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
The claim set received 08 August 2025 has been entered into the application.
Claims 1-4, 6, and 10-15 are amended.
Claim 21-25 are new.
Claims 16-20 are previously canceled.
Claims 1-15 and 21-25 are pending.
Election/Restrictions
Claims 16-20 withdrawn from further consideration pursuant to 37 CFR 1.142(b) as being drawn to a nonelected invention, there being no allowable generic or linking claim. Election was made without traverse in the reply filed on 07 February 2025.
New claims 21-25 are drawn to similar methods as elected claims 1-15.
Claims 1-15 and 21-25 are pending examination.
Priority
This Application is a 371 of PCT/CA2020/050005 filed 20 January 2020 which claims benefit to U.S. Provisional Application 62/788,682 filed 04 January 2019.
Information Disclosure Statement
The information disclosure statements (IDS’s) submitted on 12 June 2025 is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement has been considered by the examiner.
Claim Rejections - 35 USC § 112
35 USC § 112(b)
The rejection of claim 6 under 35 U.S.C. 112(b) n the Office Action mailed 20 May 2025 is withdrawn in view of the amendments filed 20 August 2025.
Claim Rejections - 35 USC § 101
The rejection of claim 6 under 35 U.S.C. 101 in the Office Action mailed 20 May 2025 is withdrawn in view of the amendments filed 20 August 2025.
Claim Rejections - 35 USC § 103
The instant rejection is maintained for reason for record in the Office Action mailed 20 May 2025 and modified in view of the amendments filed 20 August 2025. It is noted the amendments received 20 August 2025 necessitated new ground(s) of rejection.
The rejection of claim(s) 1 and 13-15 under 35 U.S.C. 103 as being unpatentable over Rayhan et al. (Scientific reports, 2017-12, Vol.7 (1), p.17731-18, Article 17731) in view of Chen et al. (Briefings in bioinformatics, 2016-07, Vol.17 (4), p.696-712) in the Office Action mailed 20 May 2025 is withdrawn in view of the amendments filed 20 August 2025.
The rejection of claim(s) 5 and 7 under 35 U.S.C. 103 as being unpatentable over Rayhan in view of Chen, as applied to claims 1 and 13-15 above, and in further view of Maggiore et al. (Journal of medicinal chemistry, 2014-04, Vol.57 (8), p.3186-3204) in the Office Action mailed 20 May 2025 is withdrawn in view of the amendments filed 20 August 2025.
The rejection of claim (s) 10 and 12 are rejected under 35 U.S.C. 103 as being unpatentable over Rayhan in view of Chen, as applied to claims 1 and 13-15 above, and in further view of Shi et al. (Genomics (San Diego, Calif.), 2019-12, Vol.111 (6), p.1839-185) in the Office Action mailed 20 May 2025 is withdrawn in view of the amendments filed 20 August 2025.
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.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claim(s) 1-2, 4, 12, 15, 21-23, and 25 are rejected under 35 U.S.C. 103 as being unpatentable over Tong et al. (Patent Pub: US 2012/0239367, Patent Pub Date: 09 September 2012 ) in view of Zhou et al. (US Patent Pub: US 2007/0244651, Patent Pub Date: 18 Oct 20074) in view of Fan et al. (Patent Pub: US 2018/0341754, Patent Pub Date: 29 Nov 2018) in view of Chiang et al. (The Journal of chemical physics, 2016-12, Vol.145 (23), p.234109-234109).
Claims 1 is drawn to a method for predicting drug-target binding using synthetically augmented data while claim 15 is drawn to non-transitory computer readable medium storing the method of claim 1 and claim 21 is drawn to a system. Claims 1, 15, and 21 recite similar limitations. Therefore, claims 1, 15, and 21 are examined similarly.
Claims 1, 15, and 21 recite generating a plurality of ghost ligands for a plurality of proteins in a protein structure database. Claims 1, 15, and 21 recite wherein generating the plurality of ghost ligands comprises Claims 1, 15, and 21 recite for a cluster of proteins, selected from the plurality of proteins: performing a three-dimensional structure alignment for the proteins in the cluster of proteins. Claims 1, 15, and 21 recite obtaining the plurality of ghost ligands by projecting a ligand of one of the proteins in the cluster onto other proteins in the cluster, after three-dimensional structure alignment. Claims 1, 15, and 21 recite obtaining, for each of the plurality of ghost ligands, a confidence score. Claims 1, 15, and 21 recite generating a plurality of drug-target interaction (DTI) features for proteins and ligands in a DTI database, using the plurality of ghost ligands. Claims 1, 15, and 21 recite generating a neural network using the plurality of DTI features. Claims 1, 15, and 21 recite predicting a likelihood of interaction for a combination of a query protein and a query ligand using the neural network.
Tong et al. (Tong) teach a ligand maybe an atom, an ion, a molecule, such as drugs, inhibitors, hormones, peptides, drugs, and small molecules [disclosure page 1 left col para 0003]. Tong discloses using at least one source ligand for generating a plurality of additional ligands [Tong, claim 1 step (i)]. Tong discloses generating additional ligands by identifying at least one base ligand-receptor interaction between the at least one source ligand and the receptor and modifying a portion of the corresponding source ligand selected according to the base ligand-receptor interaction, to produce at least one modified ligand [Tong, claim 3]. Tong discloses modifying a portion of the source ligand further comprises the sub-step of replacing the side chain coordinates of the amino acid residue of the source ligand with the side chain coordinates of a different amino acid residue [Tong, claim 6]. Tong discloses the database comprises a plurality of ligand-receptor interactions and each ligand-receptor interaction in the database is defined by ligand contact elements and receptor contact elements of the ligand-receptor interaction [Tong, claim 7], as in claims 1, 15, and 21 generating a plurality of ghost ligands for a plurality of proteins in a protein structure database.
Tong discloses wherein the database comprises a plurality of ligand-receptor interactions and each ligand-receptor interaction in the database is defined by ligand contact elements and receptor contact elements of the ligand-receptor interaction [Tong, claim 7]. Tong discloses the representation for each ligand-receptor interaction is in the form LIS: TP-RIS-BA wherein LIS represents the ligand contact elements of the interaction, TP represents the chemical bonds involved in the interaction, RIS represents the receptor contact elements of the interaction and BA represents the strength of the interaction [Tong, claim 13], as in claims 1, 15, and 21 generating a plurality of drug-target interaction (DTI) features for proteins and ligands in a DTI database, using the plurality of ghost ligands.
Tong discloses “In the embodiments of the present invention, the predictive model is trained using derived input data (or representations) characterizing instances of ligand-receptor interactions with known 3D structures or with theoretical models. In other words, the embodiments of the present invention facilitate the use of machine-learning on 3-D structures or theoretical models for prediction of binding activities between ligands and receptors. Furthermore, the predictive model is trained using non-linear statistical means such as probabilistic function, ANN, HMM, SVM, multiple regression or Bayesian network.” [disclosure page 6, right col para. 0080]. Tong discloses wherein the predictive model is trained according to the following Sub-steps: forming a representation for each ligand-receptor interaction in the database, the representation describing the characteristics of the ligand-receptor interaction training the predictive model using the representations of the ligand-receptor interactions in the database [Tong, claim 10]. Tong teaches wherein the representation for each ligand-receptor interaction is in the form LIS: TP-RIS-BA wherein LIS represents the ligand contact elements of the interaction, TP represents the chemical bonds involved in the interaction, RIS represents the receptor contact elements of the interaction and BA represents the strength of the interaction [Tong, claim 13]. Tong discloses predicting the level of interaction between the at least one specified ligand and the receptor comprises the sub-steps of forming a representation for the potential interaction between the at least one specified ligand and the receptor, the representation for the potential interaction being in a same format as the representation of each ligand receptor interaction in the database, and presenting the representation for the potential interaction to the predictive model [Tong, claim 17], as in claim generating a neural network using the plurality of DTI features. Here, it is obvious the sequence of steps of Tong teaches generating a neural network using ligand-receptor interactions (i.e., drug-target interactions (DTI)).
Tong discloses a method for predicting the interaction between at least one specified ligand and a receptor comprising presenting the specified ligand to a predictive model generated for the receptor [Tong, claim 2], as in claims 1, 15, and 21 predicting a likelihood of interaction for a combination of a query protein and a query ligand using the neural network. Here, it is obvious the predictive model of Tong is generating a quantitative measure/statistical interaction (i.e., affinity) between a ligand and receptor (i.e., protein).
Tong does not teach ghost ligands of claims 1, 15, and 21. Tong does not teach for a cluster of proteins, selected from the plurality of proteins: performing a three-dimensional structure alignment for the proteins in the cluster of proteins of claims 1, 15, and 21. Tong does not teach obtaining the plurality of ghost ligands by projecting a ligand of one of the proteins in the cluster onto other proteins in the cluster, after three-dimensional structure alignment of claims 1, 15, and 21. Tong does not teach obtaining, for each of the plurality of ghost ligands, a confidence score of claims 1, 15, and 21. Tong does not teach claims 2, 4, 9, 12, 22-23, and 25.
Zhou et al. (Zhou) disclose identifying a set of aligned three-dimensional structures, said set of aligned three-dimensional structures comprising positional information for a plurality of residues comprising a reference polypeptide, a plurality of residues comprising a target polypeptide, and a plurality of residues comprising a near-neighbor polypeptide; generating from said set of aligned three-dimensional structures a one-to-one set of corresponding residues, wherein said set of corresponding residues comprises residues from said target polypeptide, and residues from said near-neighbor polypeptide whose positions differ by less than a pre-determined distance from positions of residues in said reference polypeptide that comprise said predetermined three-dimensional structure [Zhou, claim 1]. Zhou discloses wherein said representation is a three-dimensional representation of said target [Zhou, claim 8]. Zhou discloses wherein said representation is a representation of an aligned structure [Zhou claim 9], as in claims 1, 15, and 21 for a cluster of proteins, selected from the plurality of proteins, performing a three-dimensional structure alignment for the proteins in the cluster of proteins. Here, it is obvious that the ghost ligands will be projected onto the proteins (i.e., receptors) after alignment because aligning proteins is a pre-analysis step for constructing and/or modeling 3D representation of a protein for testing ligand-receptor interactions. It is further obvious that the set of aligned 3D structures encompass clusters of proteins.
Dependent claims 12 and 22-23
Zhou discloses said set of aligned three-dimensional structures a one-to-one set of corresponding residues, wherein said set of corresponding residues comprises residues from said target polypeptide, and residues from said near-neighbor polypeptide whose positions differ by less than a pre-determined distance from positions of residues in said reference polypeptide that comprise said predetermined three-dimensional structure [Zhou, claim 1]. Zhou discloses wherein said representation is a three-dimensional representation of said target [Zhou, claim 8]. Zhou discloses wherein said representation is a representation of an aligned structure [Zhou claim 9], as in claim 12.
Zhou discloses “Using the present invention, a feature was predicted in the WNV envelope glycoprotein defined by a cluster of residues including S306, K307, T330, and T332 [disclosure page 8 left col para 0097]. Zhou discloses generating from said set of aligned three-dimensional structures a one-to-one set of corresponding residues, wherein said set of corresponding residues comprises residues from said target polypeptide, and residues from said near-neighbor polypeptide whose positions differ by less than a pre-determined distance from positions of residues in said reference polypeptide that comprise said predetermined three-dimensional structure; generating a plurality of conservation scores for corresponding reference and target residues comprising said one-to-one set, wherein said conservation scores are based on a first similarity metric [Zhou, claim 1]. Zhou discloses wherein said representation is a three-dimensional representation of said target [Zhou, claim 8]. Zhou discloses wherein said representation is a representation of an aligned structure [Zhou claim 9], as in claim 22.
Zhou discloses using BLAST [disclosure, page 1 left col para. 0006]. Zhou discloses “The proteins have been aligned using Local Global Alignment (LGA) with a reference (ricin) polypeptide. Target structures (Ricin A PDB structures) are represented by the top 3 bars. Near neighbor structures (ricin-like structures) [disclosure page 2 left col para 0019], as in claim 23. Here, it is obvious that a user creates protein clusters using alignment tools or clustering tools because a user is required to input sequence data into system (i.e., BLAST) in order to align sequence data.
Fan et al. (Fan) discloses dividing the ligand into two or more sections, anchoring a first section of the two or more ligand sections to a location of the protein, adding at least one subsequent section of the two or more ligand sections to the first section to form a growing ligand, and continuing adding subsequent sections to the growing ligand until the ligand is complete [Fan, claim 2]. Fan also discloses using a neural network [Fan, claim 1], as in claims 1, 15, and 21 obtaining the plurality of ghost ligands by projecting a ligand of one of the proteins in the cluster onto other proteins in the cluster. Here, because Fan discloses a step for projecting ligand onto a protein, Fan teaches synthetic data that can be used to make predictions as described in the specification [page 6 para 0030].
Fan discloses determining, using a neural network, scores associated with the poses [Fan, claim 1]. Fan teaches reranked scores of the poses [Fan, claim 24]. Fan discloses the lower reranked score is associated with a higher probability that the removed section of a pharmacophore [Fan, claim 25]. Fan teaches ranking the re-rank score and the second re-ranked score to determine a likelihood that the first removed section is a pharmacophore compared to the second removed section [Fan, claim 26]. Fan discloses “After the feature maps are obtained, they may be transformed into a feasibility score of the likelihood of the ligand binding to the pocket. The resulted score function is named the reranking score.” [disclosure page 8 left col para 0114], as in claim obtaining, for each of the plurality of ghost ligands, a confidence score.
Dependent claim 4
Fan discloses determining, using a neural network, scores associated with the poses [Fan, claim 1]. Fan teaches reranked scores of the poses [Fan, claim 24]. Fan discloses the lower reranked score is associated with a higher probability that the removed section of a pharmacophore [Fan, claim 25]. Fan teaches ranking the re-rank score and the second re-ranked score to determine a likelihood that the first removed section is a pharmacophore compared to the second removed section [Fan, claim 26]. Fan discloses “After the feature maps are obtained, they may be transformed into a feasibility score of the likelihood of the ligand binding to the pocket. The resulted score function is named the reranking score.” [disclosure page 8 left col para 0114], as in claim 4.
Chiang et al. (Chiang) teach “As discussed in our previous work, the topology setting, i.e., how the ghost ligand is implemented in the software, differs in NAMD and VSS. Both NAMD and VSS adopt the so-called dual topology setting, which dictates that the ghost ligand (e.g., anisole) and the original ligand (benzene) coexist during the calculation, but do not interact with each other. While in VSS this is done through keeping two complete ligands (anisole and benzene) simultaneously in memory with anisole copying the geometry of benzene, in NAMD this is done through keeping only the perturbed atoms that differ from another in memory. That is, in NAMD, the substitution –OCH3 is artificially linked to the benzene, and the perturbed atoms are –OCH3 and benzene carbon that is connected to it. The remaining part of the benzene (unperturbed atoms) is shared between the ghost ligand anisole and the original ligand benzene.” [page 234109-8 Appendix]. Chiang teaches “The four tested benzene derivatives are anisole (–OCH3), methylaniline (–NHCH3), ethylbenzene (–CH2CH3), and benzyl alcohol (–CH3OH). They are of similar physical size and each differs from benzene only by one functional group, i.e., one substitution.” [page 234109-4 right col second para].
Obvious claims 2 and 25
With respect to claim 2, the claim is rendered obvious because Zhou discloses generating two homology models of this protein that were employed as MR targets [Disclosure page 6 left col para. 0074 last sentence]. Zhou discloses wherein said representation is a three-dimensional representation of said target [Zhou, claim 8]. Zhou teaches wherein said representation is a representation of an aligned structure [Zhou claim 9]. Fan discloses dividing the ligand into two or more sections, anchoring a first section of the two or more ligand sections to a location of the protein, adding at least one subsequent section of the two or more ligand sections to the first section to form a growing ligand, and continuing adding subsequent sections to the growing ligand until the ligand is complete [Fan, claim 2].
With respect to claim 25, the claim is rendered obvious because it an inherent that utilizing neural networks or other deep learning systems to ignore statistical outliers and/or utilize normalization and optimization methods for denoising data such that noisy is ignored or filtered.
It would have been obvious for a person having ordinary skill in the art by the effective filing date of the claim invention to modify Tong in view of Zhou because Zhou discloses methods for scoring a set of residues that form a predetermined three-dimensional structure is a polypeptide that identifies aligned 3D structures that corresponds to a target, reference, or a near-neighbor polypeptide [Zhou, claim 1]. One of ordinary skill in the art would be motivated to combine Tong in view of Zhou because the alignment methods of Zhou can be utilized to align clusters of proteins for further projection with ligand structures and ligand data and, also, utilized for predicting ligand-protein (i.e., drug-target) interactions. Here, there is a reasonable expectation of success that combining the aligned 3D structure data of Zhou with the ligand generation of Tong would produce a 3D protein model (i.e., drug receptor) that can be used to quantify ligand-target interactions. As such, the combination of Tong and Fan would yield a predictable method that can be utilized for generating a plurality of ligands and aligned 3D protein structures for predicting a likelihood of an interaction between a ligand and a protein with respect to predicting drug-target interactions.
It would have been obvious for a person having ordinary skill in the art by the effective filing date of the claim invention to modify Tong in view of Zhou in view of Fan because Fan discloses methods for classifying and predicting ligand docking conformations [title]. Here, one of ordinary skill in the art would recognize that Fan discloses method for determining poses of a ligands that encompass anchoring sections of determined posed onto specific and different to locations of a protein [Fan, claims 1 and 2]. One of ordinary skill in the art would be motivated to combine Tong in view of Zhou with the anchoring methods of Fan for evaluating ligands against associated proteins targets for predicting a ligand’s ability to dock into a protein. Thus, there is a reasonable expectation that combining the ligand generation of Tong and the aligned clustered 3D proteins structures of Zhou with the anchoring of ligands to a protein methods of Fan would yield a representation comprising a ligand, protein, and associated ligand-protein data (i.e., contact elements, chemical bonds, receptor contact elements) that can be further utilized for predicting likelihood of interactions between a ligand and protein for predicting drug-target binding.
It would have been obvious for a person having ordinary skill in the art by the effective filing date of the claim invention to modify Tong in view of Zhou in view of Fan in view of Chiang because Chiang teaches methods for testing the benzene derivatives anisole, methylaniline, ethylbenzene, and benzyl alcohol because they are similar in physical size to benzene and each differs from benzene only by one functional group, i.e., one substitution [page 234109-4 right col second para] and using anisole as a ghost ligand [page 234109-8 Appendix]. One of ordinary skill in the art would recognize that although Chiang does not teach explicit drug-target interaction Chiang teaches different derivatives, conformations, and added chemical moieties can be utilize as “ghost ligand”. Therefore, as taught by Chiang the additional ligands of Tong in conjunction with the different ligand poses of Fan teach ghost ligands because the ligands do not interact with each other, but are account for. One of ordinary skill in the art would be motivated to combine Tong in view of Zhou in view of Fan in view of Chiang because Chiang teaches evaluating the potentials of ghost ligands and that ghost ligands can turn into real ligands that interact with the environment [page 234109-5 left col last sentence]. Therefore, there is a reasonable expectation of success that combining Tong in view of Zhou in view of Fan in view of Chiang would yield a predictable method for turning ghost ligands into real ligands on evaluation of drug-target binding interactions between the ghost ligands (i.e., anisole, drug, small molecule) and their associated proteins (i.e., drug receptors).
Claim(s) 3 and 24 are rejected under 35 U.S.C. 103 as being unpatentable over Tong in view of Zhou in view of Fan in view of Chiang, as applied to claims 1-2, 4, 12, 15, 21-23, and 25 above, and in further view of Shah et al. (US Patent Pub: US 2018/0312999, Patent Pub Date: 01 Nov 2018).
Tong in view of Zhou in view of Fan in view of Chiang teach 1-2, 4, 12, 15, 21-23, and 25.
Tong in view of Zhou in view of Fan in view of Chiang teach a method for predicting ligand-target (i.e., drug-target) binding that generates ghost ligands, ghost ligand drug-target interactions (DTI), and generating a neural network using DTI features for predicting a likelihood of interaction between a ligand and protein.
Tong in view of Zhou in view of Fan in view of Chiang do not teach claims 3 and 24.
With respect to claim 3, the claim is rendered obvious because Shah et al. (Shah) discloses computing similarity metric between the pair of snapshots with the snapshots including differing amino acids [Shah, claim 1 step (d) sub-step (i)]. Shah discloses “Related proteins include polypeptide members of a particular gene family, polypeptides having topologically similar binding sites, or polypeptides having at least 10% sequence identity within a domain of interest.” Shah discloses “if a protein exhibits sufficient relatedness for super-positioning based molecular modeling include, but are not limited to, sequence homology, three-dimensional relatedness ( e.g. similarity of molecular folds, or protein domains) as a function of similarity in the three dimensional configuration, order of secondary structures, or topological connections. Databases useful for assessing similarities of three-dimensional relatedness, include, but are not limited to, the Structural Comparison of Proteins (SCOP) and PROSITE.”[disclosure page 5 para 0026]. Zhou discloses “Using the present invention, a feature was predicted in the WNV envelope glycoprotein defined by a cluster of residues including S306, K307, T330, and T332 [disclosure page 8 left col para 0097] which makes obvious using existing clusters from a database.
With respect to claim 24, the claim is rendered obvious because Shah discloses computing similarity metric between the pair of snapshots with the snapshots including differing amino acids [Shah, claim 1 step (d) sub-step (i)]. Shah discloses “Related proteins include polypeptide members of a particular gene family, polypeptides having topologically similar binding sites, or polypeptides having at least 10% sequence identity within a domain of interest.” [disclosure page 5 para 0026]. Zhou discloses comprising identifying said target polypeptide using a sequence similarity comparison, a structural similarity comparison, or a taxonomic comparison to said reference polypeptide [Zhou, claim 18]. It is further obvious that a user can utilized software to create user created protein cluster from databases such as BLAST or LGA (local-global alignment) software for aligning proteins and protein clusters.
It would be obvious to one of ordinary skill in the art by the effective filing date of the claimed invention to modify Tong in view of Zhou in view of Fan in view of Chiang, and in further view of Shah because Shah discloses methods for in silico drug discovery [title] that evaluates 3D structures of a target protein and clustering snapshots based on similarity metrics [Shah, claim 1 step (d) sub step (ii))], discloses methods for selecting chemical probes (i.e., ligands), and identifying protein conformation to which the chemical probes bind [Shah, claim 2]. One of ordinary skill in the art would recognize that Tong, Zhou, Fan, Chiang, and Shah evaluate ligands and proteins for determining interactions that can be associated with drug-target interactions. Here, one of ordinary skill in the art would be motivated to combine Tong in view of Zhou in view of Fan in view of Chiang, and in further view of Shah because Shah discloses the implementation of topological similarities that can be utilized to determine potential binding pockets that could turn ghost ligands into real ligands. Thus, there is a reasonable expectation of success that combining Tong in view of Zhou in view of Fan in view of Chiang, and in further view of Shah would yield a predictable method for predicting drug-target binding using ghost ligand-target binding interactions.
Claim(s) 5 and 7 are rejected under 35 U.S.C. 103 as being unpatentable over Tong in view of Zhou in view of Fan in view of Chiang, as applied to claims 1-2, 4, 12, 15, 21-23, and 25 above, and in further view of Maggiore et al. (Journal of medicinal chemistry, 2014-04, Vol.57 (8), p.3186-3204) (Cited in the Office Action mailed 20 May 2025).
Tong in view of Zhou in view of Fan in view of Chiang teach 1-2, 4, 12, 15, 21-23, and 25.
Tong in view of Zhou in view of Fan in view of Chiang teach a method for predicting ligand-target (i.e., drug-target) binding that generates ghost ligands, ghost ligand drug-target interactions (DTI), and generating a neural network using DTI features for predicting a likelihood of interaction between a ligand and protein.
Tong in view of Zhou in view of Fan in view of Chiang do not teach claims 5 and 7.
Maggiore et al. (Maggiore) teach using the ZINC database [page 3194 figure 5].
Maggiore teaches comparing similarity coefficients between two thrombin inhibitors using Dice, Tanimoto, and Tversky coefficient and further comparing using MACCS and ECFP4 [page 3195 fig 7], as in claim 5 for each of a plurality of combinations of a ligand and a protein in the DTI database by selecting, from the plurality of ghost ligands, a ghost ligand that is most similar to the ligand considered for the combination.
Maggiore teaches using ChEMBL target identifiers for different classes of proteins such as thrombin, beta-2 adrenergic receptor, and dopamine D2 receptor, for example [page 3198 fig 10], as in claim 5 generating features for the protein considered for the combination, wherein the generated features characterize the protein considered for the combination.
With respect to claim 7, the claim is rendered obvious because Maggiore teaches using Soergel distance [page 3192]. Rayhan teaches torsion angles auto-covariance using distance factor (DF). Here, the combination of Maggiore and Rayhan would teach selecting the mot similar ghost ligands using distance metric.
It would be obvious to one of ordinary skill in the art by the effective filing date of the claimed invention to modify Tong in view of Zhou in view of Fan in view of Chiang, and in further view of Maggiore because Maggiore teaches molecular similarity in ligands for proteins such as thrombin, beta-2 adrenergic receptor, and dopamine D2 receptor, for example [page 3198 fig 10]. One of ordinary skill in the art would be motivated to combine Tong in view of Zhou in view of Fan in view of Chiang, and in further view of Maggiore because Maggiore teaches testing cyclooxygenase inhibitors and comparing similarities between the compound (i.e., ligand or drugs) [page 3189 fig 3]. There is a reasonable expectation of success combining Tong in view of Zhou in view of Fan in view of Chiang, and in further view of Maggiore because Maggiore teaches drug features such as ligand similarity. Here, substituting the similarity comparisons of Maggiore into the methods of Tong in view of Zhou in view of Fan in view of Chiang would yield a predictable method that can analyze and compare drug similarities for generating a drug-target interaction between compounds (i.e., drugs or ligands) their target (i.e., protein).
Claim(s) 10 are rejected under 35 U.S.C. 103 as being unpatentable over Tong in view of Zhou in view of Fan in view of Chiang, as applied to claims 1-2, 4, 12, 15, 21-23, and 25 above, and in further view of Shi et al. (Genomics (San Diego, Calif.), 2019-12, Vol.111 (6), p.1839-185) (Cited in the Office Action mailed 20 May 2025).
Tong in view of Zhou in view of Fan in view of Chiang teach 1-2, 4, 12, 15, 21-23, and 25.
Tong in view of Zhou in view of Fan in view of Chiang teach a method for predicting ligand-target (i.e., drug-target) binding that generates ghost ligands, ghost ligand drug-target interactions (DTI), and generating a neural network using DTI features for predicting a likelihood of interaction between a ligand and protein.
Tong in view of Zhou in view of Fan in view of Chiang does not teach claim 10.
Shi et al. (Shi) teach using chemical structure and amino acid sequences for generating positive samples [page 1844 fig 1], as in claim 10 obtaining positive training samples based on the plurality of DTI features for the proteins and the ligands. Shi teaches feature selection using known and unknown interactions and processing the interaction that are processed using Lasso, PCA, ReliefF, and Elastic net which is further processes using SMOTE method to produce positive and negative samples [page 1844 fig 1]. Shi teaches Synthetic Minority Oversampling Technique (SMOTE) is a method for randomly under sampling a large scale of samples while randomly oversampling a small scale of samples [page 1843 left col top para.]. Shi teaches SMOTE processes random oversampling on a small scale of positive samples and random under sampling on a large scale of negative samples so that the positive and negative samples are balanced [page 1842], as in claim 10 obtaining negative training samples based on the plurality of DTI features by shuffling, at least once, the plurality of DTI features for the proteins and the ligands. Here, the random undersampling and oversampling teaches the shuffling limitations. Shi teaches using positive and negative sample for prediction of drug-target prediction are [page 1844 fig 1]. Shi teaches using random forest machine learning [page 1843 section 2,8 random forest], as in claim 10 using the positive and the negative training samples, training the machine learning model for DTI prediction.
It would be obvious to one of ordinary skill in the art by the effective filing date of the claimed invention to modify Tong in view of Zhou in view of Fan in view of Chiang, and in further view of Shi because Shi teaches analyzing drug-target interactions using Lasso with random forest based on evolutionary information and chemical structure [title]. One of ordinary skill would be motivated to combine Tong in view of Zhou in view of Fan in view of Chiang, and in further view of Shi because Shi teaches using chemical data (i.e., ligand/drug) and amino acid data (i.e., target or protein), and combining and separating the interaction data into positive and negative datasets for predicting drug-target interactions (DTI). There is a reasonable expectation of success combining the methods of Tong in view of Zhou in view of Fan in view of Chiang, and in further view of Shi to construct a DTI system that can predict DTI interaction because Shi teaches methods using random forest machine learning for predicting DTI using positive and negative datasets. Here, substituting the methods of Shi into the methods of Tong, Zhou, Fan, and Chiang would yield a predictable system for predicting drug-target binding that uses positive and negative datasets.
Claim(s) 13-14 are rejected under 35 U.S.C. 103 as being unpatentable over Tong in view of Zhou in view of Fan in view of Chiang, as applied to claims 1-2, 4, 12, 15, 21-23, and 25 above, and in further view of Da et al. (Journal of chemical information and modeling, 2014-09, Vol.54 (9), p.2555-2561).
Tong in view of Zhou in view of Fan in view of Chiang teach 1-2, 4, 12, 15, 21-23, and 25.
Tong in view of Zhou in view of Fan in view of Chiang teach a method for predicting ligand-target (i.e., drug-target) binding that generates ghost ligands, ghost ligand drug-target interactions (DTI), and generating a neural network using DTI features for predicting a likelihood of interaction between a ligand and protein.
Tong in view of Zhou in view of Fan in view of Chiang does not teach claims 13-14.
With respect to claim 13, the claim is render obvious because Zhou discloses “LGA scoring function has two components, LCS (longest continuous segments) and GDT (global distance test), established for the detection of regions of local and global structure similarities between proteins. In comparing two protein structures, the LCS procedure is able to localize and superimpose the longest segments of residues that can fit under a selected RMSD cutoff.” [disclosure page 5 right col 0066]. Zhou discloses further comprising combining said plurality of composite scores with a plurality of scores with a plurality of scores indicative of frequency of the reference polypeptide residue within local sequence context occurs in a data set of polypeptide sequences [Zhou, claim 12]. Zhou teach generating from said set of aligned three-dimensional structures a one-to-one set of corresponding residues, wherein said set of corresponding residues comprises residues from said target polypeptide, and residues from said near-neighbor polypeptide whose positions differ by less than a pre-determined distance from positions of residues in said reference polypeptide that comprise said predetermined three-dimensional structure [Zhou, claim 1], as in claim 13 obtaining, for the query protein, a possible binding site and associated local features based on the plurality of ghost ligands generating features for the query protein, the features for the query protein comprising the local features. Da et al. (Da) teaches protein-ligand interactions fingerprints descriptors was implemented with MOE suite [page 2557 left col PLIF Based Similarity]. Da teaches “The PLIF descriptors for all protein-bound ligands were generated with the default parameter set in MOE. The PLIF similarity was expressed by means of the Tanimoto similarity coefficient.” [page 2557 right col top para], as in claim 13 generating features steps. Tong discloses a method for predicting the interaction between at least one specified ligand and a receptor comprising presenting the specified ligand to a predictive model generated for the receptor [Tong, claim 2]. Here, it is obvious the predictive model of Tong is generating a quantitative measure/statistical interaction (i.e., affinity) between a ligand and receptor (i.e., protein).
With respect to claim 14, the claim is rendered obvious because Zhou discloses “The LGA scoring function has two components, LCS (longest continuous segments) and GDT (global distance test), established for the detection of regions of local and global structure similarities between proteins. In comparing two protein structures, the LCS procedure is able to localize and superimpose the longest segments of residues that can fit under a selected RMSD cutoff.” [disclosure, page 5 para 0066]. Zhou discloses method of claim 1, further comprising identifying said nearest-neighbor polypeptide using a sequence similarity comparison, a structural similarity comparison, or a taxonomic comparison to said reference polypeptide [Zhou, claim 19]. Here, Zhou discloses protein features than can encompass global features and functional annotations.
It would be obvious to one of ordinary skill in the art by the effective filing date of the claimed invention to modify Tong in view of Zhou in view of Fan in view of Chiang, and in further view of Da because Da teaches evaluating structural protein-ligand interfinger prints (SPLIF) for structure-based virtual screening [title]. One of ordinary skill in the art would be motivated to combine Tong in view of Zhou in view of Fan in view of Chiang, and in further view of Da because Da teaches generating descriptors, such as Tanimoto similarity coefficients generated in MOE [page 2557 right col top para] with respect to providing ligand fingerprint and ligand descriptors Here, there is a reasonable expectation of success that combining Tong in view of Zhou in view of Fan in view of Chiang, and in further view of Da because Da teaches evaluating protein-ligand interaction using fingerprinting and descriptors methods. As such, combining Tong in view of Zhou in view of Fan in view of Chiang, and in further view of Da would yield predictable method steps using ligand fingerprints and descriptors in conjunction to a neural network for predicting likelihood of interaction between a ligand and protein for predicting drug-target binding.
Response to Arguments
Applicant’s arguments, filed 20 August 2025, have been fully considered and the rejection is maintained. However, upon further consideration, a new ground(s) of rejection is made in view of amendments received 20 August 2025. Here, Tong, Zhou, Fan, Chiang, Shah, Maggiore, Shi, and Da address the amendments received 20 August 2025.
With respect to claims 1, 15, and 21 generating a neural network step, it is recommended to amend the “generating a neural network using the plurality of DTI features” step to recite “training a neural network using the plurality of DTI features”.
Conclusion
Claims 1-15 and 21-25 are rejected.
No claims are allowed.
Finality
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 JOSEPH C PULLIAM whose telephone number is (571)272-8696. The examiner can normally be reached 0730-1700 M-F.
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.
/J.C.P./
Examiner, Art Unit 1687
/Karlheinz R. Skowronek/Supervisory Patent Examiner, Art Unit 1687