Positive-unlabeled convolutional neural networks for particle picking in cryo-electron micrographs
© 2019, The Author(s), under exclusive licence to Springer Nature America, Inc. Cryo-electron microscopy is a popular method for the determination of protein structures; however, identifying a sufficient number of particles for analysis can take months of manual effort. Current computational approac...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Springer Science and Business Media LLC
2021
|
| Online Access: | https://hdl.handle.net/1721.1/136321 |
| _version_ | 1826198281631301632 |
|---|---|
| author | Bepler, Tristan Morin, Andrew Rapp, Micah Brasch, Julia Shapiro, Lawrence Noble, Alex J Berger, Bonnie |
| author_facet | Bepler, Tristan Morin, Andrew Rapp, Micah Brasch, Julia Shapiro, Lawrence Noble, Alex J Berger, Bonnie |
| author_sort | Bepler, Tristan |
| collection | MIT |
| description | © 2019, The Author(s), under exclusive licence to Springer Nature America, Inc. Cryo-electron microscopy is a popular method for the determination of protein structures; however, identifying a sufficient number of particles for analysis can take months of manual effort. Current computational approaches find many false positives and require ad hoc postprocessing, especially for unusually shaped particles. To address these shortcomings, we develop Topaz, an efficient and accurate particle-picking pipeline using neural networks trained with a general-purpose positive-unlabeled learning method. This framework enables particle detection models to be trained with few sparsely labeled particles and no labeled negatives. Topaz retrieves many more real particles than conventional picking methods while maintaining low false-positive rates, is capable of picking challenging unusually shaped proteins (for example, small, non-globular and asymmetric particles), produces more representative particle sets and does not require post hoc curation. We demonstrate the performance of Topaz on two difficult datasets and three conventional datasets. Topaz is modular, standalone, free and open source (http://topaz.csail.mit.edu). |
| first_indexed | 2024-09-23T11:02:25Z |
| format | Article |
| id | mit-1721.1/136321 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T11:02:25Z |
| publishDate | 2021 |
| publisher | Springer Science and Business Media LLC |
| record_format | dspace |
| spelling | mit-1721.1/1363212022-10-01T00:43:56Z Positive-unlabeled convolutional neural networks for particle picking in cryo-electron micrographs Bepler, Tristan Morin, Andrew Rapp, Micah Brasch, Julia Shapiro, Lawrence Noble, Alex J Berger, Bonnie © 2019, The Author(s), under exclusive licence to Springer Nature America, Inc. Cryo-electron microscopy is a popular method for the determination of protein structures; however, identifying a sufficient number of particles for analysis can take months of manual effort. Current computational approaches find many false positives and require ad hoc postprocessing, especially for unusually shaped particles. To address these shortcomings, we develop Topaz, an efficient and accurate particle-picking pipeline using neural networks trained with a general-purpose positive-unlabeled learning method. This framework enables particle detection models to be trained with few sparsely labeled particles and no labeled negatives. Topaz retrieves many more real particles than conventional picking methods while maintaining low false-positive rates, is capable of picking challenging unusually shaped proteins (for example, small, non-globular and asymmetric particles), produces more representative particle sets and does not require post hoc curation. We demonstrate the performance of Topaz on two difficult datasets and three conventional datasets. Topaz is modular, standalone, free and open source (http://topaz.csail.mit.edu). 2021-10-27T20:34:52Z 2021-10-27T20:34:52Z 2019 2021-05-17T17:35:46Z Article http://purl.org/eprint/type/JournalArticle https://hdl.handle.net/1721.1/136321 en 10.1038/S41592-019-0575-8 Nature Methods Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf Springer Science and Business Media LLC PMC |
| spellingShingle | Bepler, Tristan Morin, Andrew Rapp, Micah Brasch, Julia Shapiro, Lawrence Noble, Alex J Berger, Bonnie Positive-unlabeled convolutional neural networks for particle picking in cryo-electron micrographs |
| title | Positive-unlabeled convolutional neural networks for particle picking in cryo-electron micrographs |
| title_full | Positive-unlabeled convolutional neural networks for particle picking in cryo-electron micrographs |
| title_fullStr | Positive-unlabeled convolutional neural networks for particle picking in cryo-electron micrographs |
| title_full_unstemmed | Positive-unlabeled convolutional neural networks for particle picking in cryo-electron micrographs |
| title_short | Positive-unlabeled convolutional neural networks for particle picking in cryo-electron micrographs |
| title_sort | positive unlabeled convolutional neural networks for particle picking in cryo electron micrographs |
| url | https://hdl.handle.net/1721.1/136321 |
| work_keys_str_mv | AT beplertristan positiveunlabeledconvolutionalneuralnetworksforparticlepickingincryoelectronmicrographs AT morinandrew positiveunlabeledconvolutionalneuralnetworksforparticlepickingincryoelectronmicrographs AT rappmicah positiveunlabeledconvolutionalneuralnetworksforparticlepickingincryoelectronmicrographs AT braschjulia positiveunlabeledconvolutionalneuralnetworksforparticlepickingincryoelectronmicrographs AT shapirolawrence positiveunlabeledconvolutionalneuralnetworksforparticlepickingincryoelectronmicrographs AT noblealexj positiveunlabeledconvolutionalneuralnetworksforparticlepickingincryoelectronmicrographs AT bergerbonnie positiveunlabeledconvolutionalneuralnetworksforparticlepickingincryoelectronmicrographs |