Mimicry Embedding Facilitates Advanced Neural Network Training for Image-Based Pathogen Detection
ABSTRACT The use of deep neural networks (DNNs) for analysis of complex biomedical images shows great promise but is hampered by a lack of large verified data sets for rapid network evolution. Here, we present a novel strategy, termed “mimicry embedding,” for rapid application of neural network arch...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
American Society for Microbiology
2020-10-01
|
| Series: | mSphere |
| Subjects: | |
| Online Access: | https://journals.asm.org/doi/10.1128/mSphere.00836-20 |
| _version_ | 1818935950667939840 |
|---|---|
| author | Artur Yakimovich Moona Huttunen Jerzy Samolej Barbara Clough Nagisa Yoshida Serge Mostowy Eva-Maria Frickel Jason Mercer |
| author_facet | Artur Yakimovich Moona Huttunen Jerzy Samolej Barbara Clough Nagisa Yoshida Serge Mostowy Eva-Maria Frickel Jason Mercer |
| author_sort | Artur Yakimovich |
| collection | DOAJ |
| description | ABSTRACT The use of deep neural networks (DNNs) for analysis of complex biomedical images shows great promise but is hampered by a lack of large verified data sets for rapid network evolution. Here, we present a novel strategy, termed “mimicry embedding,” for rapid application of neural network architecture-based analysis of pathogen imaging data sets. Embedding of a novel host-pathogen data set, such that it mimics a verified data set, enables efficient deep learning using high expressive capacity architectures and seamless architecture switching. We applied this strategy across various microbiological phenotypes, from superresolved viruses to in vitro and in vivo parasitic infections. We demonstrate that mimicry embedding enables efficient and accurate analysis of two- and three-dimensional microscopy data sets. The results suggest that transfer learning from pretrained network data may be a powerful general strategy for analysis of heterogeneous pathogen fluorescence imaging data sets. IMPORTANCE In biology, the use of deep neural networks (DNNs) for analysis of pathogen infection is hampered by a lack of large verified data sets needed for rapid network evolution. Artificial neural networks detect handwritten digits with high precision thanks to large data sets, such as MNIST, that allow nearly unlimited training. Here, we developed a novel strategy we call mimicry embedding, which allows artificial intelligence (AI)-based analysis of variable pathogen-host data sets. We show that deep learning can be used to detect and classify single pathogens based on small differences. |
| first_indexed | 2024-12-20T05:28:19Z |
| format | Article |
| id | doaj.art-d26b0508dfde4b1581e6233c5ef9daa5 |
| institution | Directory Open Access Journal |
| issn | 2379-5042 |
| language | English |
| last_indexed | 2024-12-20T05:28:19Z |
| publishDate | 2020-10-01 |
| publisher | American Society for Microbiology |
| record_format | Article |
| series | mSphere |
| spelling | doaj.art-d26b0508dfde4b1581e6233c5ef9daa52022-12-21T19:51:49ZengAmerican Society for MicrobiologymSphere2379-50422020-10-015510.1128/mSphere.00836-20Mimicry Embedding Facilitates Advanced Neural Network Training for Image-Based Pathogen DetectionArtur Yakimovich0Moona Huttunen1Jerzy Samolej2Barbara Clough3Nagisa Yoshida4Serge Mostowy5Eva-Maria Frickel6Jason Mercer7MRC-Laboratory for Molecular Cell Biology, University College London, London, United KingdomMRC-Laboratory for Molecular Cell Biology, University College London, London, United KingdomMRC-Laboratory for Molecular Cell Biology, University College London, London, United KingdomInstitute of Microbiology and Infection, University of Birmingham, Birmingham, United KingdomHost-Toxoplasma Interaction Laboratory, The Francis Crick Institute, London, United KingdomDepartment of Infection Biology, London School of Hygiene & Tropical Medicine, London, United KingdomInstitute of Microbiology and Infection, University of Birmingham, Birmingham, United KingdomMRC-Laboratory for Molecular Cell Biology, University College London, London, United KingdomABSTRACT The use of deep neural networks (DNNs) for analysis of complex biomedical images shows great promise but is hampered by a lack of large verified data sets for rapid network evolution. Here, we present a novel strategy, termed “mimicry embedding,” for rapid application of neural network architecture-based analysis of pathogen imaging data sets. Embedding of a novel host-pathogen data set, such that it mimics a verified data set, enables efficient deep learning using high expressive capacity architectures and seamless architecture switching. We applied this strategy across various microbiological phenotypes, from superresolved viruses to in vitro and in vivo parasitic infections. We demonstrate that mimicry embedding enables efficient and accurate analysis of two- and three-dimensional microscopy data sets. The results suggest that transfer learning from pretrained network data may be a powerful general strategy for analysis of heterogeneous pathogen fluorescence imaging data sets. IMPORTANCE In biology, the use of deep neural networks (DNNs) for analysis of pathogen infection is hampered by a lack of large verified data sets needed for rapid network evolution. Artificial neural networks detect handwritten digits with high precision thanks to large data sets, such as MNIST, that allow nearly unlimited training. Here, we developed a novel strategy we call mimicry embedding, which allows artificial intelligence (AI)-based analysis of variable pathogen-host data sets. We show that deep learning can be used to detect and classify single pathogens based on small differences.https://journals.asm.org/doi/10.1128/mSphere.00836-20capsule networkstransfer learningsuperresolution microscopyvaccinia virusToxoplasma gondiizebrafish |
| spellingShingle | Artur Yakimovich Moona Huttunen Jerzy Samolej Barbara Clough Nagisa Yoshida Serge Mostowy Eva-Maria Frickel Jason Mercer Mimicry Embedding Facilitates Advanced Neural Network Training for Image-Based Pathogen Detection mSphere capsule networks transfer learning superresolution microscopy vaccinia virus Toxoplasma gondii zebrafish |
| title | Mimicry Embedding Facilitates Advanced Neural Network Training for Image-Based Pathogen Detection |
| title_full | Mimicry Embedding Facilitates Advanced Neural Network Training for Image-Based Pathogen Detection |
| title_fullStr | Mimicry Embedding Facilitates Advanced Neural Network Training for Image-Based Pathogen Detection |
| title_full_unstemmed | Mimicry Embedding Facilitates Advanced Neural Network Training for Image-Based Pathogen Detection |
| title_short | Mimicry Embedding Facilitates Advanced Neural Network Training for Image-Based Pathogen Detection |
| title_sort | mimicry embedding facilitates advanced neural network training for image based pathogen detection |
| topic | capsule networks transfer learning superresolution microscopy vaccinia virus Toxoplasma gondii zebrafish |
| url | https://journals.asm.org/doi/10.1128/mSphere.00836-20 |
| work_keys_str_mv | AT arturyakimovich mimicryembeddingfacilitatesadvancedneuralnetworktrainingforimagebasedpathogendetection AT moonahuttunen mimicryembeddingfacilitatesadvancedneuralnetworktrainingforimagebasedpathogendetection AT jerzysamolej mimicryembeddingfacilitatesadvancedneuralnetworktrainingforimagebasedpathogendetection AT barbaraclough mimicryembeddingfacilitatesadvancedneuralnetworktrainingforimagebasedpathogendetection AT nagisayoshida mimicryembeddingfacilitatesadvancedneuralnetworktrainingforimagebasedpathogendetection AT sergemostowy mimicryembeddingfacilitatesadvancedneuralnetworktrainingforimagebasedpathogendetection AT evamariafrickel mimicryembeddingfacilitatesadvancedneuralnetworktrainingforimagebasedpathogendetection AT jasonmercer mimicryembeddingfacilitatesadvancedneuralnetworktrainingforimagebasedpathogendetection |