Interpreting biologically informed neural networks for enhanced proteomic biomarker discovery and pathway analysis
Abstract The incorporation of machine learning methods into proteomics workflows improves the identification of disease-relevant biomarkers and biological pathways. However, machine learning models, such as deep neural networks, typically suffer from lack of interpretability. Here, we present a deep...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2023-09-01
|
| Series: | Nature Communications |
| Online Access: | https://doi.org/10.1038/s41467-023-41146-4 |
| _version_ | 1797558380124962816 |
|---|---|
| author | Erik Hartman Aaron M. Scott Christofer Karlsson Tirthankar Mohanty Suvi T. Vaara Adam Linder Lars Malmström Johan Malmström |
| author_facet | Erik Hartman Aaron M. Scott Christofer Karlsson Tirthankar Mohanty Suvi T. Vaara Adam Linder Lars Malmström Johan Malmström |
| author_sort | Erik Hartman |
| collection | DOAJ |
| description | Abstract The incorporation of machine learning methods into proteomics workflows improves the identification of disease-relevant biomarkers and biological pathways. However, machine learning models, such as deep neural networks, typically suffer from lack of interpretability. Here, we present a deep learning approach to combine biological pathway analysis and biomarker identification to increase the interpretability of proteomics experiments. Our approach integrates a priori knowledge of the relationships between proteins and biological pathways and biological processes into sparse neural networks to create biologically informed neural networks. We employ these networks to differentiate between clinical subphenotypes of septic acute kidney injury and COVID-19, as well as acute respiratory distress syndrome of different aetiologies. To gain biological insight into the complex syndromes, we utilize feature attribution-methods to introspect the networks for the identification of proteins and pathways important for distinguishing between subtypes. The algorithms are implemented in a freely available open source Python-package ( https://github.com/InfectionMedicineProteomics/BINN ). |
| first_indexed | 2024-03-10T17:30:41Z |
| format | Article |
| id | doaj.art-6ba2704448034c108462da0c1d21a25d |
| institution | Directory Open Access Journal |
| issn | 2041-1723 |
| language | English |
| last_indexed | 2024-03-10T17:30:41Z |
| publishDate | 2023-09-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Nature Communications |
| spelling | doaj.art-6ba2704448034c108462da0c1d21a25d2023-11-20T10:02:51ZengNature PortfolioNature Communications2041-17232023-09-0114111310.1038/s41467-023-41146-4Interpreting biologically informed neural networks for enhanced proteomic biomarker discovery and pathway analysisErik Hartman0Aaron M. Scott1Christofer Karlsson2Tirthankar Mohanty3Suvi T. Vaara4Adam Linder5Lars Malmström6Johan Malmström7Division of Infection Medicine, Department of Clinical Sciences Lund, Faculty of Medicine, Lund UniversityDivision of Infection Medicine, Department of Clinical Sciences Lund, Faculty of Medicine, Lund UniversityDivision of Infection Medicine, Department of Clinical Sciences Lund, Faculty of Medicine, Lund UniversityDivision of Infection Medicine, Department of Clinical Sciences Lund, Faculty of Medicine, Lund UniversityDepartment of Perioperative and Intensive Care, University of Helsinki and Helsinki University HospitalDivision of Infection Medicine, Department of Clinical Sciences Lund, Faculty of Medicine, Lund UniversityDivision of Infection Medicine, Department of Clinical Sciences Lund, Faculty of Medicine, Lund UniversityDivision of Infection Medicine, Department of Clinical Sciences Lund, Faculty of Medicine, Lund UniversityAbstract The incorporation of machine learning methods into proteomics workflows improves the identification of disease-relevant biomarkers and biological pathways. However, machine learning models, such as deep neural networks, typically suffer from lack of interpretability. Here, we present a deep learning approach to combine biological pathway analysis and biomarker identification to increase the interpretability of proteomics experiments. Our approach integrates a priori knowledge of the relationships between proteins and biological pathways and biological processes into sparse neural networks to create biologically informed neural networks. We employ these networks to differentiate between clinical subphenotypes of septic acute kidney injury and COVID-19, as well as acute respiratory distress syndrome of different aetiologies. To gain biological insight into the complex syndromes, we utilize feature attribution-methods to introspect the networks for the identification of proteins and pathways important for distinguishing between subtypes. The algorithms are implemented in a freely available open source Python-package ( https://github.com/InfectionMedicineProteomics/BINN ).https://doi.org/10.1038/s41467-023-41146-4 |
| spellingShingle | Erik Hartman Aaron M. Scott Christofer Karlsson Tirthankar Mohanty Suvi T. Vaara Adam Linder Lars Malmström Johan Malmström Interpreting biologically informed neural networks for enhanced proteomic biomarker discovery and pathway analysis Nature Communications |
| title | Interpreting biologically informed neural networks for enhanced proteomic biomarker discovery and pathway analysis |
| title_full | Interpreting biologically informed neural networks for enhanced proteomic biomarker discovery and pathway analysis |
| title_fullStr | Interpreting biologically informed neural networks for enhanced proteomic biomarker discovery and pathway analysis |
| title_full_unstemmed | Interpreting biologically informed neural networks for enhanced proteomic biomarker discovery and pathway analysis |
| title_short | Interpreting biologically informed neural networks for enhanced proteomic biomarker discovery and pathway analysis |
| title_sort | interpreting biologically informed neural networks for enhanced proteomic biomarker discovery and pathway analysis |
| url | https://doi.org/10.1038/s41467-023-41146-4 |
| work_keys_str_mv | AT erikhartman interpretingbiologicallyinformedneuralnetworksforenhancedproteomicbiomarkerdiscoveryandpathwayanalysis AT aaronmscott interpretingbiologicallyinformedneuralnetworksforenhancedproteomicbiomarkerdiscoveryandpathwayanalysis AT christoferkarlsson interpretingbiologicallyinformedneuralnetworksforenhancedproteomicbiomarkerdiscoveryandpathwayanalysis AT tirthankarmohanty interpretingbiologicallyinformedneuralnetworksforenhancedproteomicbiomarkerdiscoveryandpathwayanalysis AT suvitvaara interpretingbiologicallyinformedneuralnetworksforenhancedproteomicbiomarkerdiscoveryandpathwayanalysis AT adamlinder interpretingbiologicallyinformedneuralnetworksforenhancedproteomicbiomarkerdiscoveryandpathwayanalysis AT larsmalmstrom interpretingbiologicallyinformedneuralnetworksforenhancedproteomicbiomarkerdiscoveryandpathwayanalysis AT johanmalmstrom interpretingbiologicallyinformedneuralnetworksforenhancedproteomicbiomarkerdiscoveryandpathwayanalysis |