pysprintppi 1.1

Protein-protein interaction predictor

PySprint

Re-implementation of SPRINT in Rust, a fast, modern programming language, and wrapped in a Python API.

Author

Francois Charih - [e-mail] - Developer

To install PySPRINT

To install PySPRINT as a Python library, you must run the following commands:

$ pip3 install pysprintppi

Note: We support Python3.8+ on the Linux platform.

Using PySPRINT

To predict interactions, you must use the PySPRINT class which is instantiated as follows:

# Instantiate the PySPRINT PPI predictor
pysprint = PySPRINT(
  "my_fasta_file_containing_my_sequences.fasta",
  "my_hsp_file.hsp", # 
  "training_pairs_file.txt"
)

# Score the interactions (all-to-all predictions)
prediction_matrix = pysprint.score_all_interactions()

You can then do things with that prediction matrix:

# Retrieve the predictions for a given pair
print(prediction_matrix.get_score_for_pair("protein1", "protein2"))

# Retrieve the scores for all proteins that interact with ProteinX
print(prediction_matrix.get_scores_for_protein("ProteinX"))

# Save the prediction matrix
prediction_matrix.save("path_to_matrix.npy")

# Load a prediction matrix
prediction_matrix = prediction_matrix.load("path_to_matrix.npy")

Running PySPRINT with Docker

$ docker build -t pysprint .
$ docker run -v $(pwd):/pysprint pysprint:latest