “determining feature representation for characterizing the molecule based on the first representation of each of the atoms”. As an example, and to not reiterate the rejection above, the broadest reasonable interpretation (BRI) of the limitation of “determining an included angle between each of the other edges except a first edge in the neighbor edge of a first atom in the plurality of atoms and the first edge to obtain a plurality of included angles” is calculating an angle, which is a mathematical concept.
The following limitations in dependent claims 6 and 8-9 recite mental processes: “wherein the three-dimensional structure information of the molecule is represented by a polar coordinate system”, “determining whether a distance between any two atoms of the plurality of atoms is less than a threshold distance”, “in response to determining that the distance between these two atoms is less than the threshold distance, constructing an edge for connecting these two atoms”, and “wherein the three-dimensional structure information of the molecule at least comprises types and space distribution of the atoms forming the molecule”. The BRI of “wherein the three-dimensional structure information of the molecule is represented by a polar coordinate system” and “wherein the three-dimensional structure information of the molecule at least comprises types and space distribution of the atoms forming the molecule” is an observation of how the 3D structural information is represented, which can practically be performed in the human mind. The BRI of “determining whether a distance between any two atoms of the plurality of atoms is less than a threshold distance” and “in response to determining that the distance between these two atoms is less than the threshold distance, constructing an edge for connecting these two atoms” is an evaluation of whether or not a value/distance is less than a threshold, which can also practically be performed in the human mind.
Therefore, the limitations reciting judicial exceptions in independent claims 1, 11, and 16 are all mathematical concepts. The limitations that recite mental process are in dependent claims 6 and 8-9. As described directly above, these mental process limitations can be practically performed in the human mind. The claims therefore recite abstract ideas and this argument is thus not persuasive.
Applicant argues that the independent claims embody at least one practical application of the invention. The subject matter of the claims embraces the generation and analysis of three-dimensional structures of molecules using a graph neural network to determine the angles and edges to characterize the molecule. Indeed, the determination of the initial representation of edges connected to the plurality of atoms in a molecule and the determination of different representations of the edges of these atoms may be considered a transformation or reduction of an atom in a molecule to a different state as the three-dimensional version of the molecule for further analysis and determination of relationship, edge distances, and angles (see MPEP 2106.04(d).I) (Applicant’s Remarks, Pg. 14-16). Applicant’s arguments are not persuasive for the following reasons:
MPEP 2106.04(d)(II) recites:
The analysis under Step 2A Prong Two is the same for all claims reciting a judicial exception, whether the exception is an abstract idea, a law of nature, or a natural phenomenon (including products of nature). Examiners evaluate integration into a practical application by: (1) identifying whether there are any additional elements recited in the claim beyond the judicial exception(s); and (2) evaluating those additional elements individually and in combination to determine whether they integrate the exception into a practical application, using one or more of the considerations introduced in subsection I supra, and discussed in more detail in MPEP §§ 2106.04(d)(1), 2106.04(d)(2), 2106.05(a) through (c) and 2106.05(e) through (h).
The limitations of determination of the initial representation of edges connected to the plurality of atoms in a molecule and the determination of different representations of the edges of these atoms have been identified as judicial exceptions in Step 2A, Prong One above. The integration of a judicial exception into a practical application can only be achieved by additional elements, not by a limitation that recites a judicial exception. Thus, the recited limitations are not considered as a practical application of the abstract ideas, or as an improvement in the analysis of structures using graph neural networks.
Additionally, the recited limitation of “wherein the feature representation for characterizing the molecule is used for prediction of various biochemical properties of the molecule” has been identified as an additional element in Step 2A, Prong Two above. Further analysis at Step 2B shows that this limitation is a WURC limitation as taught by Shui et al. This additional element does not provide an inventive concept that transforms the claimed judicial exceptions into a patent-eligible application. Therefore, the claims do not recite a practical application of the alleged abstract idea, and this argument is not persuasive.
Applicant argues that the independent claims recite significantly more than the abstract ideas. Applicant has described the transformation of the initial representation of the plurality of atoms in a molecular to different representations of the edges of these atoms may be considered a transformation or reduction of an atom in a molecule to a different state as the three-dimensional version of the molecule for further analysis and determination of relationship, edge distances, and angles. Thus, and as provided above, Applicant again resubmits that the claims recite significantly more than the alleged abstract idea (Applicant’s Remarks, Pg. 16-17). Applicant’s arguments are not persuasive for the following reasons:
As discussed in the arguments directly above, the limitations of determination of relationship, edge distances, and angles recite judicial exceptions in Step 2A, Prong One above. The integration of a judicial exception into a practical application can only be achieved by additional elements, not by a limitation that recites a judicial exception. Thus, the recited limitations are not considered to provide significantly more than the abstract ideas. This argument is thus not persuasive.
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.
Claims 1, 5, 7-9, 11, 15-16, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Shui et al. (Heterogenous Molecular Graph Neural Networks for Predicting Molecule Properties. arXiv preprint. aiXiv:2009.12710v1. https://arxiv.org/abs/2009.12710; published 9/26/2020; previously cited) in view of Chang et al. (Geometric Graph Representations and Geometric Graph Convolutions for Deep Learning on Three-Dimensional (3D) Graphs. arXiv preprint. arXiv:2006.01785v1. https://doi.org/10.48550/arXiv.2006.01785; published 6/2/2020; previously cited).
This rejection is newly recited and necessitated by claim amendment.
Regarding claim 1, Shui et al. teaches a novel graph representation of molecules, heterogeneous molecular graph (HMG), which allows graph learning methods to explicitly model many-body representation, interaction, and prediction (i.e., an information processing method) (Pg. 492, Col. 2, Para. 3 and Pg. 493, Col. 1, Para. 2). Shui et al. further teaches an example of an HMG in Fig. 1. The spatial structure of formaldehyde is shown, wherein each atom of the molecule is associated with three-dimensional coordinates in Euclidean space. In the molecular graph, edges are denoted for distances that are less than a cutoff distance of c=2 (i.e., determining initial representation of edges connected between a plurality of atoms in a molecule based on three-dimensional structure information of the molecule) (Pg. 494, Fig. 1). Shui et al. further teaches that an HMG of order two was constructed based on the molecular graph of the formaldehyde molecule. There are three types of edges (1-1, 2-2, and 1-2) in the HMG. Edges between nodes of the same order are associated with features that depict the geometric relation between the nodes (distance for 1-1 edges, angles for 2-2 edges) (i.e., determining first representation of a neighbor edge of each of the atoms based on the initial representation of the edges, the neighbor edge of each of the atoms indicating at least one edge connected with each of the atoms) (Pg. 494, Fig. 1). Shui et al. further teaches that there are two types of nodes (1-bodies and 2-bodies) in an HMG (Pg. 494, Fig. 1). Shui et al. further teaches that the input module of the Heterogeneous Molecular Graph Neural Network converts raw features of nodes to latent embeddings (i.e., determining first representation of each of the atoms based on the first representation of the neighbor edge of each of the atoms) (Pg. 295, Col. 1, Para. 3). Shui et al. further teaches that in an HMG, each node i of order p is associated with a discrete feature
Z
p
,
i
that indicates its atomic composition, and a continuous feature
x
p
.
i
that describes aspects of its geometry (i.e., determining feature representation for characterizing the molecule based on the first representation of each of the atoms) (Pg. 494, Col. 1, Para. 2). Shui et al. further teaches that their model is effective at explicitly modeling and computing many-body predictions, beneficial in chemical prediction tasks (i.e., wherein the feature representation for characterizing the molecule is used for prediction of various biochemical properties of the molecule) (Pg. 498, Col. 1, Para. 3).
Regarding claim 5, Shui et al. teaches the computational flow of heterogeneous molecular graph neural networks (HMGNN) for many-bodies up to order two in Fig. 2. For each node i of order p, an input module converts the discrete and continuous feature of the node to an initial node embedding
h
p
,
i
(
0
)
. HMGNN passes the initial embeddings through a stack of T interaction modules to encode information from its neighbor nodes of different orders to the node embedding. The outputs of the last interaction module, and the final node embedding
h
p
,
i
(
T
+
1
)
, are then fed into a fusion module and an output module to compute a weight vector α and prediction
y
^
p
,
i
, respectively (i.e., the second and third representations are part of the stack of T interaction modules, determining second representation of the edges based on the first representation of each of the atoms; determining third representation of the neighbor edge of each of the atoms based on the second representation of the edges; determining second representation of each of the atoms based on the third representation of the neighbor edge of each of the atoms) (Pg. 496, Fig. 2). Shui et al. further teaches that each many-body order p possesses a specific output module that passes the output of its interaction module, final node embeddings
h
p
,
i
(
T
+
1
)
, through a sequence of linear mappings and an aggregation process to compute the estimated value of the target property (Pg. 495, Col. 2, Para. 2). Shui et al. further teaches that an embedding lookup table is used to map the discrete feature
Z
p
,
i
and the continuous feature
x
p
.
i
(i.e., determining the feature representation for characterizing the molecule based on the second representation of each of the atoms) (Pg. 495, Col. 1, Para. 3).
Regarding claim 7, Shui et al. teaches an example of a heterogenous molecular graph for the formaldehyde molecule. Each atom in the molecule is associated with three-dimensional coordinates in Euclidean space. The molecular graph shows the atom types (C, O, H) and the bond/edge distances (i.e., determining initial representation of the plurality of atoms in the molecule based on the three-dimensional structure of the molecule) (Pg. 494, Fig. 1). Shui et al. further teaches that the molecular graph of the molecule is generated with a cutoff distance of c=2 for the edges. They converted atom coordinates to pair-wise distances to guarantee translation and rotation invariance of the representation (Pg. 494, Fig 1). Shui et al. further teaches that the edge feature characterizes geometric relation between two nodes, e.g., the distance between atoms or the angles between bonds (i.e., determining characterization of a distance between the plurality of atoms connected by the edge based on the three-dimensional structure of the molecule) (Pg. 494, Col. 1, Para. 2). Shui et al. further teaches that they denoted edges whose distances are less than c using black solid black lines and the edges that are broke by the cutoff using black dotted lines (i.e., determining initial representation of edges connected between the plurality of atoms based on the initial representation of the plurality of atoms and the characterization of the distance between plurality of atoms) (Pg. 494, Fig. 1).
Regarding claim 8, Shui et al. teaches that two atoms are connected in a molecular graph when the Euclidean distance between them is less than a cutoff threshold (i.e., determining whether a distance between any two atoms of the plurality of atoms is less than a threshold distance) (Pg. 493, Col. 2, Para. 3). Shui et al. further teaches that each edge in the graph is associated with a distance to store the geometric structure of the molecule (i.e., in response to determining that the distance between two atoms is less than the threshold distance, constructing an edge for connecting these two atoms) (Pg. 493, Col. 2, Para. 3).
Regarding claim 9, Shui et al. teaches an example of a heterogenous molecular graph for a formaldehyde molecule. The molecular graph takes into account the spatial structure based on the three-dimensional coordinates in Euclidean space, as well as the atom types (e.g., C, O, and H) (i.e., wherein the three-dimensional structure information of the molecule at least comprises types and space distribution of the atoms forming the molecule) (Pg. 494, Fig. 1).
Regarding claim 11, Shui et al. teaches the computational workflow of heterogeneous molecular graph neural networks (Pg. 496, Fig. 2). Shui et al. further teaches that modern computing architectures such as graphics processing unit (GPU) and tensor processing unit (TPU) are optimized to accelerate this computation (i.e., on a computing device containing at least one processor; a memorizer, in communication connection with the at least one processor; and a memorizer stores an instruction capable of being performed by the at least one processor, and the instruction is performed by the at least one processor, to cause the at least one processor to perform steps) (Pg. 497, Col. 1, Para. 3). Shui et al further teaches the limitations of determining initial representation of edges connected between a plurality of atoms in a molecule based on three-dimensional structure information of the molecule; determining first representation of a neighbor edge of each of the atoms based on the initial representation of the edges, the neighbor edge of each of the atoms indicating at least one edge connected with each of the atoms; determining first representation of each of the atoms based on the first representation of the neighbor edge of each of the atoms; and determining feature representation for characterizing the molecule based on the first representation of each of the atoms as described for claim 1 above.
Regarding claim 15, Shui et al. teaches the limitations of determining second representation of the edges based on the first representation of each of the atoms; determining third representation of the neighbor edge of each of the atoms based on the second representation of the edges; determining second representation of each of the atoms based on the third representation of the neighbor edge of each of the atoms; and determining the feature representation for characterizing the molecule based on the second representation of each of the atoms as described for claim 5 above.
Regarding claim 16, Shui et al. teaches that modern computing architectures such as graphics processing unit (GPU) and tensor processing unit (TPU) are optimized to accelerate the computation (i.e., on a computing device containing a non-transitory storage medium, storing a computer instruction, wherein the computer instruction is used for a computer to perform steps) (Pg. 497, Col. 1, Para. 3). Shui et al further teaches the limitations of determining initial representation of edges connected between a plurality of atoms in a molecule based on three-dimensional structure information of the molecule; determining first representation of a neighbor edge of each of the atoms based on the initial representation of the edges, the neighbor edge of each of the atoms indicating at least one edge connected with each of the atoms; determining first representation of each of the atoms based on the first representation of the neighbor edge of each of the atoms; and determining feature representation for characterizing the molecule based on the first representation of each of the atoms as described for claim 1 above.
Regarding claim 20, Shui et al. teaches the limitations of determining second representation of the edges based on the first representation of each of the atoms; determining third representation of the neighbor edge of each of the atoms based on the second representation of the edges; determining second representation of each of the atoms based on the third representation of the neighbor edge of each of the atoms; and determining the feature representation for characterizing the molecule based on the second representation of each of the atoms as described for claim 5 above.
Shui et al. does not teach determining an included angle between each of the other edges except a first edge in the neighbor edge of a first atom in the plurality of atoms and the first edge to obtain a plurality of included angles; and determining the first representation of the first edge based on the plurality of included angles and initial representation of the other edges.
Regarding claims 1, 11, and 16, Chang et al. teaches the incorporation of the full geometry of 3D graph convolutions using standard graph convolutional networks by (1) expanding the kinds of edges involved to include not just edges (e) with neighbor nodes, but also angle edges (eθ) with second-neighbor nodes and dihedral edges (eφ) with third-neighbor nodes and (2) assigning different weights to different edges based on their kind and distance (Pg. 9, Para. 4). Chang et al. further teaches that the angle edge is the angle formed between nodes across two edges and the dihedral edge is the angle between the plane formed by the first two edges and the plane formed by the last two edges for three edges connected in a chain (i.e., determining an included angle between each of the other edges except a first edge in the neighbor edge of a first atom in the plurality of atoms and the first edge to obtain a plurality of included angles) (Pg. 3, Para. 3). Chang et al. further teaches that a simple edge weight/edge distance correlation scheme was employed, wherein the parameters can be fixed using a reference value or determined using Bayesian hyperparameter optimization. The correlations for edges, angle edges, and dihedral edges were calculated (i.e., determining the first representation of the first edge based on the plurality of included angles and initial representation of the other edges) (Pg. 10, Para. 1-4).
Therefore, regarding claims 1, 5, 7-9, 11, 15-16, and 20, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the neural network for predicting molecular properties shown by Shui et al. with the teachings of Chang et al. because incorporating the distance-geometric representation for angles and dihedrals showed significant improvement in accuracy over standard graph convolutional networks (Chang et al., Pg. 12, Para. 1 and Pg. 14, Para. 2-3). One of ordinary skill in the art would be able to combine the teachings of Shui et al. with Chang et al. with reasonable expectation of success due to the same nature of the problem to be solved, since both are drawn towards a method for implementing 3D structural data into a graph network. Therefore, regarding claims 1, 5, 7-9, 11, 15-16, and 20, the instant invention is prima facie obvious (MPEP § 2142).
Claims 4, 10, 14, and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Shui et al. in view of Chang et al. as applied to claims 1, 5, 7-9, 11, 15-16, and 20 above, and further in view of Xiong et al. (Pushing the Boundaries of Molecular Representation for Drug Discovery with the Graph Attention Mechanism. J. Med. Chem. 63(16): 8749-8760 (2020); published 8/13/2019; previously cited).
This rejection is newly recited and necessitated by claim amendment.
Regarding claims 4, 14, and 19, Shui et al. teaches an example molecular graph of the formaldehyde molecule in Fig. 1. The distances between all nodes, and neighboring nodes/edges are shown (i.e., determining a distance between the neighbor edge of the first atom in the plurality of atoms and the first atom, the distance between the neighbor edge and the first atom indicating a distance between a second atom connected with the neighbor edge and the first atom) (Pg. 494, Fig. 1).
Regarding claim 10, Shui et al. teaches that each edge in the graph is associated with a distance to store the geometric structure of the molecule (Pg. 493, Col. 2, Para. 3). Shui et al. further teaches that edges between nodes depict the geometric relation between the nodes (i.e., the initial representation of edges is a one-dimensional vector) (Pg. 494, Fig. 1).
Shui et al. in view of Chang et al., as applied to claims 1, 5, 7-9, 11, 15-16, and 20 above, does not teach determining an attention weight of the neighbor edge of the first atom to the first atom based on the distance; determining a weighted average of the first representation of the neighbor edge of the first atom based on the attention weight as a first representation of the first atom; and the feature representation for characterizing the molecule is one-dimensional vector.
Regarding claims 4, 14, and 19, Xiong et al. further teaches the graph attention mechanism in a single attentive layer in Figure 1c. When applying attention to atom 3, the state vector of atom 3 is aligned with the state vector of its neighbors 2, 4, and 5, in which the features of connecting bonds must have also been embedded by a fully connected layer. Then, the weight that measures how much attention they want to assign to the neighbors is calculated by a softmax function (i.e., determining an attention weight of the neighbor edge of the first atom to the first atom based on the distance) (Pg. 8753, Col. 1, Para. 2). Xiong et al. further teaches that the state vector is then subjected to weighting to obtain the attention context (
C
v
0
) of atom v (i.e., determining a weighted average of the first representation of the neighbor edge of the first atom based on the attention weight as first representation of the first atom) (Pg. 8753, Col. 1, Para. 1).
Regarding claim 10, Xiong et al. teaches that to create a one-hot encoding feature, all of the candidate categorical variables of the feature are listed and marked as either 1 or 0 by their matches to those variables. For example, a vector of 16 bits is defined to encode atomic symbols, and a vector of 6 bits is defined to encode the hybridization state (i.e., the feature representation for characterizing a molecule is one-dimensional vector) (Pg. 8752, Col. 1, Para. 3 – Col. 2, Para. 1).
Therefore, regarding claims 4, 10, 14, and 19, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the neural network for predicting molecular properties shown by Shui et al. in view of Chang et al. with the teachings of Xiong et al. because the adoption of graph attention mechanisms at both the atom and molecule levels allows the new representation framework to learn both local and nonlocal properties of a given chemical structure. Accordingly, it captures subtle substructure patterns such as intramolecular hydrogen bonding and aromatic systems, contributing to its excellent learning capability for a wide range of different molecular properties (Xiong et al., Pg. 8758, Col. 2, Para. 3). One of ordinary skill in the art would be able to combine the teachings of Shui et al. in view of Chang et al. with Xiong et al. with reasonable expectation of success due to the same nature of the problem to be solved, since both are drawn towards a method for implementing molecular and structural data into graph neural networks. Therefore, regarding claims 4, 10, 14, and 19, the instant invention is prima facie obvious (MPEP § 2142).
Claim 6 is rejected under 35 U.S.C. 103 as being unpatentable over Shui et al. in view of Chang et al. as applied to claims 1, 5, 7-9, 11, 15-16, and 20 above, and further in view of Kim et al. (CyCNN: A Rotational Invariant CNN using Polar Mapping and Cylindrical Convolution Layers. arXiv preprint. aiXiv:2007.10588v1. https://doi.org/10.48550/arXiv.2007.10588; published 6/21/2020; previously cited).
This rejection is newly recited and necessitated by claim amendment.
Shui et al. in view of Chang et al., as applied to claims 1, 5, 7-9, 11, 15-16, and 20 above, does not teach wherein the three-dimensional information of the molecule is represented by a polar coordinate system.
Regarding claim 6, Kim et al. teaches a deep convolutional neutral network, called CyCNN, which exploits polar mapping of input images to convert rotation to translation (i.e., wherein the three-dimensional structure information of the molecule is represented by a polar coordinate system) (Abstract).
Therefore, regarding claim 6, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the neural network for predicting molecular properties of Shui et al. in view of Chang et al. with the teachings of Kim et al. because one major advantage of the CyCNN is that the polar coordinate conversion and cylindrical convolution can be easily applied to any conventional convolutional neural network without significant slowdown nor the need for more memory (Kim et al., Pg. 8, Col. 2, Para. 3). One of ordinary skill in the art would be able to combine the teachings of Shui et al. in view of Chang et al. with Kim et al. with reasonable expectation of success due to the same nature of the problem to be solved, since both are drawn towards a neural network wherein the input representation aims to achieve rotational invariance. Therefore, regarding claim 6, the instant invention is prima facie obvious (MPEP § 2142).
Response to Arguments under 35 U.S.C. 102/103
Applicant’s arguments filed 11/12/2025 have been fully considered but they are not persuasive.
Applicant argues that none of the references disclose the limitations, "determining the first representation of the neighbor edge of each of the atoms comprises: determining an included angle between each of the other edges except a first edge in the neighbor edge of a first atom in the plurality of atoms and the first edge to obtain a plurality of included angles; and determining the first representation of the first edge based on the plurality of included angles and initial representation of the other edges." Chang only discloses an angle distance based on the measure of distance derived naturally from the polar coordinate system by first converting 3D molecular diagram to the polar representation. See Chang at page 5, paragraph 3. The angle distance described in Chang only describes the characteristics of the edge determined by the angle information, and more specifically, the angle distance refers to the angle formed by the edges, but not directly to determine the first representation of the edges. Therefore, the angle distance of Chang focuses on how to encode angles from distance information instead of directly measuring or using angle information between two edges as described and claimed in the instant application. (Applicant’s Remarks, Pg. 17-19). Applicant’s arguments are not persuasive for the following reasons:
As described in the rejection above, Chang et al. teaches that that the angle edge is the angle formed between nodes across two edges and the dihedral edge is the angle between the plane formed by the first two edges and the plane formed by the last two edges for three edges connected in a chain (Pg.3, Para. 3). Chang et al. further teaches that a simple edge weight/edge distance correlation scheme, wherein the correlations for edges, angle edges and dihedral edges were calculated (Pg. 10, Para. 1-4). Chang et al. further teaches a feasibility study using the ESOL and FreeSolv datasets to compare the use of standard graph convolutions with the incorporation of geometry (i.e., edge distance, angle distance, dihedral distances, etc.) The incorporation of geometry significantly improves the results compared to those of standard graph convolutions (Abstract and Pg. 11, Para. 5 – Pg. 14, Para. 1). Chang et al. therefore teaches the limitations of “determining an included angle between each of the other edges except a first edge in the neighbor edge of a first atom in the plurality of atoms and the first edge to obtain a plurality of included angles” (Pg. 3, Para. 3 and Pg. 5, Para. 3) and “determining the first representation of the first edge based on the plurality of included angles and initial representation of the other edges” (Pg. 10, Para. 1-4 and Pg. 11, Para. 5 – Pg. 14, Para. 1). As indicated by Applicant, Chang teaches how to encode the angles from distance information (Applicant’s Remarks, Pg. 19, Para. 1). Chang et al. also teaches “the use of angle information between two edges” (Applicant’s Remarks, Pg. 19, Para. 1), because Chang et al. uses the encoded information in a feasibility study, with improved results over standard methods (Abstract and Pg. 11, Para. 5 – Pg. 14, Para. 1). Therefore, Shui et al. in view of Chang et al. teaches all the limitations in independent claims 1, 11, and 16. This argument is thus not persuasive.
Conclusion
No claims allowed.
Claims 3, 13, and 18 are free of the prior art as described in the Non-Final Office Action mailed 8/12/2025.
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Inquiries
Any inquiry concerning this communication or earlier communications from the examiner should be directed to DIANA P SANFORD whose telephone number is (571)272-6504. The examiner can normally be reached Mon-Fri 8am-5pm 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, Karlheinz Skowronek can be reached at (571)272-9047. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/D.P.S./Examiner, Art Unit 1687
/Karlheinz R. Skowronek/Supervisory Patent Examiner, Art Unit 1687
Prosecution Insights