Screen_Covalent_Compound_by_LISNN

Title: Discovery of Covalent Lead Compounds Targeting 3CL Protease with Lateral Interactions Spiking Neural Network

Introduction

Apply Lateral Interactions Spiking Neural Network to screening the covalent Lead Compounds targeting 3CL Pro.

Requirements:

Python        3.7
torch-gpu     1.11.0
numpy         1.21.5
jieba         0.42.1
pandas        1.3.5
scikit-learn  1.0.2
scipy         1.7.3
seaborn       0.12.2
torchaudio    0.11.0
torchvision   0.12.0
gensim        3.8.3
matplotlib    3.5.3
seaborn       0.12.2
Other packages and their versions are shown in list.txt

File description

list.txt                      Packages and their versions required to run the environment

compare_protein.xlsx          Protein amino acid sequence represented by numbers in figure S3

3CL.csv                       Raw data on inhibitors targeting 3CL Pro

model_human                   The file that evaluates the classification performance of the model


model_Compounds_Inhibitory_Activity_Dataset_Targeting_3CL_Pro       The file of train model and screen inhibitors target 3CL pro

model_Covalent_Complex_Dataset_Targeting_Cys The file of training and validating model and screen covalent compound targeting Cys

../LISNN.py                      The model of LISNN

../pre_data_embedding_data_interaction.py    ../pre_data_embedding.py   ../predict_embedding.py   ../Validation_predict_embedding.py               SMILES sequences of compounds and amino acid sequences of proteins are converted into vectors by Word2Vec.

../train.py                      Training models are based on different datasets

../predict.py   ../Validation_predict.py                    The probability of predicting the positive result

../model_Covalent_Complex_Dataset_Targeting_Cys/T-SNE.py    The model of t-SNE

../../data                          Relevant data to bulid the model

../../data/gensim-model-...         Word2Vec model

../../screen                        Relevant data for application model screening

../../seed                          The trained model represented by different seed number

../model_Covalent_Complex_Dataset_Targeting_Cys/screen/screen_specs           Processed commercial screening library