articleNature CommunicationsJul 27, 2022GOLD OA

ProtGPT2 is a deep unsupervised language model for protein design

Universitat de Girona · University of Bayreuth

PubMed
Indexed incrossrefdoajpubmed

Abstract

Protein design aims to build novel proteins customized for specific purposes, thereby holding the potential to tackle many environmental and biomedical problems. Recent progress in Transformer-based architectures has enabled the implementation of language models capable of generating text with human-like capabilities. Here, motivated by this success, we describe ProtGPT2, a language model trained on the protein space that generates de novo protein sequences following the principles of natural ones. The generated proteins display natural amino acid propensities, while disorder predictions indicate that 88% of ProtGPT2-generated proteins are globular, in line with natural sequences. Sensitive sequence searches…

Citation impact

760
total citations
FWCI
57.43
Percentile
100%
References
71
Citations per year

Authors

3

Topics & keywords

Keywords
  • Computer science
  • Protein design
  • Computational biology
  • Natural language
  • Protein structure
  • Similarity (geometry)
  • Bridging (networking)
  • Protein sequencing
No related works found for this paper.

Funding