# US 12,224,045 B2 · Systems and methods for prediction of protein formulation properties > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,224,045 B2 - Document ID: US12224045B2 - Status: Granted - Kind: B2 - Application number: 17/638,395 - Filed: 2020-08-25 - Published: 2022-09-15 - Granted: 2025-02-11 - Earliest claimed: 2019-08-26 - Assignees: Amgen Inc. - Applicants: Amgen Inc. - Inventors: Cindy Ren; Mohan B. Boggara; Nitin Rathore; Behnam Partopour - Web page: https://patentgrep.com/patent/US12224045B2 ## CPC - G16B 15/20: Protein or domain folding - G16C 20/30: Prediction of properties of chemical compounds, compositions or mixtures - G16C 20/50: Molecular design, e.g. of drugs - G16C 20/70: Machine learning, data mining or chemometrics - G16B: BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY - G16C: COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE ## Abstract In a method for predicting a property of potential protein formulations, a set of formulation descriptors is classified as belonging to a specific one of a plurality of predetermined groups that each correspond to a different value range for a protein formulation property. Classifying the set of descriptors includes applying at least a first portion of the set of descriptors as inputs to a first machine learning model. The method also includes selecting, based on the classification, a second machine learning model from among multiple models corresponding to different groups. The method also includes predicting a value of the protein formulation property that corresponds to the set of descriptors, by applying at least a second portion of the set of formulation descriptors as inputs to the selected model. The method further includes causing the value of the protein formulation property to be displayed to a user and/or stored in a memory. ## Claims ### Claim 1 A method of predicting a property of potential protein formulations, the method comprising: classifying, by one or more processors of a computing system, a set of formulation descriptors as belonging to a specific group of a plurality of predetermined groups that each correspond to a different value range for a protein formulation property, wherein classifying the set of formulation descriptors includes applying at least a first portion of the set of formulation descriptors as inputs to a first machine learning model; selecting, by the one or more processors and based on the specific group, a second machine learning model from among a plurality of machine learning models that each correspond to a different one of the plurality of predetermined groups; predicting, by the one or more processors, a value of the protein formulation property by applying at least a second portion of the set of formulation descriptors as inputs to the second machine learning model; and causing, by the one or more processors, the value of the protein formulation property to be one or both of (i) displayed to a user, and (ii) stored in a memory. ### Claim 2 (depends on 1) The method of claim 1, wherein the protein formulation property is viscosity, and wherein each group of the plurality of predetermined groups corresponds to a different range of viscosity values. ### Claim 3 (depends on 1) The method of claim 1, wherein the set of formulation descriptors includes (i) one or more molecule descriptors, (ii) one or more excipient descriptors, and (iii) one or more solution descriptors. ### Claim 4 (depends on 1) The method of claim 1, wherein the set of formulation descriptors includes one or more numerical values. ### Claim 5 (depends on 1) The method of claim 1, wherein the set of formulation descriptors includes one or more categories. ### Claim 6 (depends on 1) The method of claim 1, wherein the set of formulation descriptors includes one or more of (i) a molecule identifier, (ii) an antibody type, (iii) a target pH level, and (iv) a target protein concentration. ### Claim 7 (depends on 1) The method of claim 1, wherein the set of formulation descriptors includes one or more of (i) molecular weight of an excipient, (ii) charge of the excipient, (iii) solubility of the excipient, (iv) bonds of the excipient, and (v) a type of the excipient. ### Claim 8 (depends on 1) The method of claim 1, wherein the plurality of predetermined groups consists of three groups. ### Claim 9 (depends on 1) The method of claim 1, wherein each of the plurality of machine learning models includes a regression and/or classification model. ### Claim 10 (depends on 9) The method of claim 9, wherein at least one of the plurality of machine learning models is a random forest model. ### Claim 11 (depends on 1) The method of claim 1, wherein: classifying the set of formulation descriptors includes applying the set of formulation descriptors as inputs to the first machine learning model; and predicting the value of the protein formulation property includes applying the set of formulation descriptors as inputs to the second machine learning model. ### Claim 12 (depends on 1) The method of claim 1, further comprising: receiving, by the one or more processors and via a first graphical user interface, the set of formulation descriptors from a user. ### Claim 13 (depends on 12) The method of claim 12, further comprising: determining, by the one or more processors and based in part on the predicted value of the protein formulation property, a rank-ordering of a plurality of candidate protein formulations; and causing, by the one or more processors, an indication of the rank-ordering to be displayed via the first graphical user interface or a second graphical user interface. ### Claim 14 A computing system comprising: one or more processors; and one or more memories storing instructions that, when executed by the one or more processors, cause the computing system to classify a set of formulation descriptors as belonging to a specific group of a plurality of predetermined groups that each correspond to a different value range for a protein formulation property, wherein classifying the set of formulation descriptors includes applying at least a first portion of the set of formulation descriptors as inputs to a first machine learning model, select, based on the specific group, a second machine learning model from among a plurality of machine learning models that each correspond to a different one of the plurality of predetermined groups, predict a value of the protein formulation property by applying at least a second portion of the set of formulation descriptors as inputs to the second machine learning model, and cause the value of the protein formulation property to be one or both of (i) displayed to a user, and (ii) stored in a memory. ### Claim 15 (depends on 14) The computing system of claim 14, wherein the protein formulation property is viscosity, and wherein each group of the plurality of predetermined groups corresponds to a different range of viscosity values. ### Claim 16 (depends on 14) The computing system of claim 14, wherein the set of formulation descriptors includes (i) one or more molecule descriptors, (ii) one or more excipient descriptors, and (iii) one or more solution descriptors. ### Claim 17 (depends on 14) The computing system of claim 14, wherein each of the plurality of machine learning models is a regression and/or classification model. ### Claim 18 (depends on 17) The computing system of claim 17, wherein at least one of the plurality of machine learning models is a random forest model. ### Claim 19 A non-transitory computer-readable medium storing instructions that, when executed by one or more processors of a computing system, cause the computing system to: classify a set of formulation descriptors as belonging to a specific group of a plurality of predetermined groups that each correspond to a different value range for a protein formulation property, wherein classifying the set of formulation descriptors includes applying at least a first portion of the set of formulation descriptors as inputs to a first machine learning model; select, based on the specific group, a second machine learning model from among a plurality of machine learning models that each correspond to a different one of the plurality of predetermined groups; predict a value of the protein formulation property by applying at least a second portion of the set of formulation descriptors as inputs to the second machine learning model; and cause the value of the protein formulation property to be one or both of (i) displayed to a user, and (ii) stored in a memory. ### Claim 20 (depends on 19) The non-transitory computer-readable medium of claim 19, wherein: the protein formulation property is viscosity; each group of the plurality of predetermined groups corresponds to a different range of viscosity values; and the set of formulation descriptors includes (i) one or more molecule descriptors, (ii) one or more excipient descriptors, and (iii) one or more solution descriptors. ### Claim 21 (depends on 19) The non-transitory computer-readable medium of claim 19, wherein: each of the plurality of machine learning models is a regression and/or classification model; and at least one of the plurality of machine learning models is a random forest model.