Published November 19, 2021 | Version v0.2.0

gitter-lab/nn4dms: Version 0.2.0

Description

  • Add notebook with the number of trainable parameters in each model
  • Add notebook showing how to visualize the latent space of a model with UMAP
  • Add notebook and code for GB1 sequence design and corresponding outputs
  • Add notebook and support for mutation- and position-based data splits
  • Add additional data splits: main experiment replicates, reduced training set sizes, mutation extrapolation, position extrapolation
  • Update conda environments

Files

gitter-lab/nn4dms-v0.2.0.zip

Files (244.9 MB)

Name Size
md5:8d4b1c221e371382638ba57dbdd53d92
244.9 MB Preview Download

Additional details

Related works

References

  • Neural networks to learn protein sequence-function relationships from deep mutational scanning data. Sam Gelman, Sarah A Fahlberg, Pete Heinzelman, Philip A Romero, Anthony Gitter. bioRxiv 2021. doi:10.1101/2020.10.25.353946