Reduction of monoclonal antibody viscosity using interpretable machine learning
Early identification of antibody candidates with drug-like properties is essential for simplifying the development of safe and effective antibody therapeutics. For subcutaneous administration, it is important to identify candidates with low self-association to enable their formulation at high concen...
Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Taylor & Francis Group
2024-12-01
|
| Series: | mAbs |
| Subjects: | |
| Online Access: | https://www.tandfonline.com/doi/10.1080/19420862.2024.2303781 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832576536939593728 |
|---|---|
| author | Emily K. Makowski Hsin-Ting Chen Tiexin Wang Lina Wu Jie Huang Marissa Mock Patrick Underhill Emma Pelegri-O’Day Erick Maglalang Dwight Winters Peter M. Tessier |
| author_facet | Emily K. Makowski Hsin-Ting Chen Tiexin Wang Lina Wu Jie Huang Marissa Mock Patrick Underhill Emma Pelegri-O’Day Erick Maglalang Dwight Winters Peter M. Tessier |
| author_sort | Emily K. Makowski |
| collection | DOAJ |
| description | Early identification of antibody candidates with drug-like properties is essential for simplifying the development of safe and effective antibody therapeutics. For subcutaneous administration, it is important to identify candidates with low self-association to enable their formulation at high concentration while maintaining low viscosity, opalescence, and aggregation. Here, we report an interpretable machine learning model for predicting antibody (IgG1) variants with low viscosity using only the sequences of their variable (Fv) regions. Our model was trained on antibody viscosity data (>100 mg/mL mAb concentration) obtained at a common formulation pH (pH 5.2), and it identifies three key Fv features of antibodies linked to viscosity, namely their isoelectric points, hydrophobic patch sizes, and numbers of negatively charged patches. Of the three features, most predicted antibodies at risk for high viscosity, including antibodies with diverse antibody germlines in our study (79 mAbs) as well as clinical-stage IgG1s (94 mAbs), are those with low Fv isoelectric points (Fv pIs < 6.3). Our model identifies viscous antibodies with relatively high accuracy not only in our training and test sets, but also for previously reported data. Importantly, we show that the interpretable nature of the model enables the design of mutations that significantly reduce antibody viscosity, which we confirmed experimentally. We expect that this approach can be readily integrated into the drug development process to reduce the need for experimental viscosity screening and improve the identification of antibody candidates with drug-like properties. |
| format | Article |
| id | doaj-art-36271266403a4691a2dd72e35b369f74 |
| institution | Kabale University |
| issn | 1942-0862 1942-0870 |
| language | English |
| publishDate | 2024-12-01 |
| publisher | Taylor & Francis Group |
| record_format | Article |
| series | mAbs |
| spelling | doaj-art-36271266403a4691a2dd72e35b369f742025-01-31T04:19:38ZengTaylor & Francis GroupmAbs1942-08621942-08702024-12-0116110.1080/19420862.2024.2303781Reduction of monoclonal antibody viscosity using interpretable machine learningEmily K. Makowski0Hsin-Ting Chen1Tiexin Wang2Lina Wu3Jie Huang4Marissa Mock5Patrick Underhill6Emma Pelegri-O’Day7Erick Maglalang8Dwight Winters9Peter M. Tessier10Department of Pharmaceutical Sciences, University of Michigan, Ann Arbor, MI, USABiointerfaces Institute, University of Michigan, Ann Arbor, MI, USABiointerfaces Institute, University of Michigan, Ann Arbor, MI, USABiointerfaces Institute, University of Michigan, Ann Arbor, MI, USADepartment of Pharmaceutical Sciences, University of Michigan, Ann Arbor, MI, USATherapeutic Discovery, Research, Amgen Inc, Thousand Oaks, CA, USADepartment of Chemical and Biological Engineering, Rensselaer Polytechnic Institute, Troy, NY, USATherapeutic Discovery, Research, Amgen Inc, Thousand Oaks, CA, USADrug Product Technologies, Amgen Inc, Thousand Oaks, CA, USATherapeutic Discovery, Research, Amgen Inc, Thousand Oaks, CA, USADepartment of Pharmaceutical Sciences, University of Michigan, Ann Arbor, MI, USAEarly identification of antibody candidates with drug-like properties is essential for simplifying the development of safe and effective antibody therapeutics. For subcutaneous administration, it is important to identify candidates with low self-association to enable their formulation at high concentration while maintaining low viscosity, opalescence, and aggregation. Here, we report an interpretable machine learning model for predicting antibody (IgG1) variants with low viscosity using only the sequences of their variable (Fv) regions. Our model was trained on antibody viscosity data (>100 mg/mL mAb concentration) obtained at a common formulation pH (pH 5.2), and it identifies three key Fv features of antibodies linked to viscosity, namely their isoelectric points, hydrophobic patch sizes, and numbers of negatively charged patches. Of the three features, most predicted antibodies at risk for high viscosity, including antibodies with diverse antibody germlines in our study (79 mAbs) as well as clinical-stage IgG1s (94 mAbs), are those with low Fv isoelectric points (Fv pIs < 6.3). Our model identifies viscous antibodies with relatively high accuracy not only in our training and test sets, but also for previously reported data. Importantly, we show that the interpretable nature of the model enables the design of mutations that significantly reduce antibody viscosity, which we confirmed experimentally. We expect that this approach can be readily integrated into the drug development process to reduce the need for experimental viscosity screening and improve the identification of antibody candidates with drug-like properties.https://www.tandfonline.com/doi/10.1080/19420862.2024.2303781Antibody engineeringchargecomputationdevelopabilityformulationFv |
| spellingShingle | Emily K. Makowski Hsin-Ting Chen Tiexin Wang Lina Wu Jie Huang Marissa Mock Patrick Underhill Emma Pelegri-O’Day Erick Maglalang Dwight Winters Peter M. Tessier Reduction of monoclonal antibody viscosity using interpretable machine learning mAbs Antibody engineering charge computation developability formulation Fv |
| title | Reduction of monoclonal antibody viscosity using interpretable machine learning |
| title_full | Reduction of monoclonal antibody viscosity using interpretable machine learning |
| title_fullStr | Reduction of monoclonal antibody viscosity using interpretable machine learning |
| title_full_unstemmed | Reduction of monoclonal antibody viscosity using interpretable machine learning |
| title_short | Reduction of monoclonal antibody viscosity using interpretable machine learning |
| title_sort | reduction of monoclonal antibody viscosity using interpretable machine learning |
| topic | Antibody engineering charge computation developability formulation Fv |
| url | https://www.tandfonline.com/doi/10.1080/19420862.2024.2303781 |
| work_keys_str_mv | AT emilykmakowski reductionofmonoclonalantibodyviscosityusinginterpretablemachinelearning AT hsintingchen reductionofmonoclonalantibodyviscosityusinginterpretablemachinelearning AT tiexinwang reductionofmonoclonalantibodyviscosityusinginterpretablemachinelearning AT linawu reductionofmonoclonalantibodyviscosityusinginterpretablemachinelearning AT jiehuang reductionofmonoclonalantibodyviscosityusinginterpretablemachinelearning AT marissamock reductionofmonoclonalantibodyviscosityusinginterpretablemachinelearning AT patrickunderhill reductionofmonoclonalantibodyviscosityusinginterpretablemachinelearning AT emmapelegrioday reductionofmonoclonalantibodyviscosityusinginterpretablemachinelearning AT erickmaglalang reductionofmonoclonalantibodyviscosityusinginterpretablemachinelearning AT dwightwinters reductionofmonoclonalantibodyviscosityusinginterpretablemachinelearning AT petermtessier reductionofmonoclonalantibodyviscosityusinginterpretablemachinelearning |