GitHub - jiaqingxie/DeepProtein: A Deep Learning Toolkit for Protein Property Prediction, Localization Prediction, Protein-Protein Interaction, antigen epitope prediction, antibody paratope prediction, antibody developability prediction, etc. [NeurIPS 2024 AIDrugX Spotlight]

[DeepProtein: Deep Learning Library and Benchmark for Protein Sequence Learning](https://arxiv.org/abs/2410.02023, NeurIPS 2024 AIDrugX Spotlight)

Applications in Protein Property Prediction, Localization Prediction, Protein-Protein Interaction, Antigen Epitope Prediction, Antibody Paratope Prediction, Antibody Developability Prediction, and more.

Introduction

Understanding proteomics is critical for advancing biology, genomics, and medicine. Proteins perform essential roles, such as catalyzing biochemical reactions and providing immune responses. With the rise of 3D databases like AlphaFold 2.0, machine learning has become a powerful tool for studying protein mechanisms.

Why DeepProtein?

Deep learning has revolutionized tasks such as:

Protein-protein interaction
Protein folding
Protein-ligand interaction
Protein function and property prediction

However, current benchmarks often focus on sequential methods like CNNs and transformers, overlooking graph-based models and lacking user-friendly interfaces.

What is DeepProtein?

DeepProtein is a comprehensive deep learning library and benchmark designed to fill these gaps:

Comprehensive Benchmarking: Evaluating CNNs, RNNs, transformers, and GNNs on 7 essential protein learning tasks, such as function prediction and antibody developability.
User-friendly Interface: Simplifying execution with one command for all tasks.
Enhanced Accessibility: Extensive documentation and tutorials for reproducible research.

News

[12/24] The documentation of DeepProtein is still under construction. It's at https://deepprotein.readthedocs.io/
[12/24] DeepProtein is going to be supported with pretrained shallow DL models.
[12/24] DeepProtein now supports BioMistral-7B model, working on [BioT5+, BioT5, ChemLLM, and LlaSMol]
[12/24] DeepProtein now supports four new Protein Language Models: ESM-1-650M, ESM-2-650M, Prot-Bert and Prot-T5 Models for Protein Function Prediction.
[11/24] DeepProtein is accepted at NeurIPS AI4DrugX as Spotlight. It's under revision at Bioinformatics.
[09/24] DeepProtein is online.

Installation

First, we recommend you follow the instructions on how DeepPurpose's dependencies are installed.

conda create -n DeepProtein python=3.9
conda activate DeepProtein
pip install git+https://github.com/bp-kelley/descriptastorus
pip install lmdb seaborn wandb pydantic DeepPurpose
conda install -c conda-forge pytdc

A version of torch 2.1+ is required to be installed since Jul requires a version of torch >=2.1.0.

If you want to use GPU, then first find a matched torch version, then install duel with cuda version. We give an example of torch 2.3.0 with cuda 11.8:

pip install torch==2.3.0 torchvision==0.18.0 torchaudio==2.3.0 --index-url https://download.pytorch.org/whl/cu118
conda install -c dglteam/label/th23_cu118 dgl

If you are not using a GPU, then follow this:

pip install torch==2.3.0 torchvision==0.18.0 torchaudio==2.3.0 --index-url https://download.pytorch.org/whl/cpu
conda install -c dglteam/label/th23_cpu dgl

Demos

Checkout some demos & tutorials to start:

Name	Description
Dataset Tutorial	Tutorial on how to use the dataset loader and read customized data
Protein Property Prediction Tutorial	Example of CNN on Beta-lactamase property prediction
Protein Protein Interaction Tutorial	Example of CNN on PPI Affinity prediction

Example

We give two examples for each case study. One is trained with fixed parameters (a) and one is trained with argument. The argument list is given below.

Argument	Description
target_encoding	'CNN' / 'Transformer' for sequential learning, or 'DGL_GCN' for 'DGL_AttentiveFP' for structure learning. Current available protein encoding belongs to this full list: ['CNN', 'Transformer', 'CNN_RNN', 'DGL_GCN', 'DGL_GAT', 'DGL_AttentiveFP', 'DGL_NeuralFP', 'DGL_MPNN', 'PAGTN', 'Graphormer']. For residue level tasks, the protein encoding list is ['Token_CNN', 'Token_CNN_RNN, 'Token_Transformer']
seed	For paper: 7 / 42 /100. You could try your own seed.
wandb_proj	The name of your wandb project that you wish to save the results into.
lr	Learning rate. We recommend 1e-4 for non-GNN learning and 1e-5 for GNN learning.
epochs	Number of training epochs. Generally setting 60 - 100 epochs leads to convergence.
compute_pos_enc *	Compute positional encoding for using graph transformers. We dont recommend add this inductive bias into GNN as GNN itself already encoded it. This don't work effectively on large scale graphs with dgl so the implementation is still under test.
batch_size	Batch size of 8 - 32 is good for protein sequence learning.

Case Study 1(a): A Framework for Protein Function (Property) Prediction

Click here for the code!

import os, sys, argparse, torch, wandb

### Our library DeepProtein
from DeepProtein.dataset import *
import DeepProtein.utils as utils
import DeepProtein.ProteinPred as models

### Load Beta lactamase dataset
path = os.getcwd()
train_beta = Beta_lactamase(path + '/DeepProtein/data', 'train')
valid_beta = Beta_lactamase(path + '/DeepProtein/data', 'valid')
test_beta = Beta_lactamase(path + '/DeepProtein/data', 'test')

train_protein_processed, train_target, train_protein_idx = collate_fn(train_beta)
valid_protein_processed, valid_target, valid_protein_idx = collate_fn(valid_beta)
test_protein_processed, test_target, test_protein_idx = collate_fn(test_beta)

### Train Valid Test Split
target_encoding = 'CNN'
train, _, _ = utils.data_process(X_target=train_protein_processed, y=train_target, 
    target_encoding=target_encoding, split_method='random', frac=[0.99998, 1e-5, 1e-5])

_, val, _ = utils.data_process(X_target=valid_protein_processed, y=valid_target,        
    target_encoding=target_encoding, split_method='random', frac=[1e-5, 0.99998, 1e-5])

_, _, test = utils.data_process(X_target=test_protein_processed, y=test_target,         
    target_encoding=target_encoding,split_method='random', frac=[1e-5, 1e-5, 0.99998])
                            
### Load configuration for model
config = generate_config(target_encoding=target_encoding,
                         cls_hidden_dims=[1024, 1024],
                         train_epoch=20,
                         LR=0.0001,
                         batch_size=32,
                         )
config['multi'] = False
torch.manual_seed(args.seed)
model = models.model_initialize(**config)

### Train our model
model.train(train, val, test, compute_pos_enc = False)

If you want to use structure learning methods such as graph neural network, please set the second parameters in the collate_fn() into True.

(b) If you wish to use arguments, this could be trained in one line. All mentioned GNN variants above is available for training.

CNN Case

python train/beta.py --target_encoding CNN --seed 7 --wandb_proj DeepProtein --lr 0.0001 --epochs 100

GNN Case

python train/beta.py --target_encoding DGL_GCN --seed 7 --wandb_proj DeepProtein --lr 0.00001 --epochs 100

Case Study 1(b): A Framework for Protein Protein Interaction Prediction

Click here for the code!

### package import
import os, sys, argparse, torch, wandb

### Our library DeepProtein
from DeepProtein.dataset import *
import DeepProtein.utils as utils
import DeepProtein.PPI as models

### Load PPI Affinity dataset
path = os.getcwd()
train_ppi = PPI_Affinity(path + '/DeepProtein/data', 'train')
valid_ppi = PPI_Affinity(path + '/DeepProtein/data', 'valid')
test_ppi = PPI_Affinity(path + '/DeepProtein/data', 'test')

train_protein_1, train_protein_2, train_target, train_protein_idx = collate_fn_ppi(train_ppi, graph=False, unsqueeze=False)
valid_protein_1, valid_protein_2, valid_target, valid_protein_idx = collate_fn_ppi(valid_ppi, graph=False, unsqueeze=False)
test_protein_1, test_protein_2, test_target, test_protein_idx = collate_fn_ppi(test_ppi, graph=False, unsqueeze=False)

### Train Valid Test Split
target_encoding = 'CNN'
train, _, _ = data_process(X_target = train_protein_1, X_target_ = train_protein_2, y = train_target,
                target_encoding = target_encoding,
                split_method='random', frac=[0.99998, 1e-5, 1e-5],
                random_seed = 1)
_, val, _ = data_process(X_target = valid_protein_1, X_target_ = valid_protein_2, y = valid_target,
                target_encoding = target_encoding,
                split_method='random',frac=[1e-5, 0.99998, 1e-5],
                random_seed = 1)

_, _, test = data_process(X_target = test_protein_1, X_target_ = test_protein_2, y = test_target,
                target_encoding = target_encoding,
                split_method='random',frac=[1e-5, 1e-5, 0.99998],
                random_seed = 1)
                            
### Load configuration for model
config = generate_config(target_encoding=target_encoding,
                         cls_hidden_dims=[512],
                         train_epoch=20,
                         LR=0.0001,
                         batch_size=32,
                         )
# config['multi'] = False
torch.manual_seed(args.seed)
model = models.model_initialize(**config)

### Train our model
model.train(train, val, test)

If you want to use structure learning methods such as graph neural network, please set the second parameters in the collate_fn_ppi() into True.

(b) If you wish to use arguments, this could be trained in one line. For GNN, only DGL_GCN, DGL_GAT and DGL_NeuralFP is available currently.

CNN Case

python train/ppi_affinity.py --target_encoding CNN --seed 42 --wandb_proj DeepProtein --lr 0.0001 --epochs 100

GNN Case

python train/ppi_affinity.py --target_encoding DGL_GCN --seed 42 --wandb_proj DeepProtein --lr 0.00001 --epochs 100

Case Study 1(c): A Framework for Protein Localization Prediction

Click here for the code!

### package import
import os, sys, argparse, torch, wandb

### Our library DeepProtein
from DeepProtein.dataset import *
import DeepProtein.utils as utils
import DeepProtein.ProteinPred as models

### Load Subcellular Dataset
path = os.getcwd()
train_sub = Subcellular(path + '/DeepProtein/data', 'train')
valid_sub = Subcellular(path + '/DeepProtein/data', 'valid')
test_sub = Subcellular(path + '/DeepProtein/data', 'test')

train_protein_processed, train_target, train_protein_idx = collate_fn(train_sub)
valid_protein_processed, valid_target, valid_protein_idx = collate_fn(valid_sub)
test_protein_processed, test_target, test_protein_idx = collate_fn(test_sub)

### Train Valid Test Split
target_encoding = 'CNN'
train, _, _ = utils.data_process(X_target=train_protein_processed, y=train_target, 
    target_encoding=target_encoding, split_method='random', frac=[0.99998, 1e-5, 1e-5])

_, val, _ = utils.data_process(X_target=valid_protein_processed, y=valid_target,        
    target_encoding=target_encoding, split_method='random', frac=[1e-5, 0.99998, 1e-5])

_, _, test = utils.data_process(X_target=test_protein_processed, y=test_target,         
    target_encoding=target_encoding,split_method='random', frac=[1e-5, 1e-5, 0.99998])
                            
### Load configuration for model
config = generate_config(target_encoding=target_encoding,
                         cls_hidden_dims=[1024, 1024],
                         train_epoch=20,
                         LR=0.0001,
                         batch_size=32,
                         )
config['binary'] = False
config['multi'] = True
config['classes'] = 10
torch.manual_seed(args.seed)
model = models.model_initialize(**config)

### Train our model
model.train(train, val, test, compute_pos_enc = False)

If you want to use structure learning methods such as graph neural network, please set the second parameters in the collate_fn() into True. Note that SubCellular is multi-class classification problem, therefore you should set config['multi'] to True.

(b) If you wish to use arguments, this could be trained in one line. All mentioned GNN variants above is available for training.

CNN Case

python train/subcellular.py --target_encoding CNN --seed 7 --wandb_proj DeepProtein --lr 0.0001 --epochs 100

GNN Case

python train/subcellular.py --target_encoding DGL_GCN --seed 7 --wandb_proj DeepProtein --lr 0.00001 --epochs 100

Case Study 1(d): A Framework for Antigen Epitope Prediction

Make sure that tdc is installed, if not

pip install PyTDC

Click here for the code!

### package import
import os, sys, argparse, torch, wandb

### Our library DeepProtein
from DeepProtein.dataset import *
import DeepProtein.utils as utils
import DeepProtein.TokenPred as models
from tdc.single_pred import Epitope

### Load Epitope  Dataset
data_class, name, X = Epitope, 'IEDB_Jespersen', 'Antigen'
data = data_class(name=name)
split = data.get_split()

train_data, valid_data, test_data = split['train'], split['valid'], split['test']
vocab_set = set()

train_vocab, train_positive_ratio = data2vocab(train_data, train_data, X)
valid_vocab, valid_positive_ratio = data2vocab(valid_data, train_data, X)
test_vocab, test_positive_ratio = data2vocab(test_data, train_data, X)

vocab_set = train_vocab.union(valid_vocab)
vocab_set = vocab_set.union(test_vocab)
vocab_lst = list(vocab_set)

### Train Valid Test Split
train_data = standardize_data(train_data, vocab_lst, X)
valid_data = standardize_data(valid_data, vocab_lst, X)
test_data = standardize_data(test_data, vocab_lst, X)

train_set = data_process_loader_Token_Protein_Prediction(train_data)
valid_set = data_process_loader_Token_Protein_Prediction(valid_data)
test_set = data_process_loader_Token_Protein_Prediction(test_data)

### Load configuration for model
config = generate_config(target_encoding=target_encoding,
                         cls_hidden_dims=[1024, 1024],
                         train_epoch=20,
                         LR=0.0001,
                         batch_size=32,
                         )
config['multi'] = False
config['binary'] = True
config['token'] = True
config['in_channels'] = 24
torch.manual_seed(args.seed)
model = models.model_initialize(**config)


### Train our model
model.train(train_set, valid_set, test_set, batch_size=batch_size)

(b) If you wish to use arguments, this could be trained in one line. All mentioned GNN variants above is not available for training. For an important notice, this task is residue (token) level classification therefore we deploy token level CNN, CNN_RNN and Transformer models.

Current version only supports Token_CNN, Token_CNN_RNN and Token_Transformer for target_encoding.

CNN Case

python train/IEDB.py --target_encoding Token_CNN --seed 7 --wandb_proj DeepProtein --lr 0.0001 --epochs 100

Case Study 1(e): A Framework for Antibody Paratope Prediction

Make sure that tdc is installed, if not

pip install PyTDC

Click here for the code!

### package import
import os, sys, argparse, torch, wandb 

### Our library DeepProtein
from DeepProtein.dataset import *
import DeepProtein.utils as utils
import DeepProtein.TokenPred as models
from tdc.single_pred import Paratope

### Load Paratope Dataset
data_class, name, X = Paratope, 'SAbDab_Liberis', 'Antibody'
data = data_class(name=name)
split = data.get_split()
train_data, valid_data, test_data = split['train'], split['valid'], split['test']
vocab_set = set()

train_vocab, train_positive_ratio = data2vocab(train_data, train_data, X)
valid_vocab, valid_positive_ratio = data2vocab(valid_data, train_data, X)
test_vocab, test_positive_ratio = data2vocab(test_data, train_data, X)

vocab_set = train_vocab.union(valid_vocab)
vocab_set = vocab_set.union(test_vocab)
vocab_lst = list(vocab_set)



### Train Valid Test Split
train_data = standardize_data(train_data, vocab_lst, X)
valid_data = standardize_data(valid_data, vocab_lst, X)
test_data = standardize_data(test_data, vocab_lst, X)

train_set = data_process_loader_Token_Protein_Prediction(train_data)
valid_set = data_process_loader_Token_Protein_Prediction(valid_data)
test_set = data_process_loader_Token_Protein_Prediction(test_data)

### Load configuration for model
config = generate_config(target_encoding=target_encoding,
                         cls_hidden_dims=[1024, 1024],
                         train_epoch=20,
                         LR=0.0001,
                         batch_size=32,
                         )
config['multi'] = False
config['binary'] = True
config['token'] = True
config['in_channels'] = 20
torch.manual_seed(args.seed)
model = models.model_initialize(**config)


### Train our model
model.train(train_set, valid_set, test_set, batch_size=batch_size)

(b) If you wish to use arguments, this could be trained in one line. All mentioned GNN variants above is not available for training. For an important notice, this task is residue (token) level classification therefore we deploy token level CNN, CNN_RNN and Transformer models.

Current version only supports Token_CNN, Token_CNN_RNN and Token_Transformer for target_encoding.

CNN Case

python train/SAbDab_Liberis.py --target_encoding Token_CNN --seed 7 --wandb_proj DeepProtein --lr 0.0001 --epochs 100

Case Study 1(f): A Framework for Antibody Developability Prediction (TAP)

Make sure that tdc is installed, if not

pip install PyTDC

Click here for the code!

### package import
import os, sys, argparse, torch, wandb


### Our library DeepProtein
from DeepProtein.dataset import *
import DeepProtein.utils as utils
import DeepProtein.PPI as models
from tdc.utils import retrieve_label_name_list
from tdc.single_pred import Develop

### Load TAP Dataset
label_list = retrieve_label_name_list('TAP')

data = Develop(name='TAP', label_name=label_list[0])
split = data.get_split()

train_antibody_1, train_antibody_2 = to_two_seq(split, 'train', 'Antibody')
valid_antibody_1, valid_antibody_2 = to_two_seq(split, 'valid', 'Antibody')
test_antibody_1, test_antibody_2 = to_two_seq(split, 'test', 'Antibody')

y_train, y_valid, y_test = split['train']['Y'], split['valid']['Y'], split['test']['Y']

train_TAP = list(zip(train_antibody_1, train_antibody_2, y_train))
valid_TAP = list(zip(valid_antibody_1, valid_antibody_2, y_valid))
test_TAP = list(zip(test_antibody_1, test_antibody_2, y_test))


### Train Valid Test Split
target_encoding = 'CNN'
train_protein_1, train_protein_2, train_target, train_protein_idx = collate_fn_ppi(train_TAP, graph=False, unsqueeze=False)
valid_protein_1, valid_protein_2, valid_target, valid_protein_idx = collate_fn_ppi(valid_TAP, graph=False, unsqueeze=False)
test_protein_1, test_protein_2, test_target, test_protein_idx = collate_fn_ppi(test_TAP, graph=False, unsqueeze=False)

train, _, _ = data_process(X_target=train_protein_1, X_target_=train_protein_2, y=train_target,
                               target_encoding=target_encoding,
                               split_method='random', frac=[0.99998, 1e-5, 1e-5],
                               random_seed=1)

_, val, _ = data_process(X_target=valid_protein_1, X_target_=valid_protein_2, y=valid_target,
                            target_encoding=target_encoding,
                            split_method='random', frac=[1e-5, 0.99998, 1e-5],
                            random_seed=1)

_, _, test = data_process(X_target=test_protein_1, X_target_=test_protein_2, y=test_target,
                            target_encoding=target_encoding,
                            split_method='random', frac=[1e-5, 1e-5, 0.99998],
                            random_seed=1)

### Load configuration for model
config = generate_config(target_encoding=target_encoding,
                         cls_hidden_dims=[1024, 1024],
                         train_epoch=20,
                         LR=0.0001,
                         batch_size=32,
                         )
config['binary'] = False
config['multi'] = False
torch.manual_seed(args.seed)
model = models.model_initialize(**config)


### Train our model
model.train(train_set, valid_set, test_set, batch_size=batch_size)

If you want to use structure learning methods such as graph neural network, please set the second parameters in the collate_fn_ppi() into True.

(b) If you wish to use arguments, this could be trained in one line. For GNN, only DGL_GCN, DGL_GAT and DGL_NeuralFP is available currently.

CNN Case

python train/TAP.py --target_encoding CNN --seed 7 --wandb_proj DeepProtein --lr 0.0001 --epochs 100

GNN Case

python train/TAP.py --target_encoding DGL_GCN --seed 7 --wandb_proj DeepProtein --lr 0.00001 --epochs 100

Case Study 1(g): A Framework for Repair Outcome Prediction (CRISPR)

Make sure that tdc is installed, if not

pip install PyTDC

Click here for the code!

### package import
import os, sys, argparse, torch, wandb


### Our library DeepProtein
from DeepProtein.dataset import *
import DeepProtein.utils as utils
import DeepProtein.ProteinPred as models
from tdc.utils import retrieve_label_name_list
from tdc.single_pred import Develop, CRISPROutcome

### Load CRISPR Leenay Dataset
label_list = retrieve_label_name_list('Leenay')

data = CRISPROutcome(name='Leenay', label_name=label_list[0])
split = data.get_split()

train_GuideSeq, y_train = list(split['train']['GuideSeq']), list(split['train']['Y'])
val_GuideSeq, y_valid = list(split['valid']['GuideSeq']), list(split['valid']['Y'])
test_GuideSeq, y_test = list(split['test']['GuideSeq']), list(split['test']['Y'])

train_CRISPR = list(zip(train_GuideSeq, y_train))
valid_CRISPR = list(zip(val_GuideSeq, y_valid))
test_CRISPR = list(zip(test_GuideSeq, y_test))


### Train Valid Test Split
target_encoding = 'CNN'
train_protein_1, train_target, train_protein_idx = collate_fn(train_CRISPR, graph=False, unsqueeze=True)
valid_protein_1, valid_target, valid_protein_idx = collate_fn(valid_CRISPR, graph=False, unsqueeze=True)
test_protein_1, test_target, test_protein_idx = collate_fn(test_CRISPR, graph=False, unsqueeze=True)

train, _, _ = data_process(X_target=train_protein_1, y=train_target,
                               target_encoding=target_encoding,
                               split_method='random', frac=[0.99998, 1e-5, 1e-5],
                               random_seed=1)
_, val, _ = data_process(X_target=valid_protein_1, y=valid_target,
                            target_encoding=target_encoding,
                            split_method='random', frac=[1e-5, 0.99998, 1e-5],
                            random_seed=1)

_, _, test = data_process(X_target=test_protein_1, y=test_target,
                            target_encoding=target_encoding,
                            split_method='random', frac=[1e-5, 1e-5, 0.99998],
                            random_seed=1)

### Load configuration for model
config = generate_config(target_encoding=target_encoding,
                         cls_hidden_dims=[1024, 1024],
                         train_epoch=20,
                         LR=0.0001,
                         batch_size=32,
                         )
config['binary'] = False
config['multi'] = False
torch.manual_seed(args.seed)
model = models.model_initialize(**config)


### Train our model
model.train(train_set, valid_set, test_set, batch_size=batch_size)

If you want to use structure learning methods such as graph neural network, please set the second parameters in the collate_fn() into True.

(b) If you wish to use arguments, this could be trained in one line. All mentioned GNN above is available currently.

CNN Case

python train/CRISPR.py --target_encoding CNN --seed 7 --wandb_proj DeepProtein --lr 0.0001 --epochs 100

GNN Case

python train/CRISPR.py --target_encoding DGL_GCN --seed 7 --wandb_proj DeepProtein --lr 0.00001 --epochs 100

Encodings

Thanks to DeepPurpose and dgllife, we could borrow some of the encodings from DeepPurpose. The newly added encodings are PAGTN, EGT and Graphormer which belong to graph transformer modules that are prevailing methods these years for encoding protein graphs.

Currently, we support the following encodings:

Protein Encodings	Description
CNN	Convolutional Neural Network on SMILES
CNN_RNN	A GRU/LSTM on top of a CNN on SMILES
Transformer	Transformer Encoder on ESPF
MPNN	Message-passing neural network
DGL_GCN	Graph Convolutional Network
DGL_NeuralFP	Neural Fingerprint
DGL_AttentiveFP	Attentive FP, Xiong et al. 2020
DGL_GAT	Graph Attention Network
PAGTN	Path Augmented Graph Transformer Network
Graphormer	Do Transformers Really Perform Bad, Ying et al.

Note that we've tried EGT, however, it would lead to memory error if we want to construct a large batched edge feature matrix therefore we ignore the implementation of EGT. This could be solved if applied to small graphs so it will be our future work.

Data

We provided the data under the folder DeepProtein/data (Besides TDC) and the folder data (TDC).

Data	Source	Form	Task
Fluorescence	PEER	LMDB	Protein Property Prediction
Solubility	PEER	LMDB	Protein Property Prediction
Stability	PEER	LMDB	Protein Property Prediction
Beta-lactamase	PEER	LMDB	Protein Property Prediction
SubCellular	PEER	LMDB	Protein Localization Prediction
SubCellular-Binary	PEER	LMDB	Protein Localization Prediction
PPI_Affinity	PEER	LMDB	Protein-Protein Interaction
Human_PPI	PEER	LMDB	Protein-Protein Interaction
Yeast_PPI	PEER	LMDB	Protein-Protein Interaction
IEDB	TDC	PICKLE	Antigen Epitope Prediction
PDB-Jespersen	TDC	PICKLE	Antigen Epitope Prediction
SAbDab-Liberis	TDC	PICKLE	Antibody Paratope Prediction
TAP	TDC	TAB	Antibody Developability Prediction
SAbDab-Chen	TDC	TAB	Antibody Developability Prediction
CRISPR-Leenay	TDC	TAB	CRISPR Repair Outcome Prediction

Cite Us

If you found this package useful, please cite our paper:

@article{xie2024deepprotein,
  title={DeepProtein: Deep Learning Library and Benchmark for Protein Sequence Learning},
  author={Xie, Jiaqing and Zhao, Yue and Fu, Tianfan},
  journal={arxiv},
  year={2024}
}

Name		Name	Last commit message	Last commit date
Latest commit History 240 Commits
DEMO		DEMO
DeepProtein		DeepProtein
data		data
docs		docs
figs		figs
plot		plot
result		result
train		train
train_sh		train_sh
.gitattributes		.gitattributes
.gitignore		.gitignore
.readthedocs.yaml		.readthedocs.yaml
LICENSE		LICENSE
MANIFEST.in		MANIFEST.in
Readme.md		Readme.md
requirements.txt		requirements.txt

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

[DeepProtein: Deep Learning Library and Benchmark for Protein Sequence Learning](https://arxiv.org/abs/2410.02023, NeurIPS 2024 AIDrugX Spotlight)

Applications in Protein Property Prediction, Localization Prediction, Protein-Protein Interaction, Antigen Epitope Prediction, Antibody Paratope Prediction, Antibody Developability Prediction, and more.

Introduction

Why DeepProtein?

What is DeepProtein?

News

Installation

Demos

Example

Case Study 1(a): A Framework for Protein Function (Property) Prediction

Case Study 1(b): A Framework for Protein Protein Interaction Prediction

Case Study 1(c): A Framework for Protein Localization Prediction

Case Study 1(d): A Framework for Antigen Epitope Prediction

Case Study 1(e): A Framework for Antibody Paratope Prediction

Case Study 1(f): A Framework for Antibody Developability Prediction (TAP)

Case Study 1(g): A Framework for Repair Outcome Prediction (CRISPR)

Encodings

Data

Cite Us

About

Releases

Packages

Contributors 3

Languages

License

jiaqingxie/DeepProtein

Folders and files

Latest commit

History

Repository files navigation

[DeepProtein: Deep Learning Library and Benchmark for Protein Sequence Learning](https://arxiv.org/abs/2410.02023, NeurIPS 2024 AIDrugX Spotlight)

Applications in Protein Property Prediction, Localization Prediction, Protein-Protein Interaction, Antigen Epitope Prediction, Antibody Paratope Prediction, Antibody Developability Prediction, and more.

Introduction

Why DeepProtein?

What is DeepProtein?

News

Installation

Demos

Example

Case Study 1(a): A Framework for Protein Function (Property) Prediction

Case Study 1(b): A Framework for Protein Protein Interaction Prediction

Case Study 1(c): A Framework for Protein Localization Prediction

Case Study 1(d): A Framework for Antigen Epitope Prediction

Case Study 1(e): A Framework for Antibody Paratope Prediction

Case Study 1(f): A Framework for Antibody Developability Prediction (TAP)

Case Study 1(g): A Framework for Repair Outcome Prediction (CRISPR)

Encodings

Data

Cite Us

About

Topics

Resources

License

Stars

Watchers

Forks

Releases

Packages 0

Contributors 3

Languages

Packages