Low-N protein engineering with data-efficient deep learning
Protein engineering has enormous academic and industrial potential. However, it is limited by the lack of experimental assays that are consistent with the design goal and sufficiently high throughput to find rare, enhanced variants. Here we introduce a machine learning-guided paradigm that can use a...
| Main Authors: | , , , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
Springer Science and Business Media LLC
2021
|
| Online Access: | https://hdl.handle.net/1721.1/134193 |
| _version_ | 1811078386906300416 |
|---|---|
| author | Biswas, Surojit Khimulya, Grigory Alley, Ethan C Esvelt, Kevin M Church, George M |
| author2 | Massachusetts Institute of Technology. Media Laboratory |
| author_facet | Massachusetts Institute of Technology. Media Laboratory Biswas, Surojit Khimulya, Grigory Alley, Ethan C Esvelt, Kevin M Church, George M |
| author_sort | Biswas, Surojit |
| collection | MIT |
| description | Protein engineering has enormous academic and industrial potential. However, it is limited by the lack of experimental assays that are consistent with the design goal and sufficiently high throughput to find rare, enhanced variants. Here we introduce a machine learning-guided paradigm that can use as few as 24 functionally assayed mutant sequences to build an accurate virtual fitness landscape and screen ten million sequences via in silico directed evolution. As demonstrated in two dissimilar proteins, GFP from Aequorea victoria (avGFP) and E. coli strain TEM-1 β-lactamase, top candidates from a single round are diverse and as active as engineered mutants obtained from previous high-throughput efforts. By distilling information from natural protein sequence landscapes, our model learns a latent representation of 'unnaturalness', which helps to guide search away from nonfunctional sequence neighborhoods. Subsequent low-N supervision then identifies improvements to the activity of interest. In sum, our approach enables efficient use of resource-intensive high-fidelity assays without sacrificing throughput, and helps to accelerate engineered proteins into the fermenter, field and clinic. |
| first_indexed | 2024-09-23T10:58:42Z |
| format | Article |
| id | mit-1721.1/134193 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T10:58:42Z |
| publishDate | 2021 |
| publisher | Springer Science and Business Media LLC |
| record_format | dspace |
| spelling | mit-1721.1/1341932023-09-15T18:47:01Z Low-N protein engineering with data-efficient deep learning Biswas, Surojit Khimulya, Grigory Alley, Ethan C Esvelt, Kevin M Church, George M Massachusetts Institute of Technology. Media Laboratory Protein engineering has enormous academic and industrial potential. However, it is limited by the lack of experimental assays that are consistent with the design goal and sufficiently high throughput to find rare, enhanced variants. Here we introduce a machine learning-guided paradigm that can use as few as 24 functionally assayed mutant sequences to build an accurate virtual fitness landscape and screen ten million sequences via in silico directed evolution. As demonstrated in two dissimilar proteins, GFP from Aequorea victoria (avGFP) and E. coli strain TEM-1 β-lactamase, top candidates from a single round are diverse and as active as engineered mutants obtained from previous high-throughput efforts. By distilling information from natural protein sequence landscapes, our model learns a latent representation of 'unnaturalness', which helps to guide search away from nonfunctional sequence neighborhoods. Subsequent low-N supervision then identifies improvements to the activity of interest. In sum, our approach enables efficient use of resource-intensive high-fidelity assays without sacrificing throughput, and helps to accelerate engineered proteins into the fermenter, field and clinic. 2021-10-27T20:03:55Z 2021-10-27T20:03:55Z 2021 2021-06-23T16:14:40Z Article http://purl.org/eprint/type/JournalArticle https://hdl.handle.net/1721.1/134193 en 10.1038/s41592-021-01100-y Nature Methods Article is made available in accordance with the publisher's policy and may be subject to US copyright law. Please refer to the publisher's site for terms of use. application/pdf Springer Science and Business Media LLC bioRxiv |
| spellingShingle | Biswas, Surojit Khimulya, Grigory Alley, Ethan C Esvelt, Kevin M Church, George M Low-N protein engineering with data-efficient deep learning |
| title | Low-N protein engineering with data-efficient deep learning |
| title_full | Low-N protein engineering with data-efficient deep learning |
| title_fullStr | Low-N protein engineering with data-efficient deep learning |
| title_full_unstemmed | Low-N protein engineering with data-efficient deep learning |
| title_short | Low-N protein engineering with data-efficient deep learning |
| title_sort | low n protein engineering with data efficient deep learning |
| url | https://hdl.handle.net/1721.1/134193 |
| work_keys_str_mv | AT biswassurojit lownproteinengineeringwithdataefficientdeeplearning AT khimulyagrigory lownproteinengineeringwithdataefficientdeeplearning AT alleyethanc lownproteinengineeringwithdataefficientdeeplearning AT esveltkevinm lownproteinengineeringwithdataefficientdeeplearning AT churchgeorgem lownproteinengineeringwithdataefficientdeeplearning |