Use the DeepFRI webserver to get sequence and structure based gene function prediction

Using the https://beta.deepfri.flatironinstitute.org/ webserver Use Structure-based gene function predictions using graph convolutional networks (Gligorijevic, et al., 2019)

get_deepfri_predictions(
  pdb_fname,
  tag = NULL,
  output_path = NULL,
  max_retries = 120,
  verbose = TRUE
)

Arguments

pdb_fname	proteins structure .pdb file
tag	used to tracking the prediction on the DeepFRI webserver and for the output path if not using use the basename without .pdb extension. (default: NULL)
output_path	folder where the output predictions should be written if null, then don't save output to file. (default: NULL)
max_retries	(default: 120)
verbose	use verbose output (default: TRUE)

Value

data.frame with columns structure_tag: <tag> prediction_type: [cnn, gcn] for sequence and structure based predictions go_term: e.g. GO:0044237 go_term_name: e.g. "cellular metabolic process" go_termscore: prediction score

Details

If you use for more than few predictions please coordinate with the Bonneau lab to not abuse their resources

usage example: On disk: <tag>.pdb In R:

predictions <- get_deepfri_predictions( pdb_fname = "<tag>.pdb", output_path = "intermediate_data")

If you use please cite:

Gligorijevic, Vladimir and Renfrew, P. Douglas and Kosciolek, Tomasz and Leman, Julia Koehler and Cho, Kyunghyun and Vatanen, Tommi and Berenberg, Daniel and Taylor, Bryn and Fisk, Ian M. and Xavier, Ramnik J. and Knight, Rob and Bonneau, Richard Structure-Based Function Prediction using Graph Convolutional Networks 2019, https://www.biorxiv.org/content/early/2019/10/04/786236