Menu ▾ ▴

Tree [94cfa6] main /
 History

HTTPS git://
HTTPS access

File Date Author Commit
 LICENSE 2021-05-21 Gabriela Merino Gabriela Merino [395061] Initial commit
 README.md 2021-05-21 Gabriela Merino Gabriela Merino [94cfa6] Create Readme.md

Read Me

Predicting protein functions using DeeProtGO

This readme contains a detailed example of a case-of-use of DeeProtGO, a deep learning model for automatic function prediction by integrating protein knowledge from multiple databases.

G.A. Merino, R. Saidi, D.H. Milone, G. Stegmayer, M. Martin. Hierarchical deep learning for predicting GO annotations by integrating heterogeneous protein knowledge, XXX, 2021,.

The following sections contain the steps and the libraries used in each of them. All datasets and libraries are open sourced and free to use. If you use any of the following in your pipelines, please cite them properly.

1. Layout

This worklfow describes the steps required to predict GO terms for proteins called No-knowledge (NK): Proteins that do not have experimental annotations in any of the GO sub-ontologies at a reference time, but have accumulated at least one GO term with an experimental evidence code during a growth period.
Data used for developing and evaluating the proposed models were obtained from different protein-knowledge databases, considering the proteins provided in the CAFA3 challenge. For CAFA3, $t_{-1}$ is when the challenge was released (09/2016), in which the sets of training and target proteins were provided to the participants; $t_0$, the deadline for participants submissions of the predictions for the target proteins (02/2017); and $t_1$ is when benchmark proteins were collected for assessment (11/2017). Thus, the CAFA3 benchmark dataset is composed of those target proteins that have, at least, one new functional annotation added during the growth period between $t_0$ and $t_1$.
Since DeeProtGO was developed to learn new annotations gained during a time gap, it is trained here with proteins that gained GO terms during the growth period defined between $t{-1}$ and $t_0$, and evaluated on NK proteins of the CAFA3 benchmark dataset.

Protein function prediction
│
├── data             -> All data for preparing files for DeeProtGO training and testing.
│   │
│   ├── extras       -> Files useful for datasets preparation.
│   │   │
│   │   └── gene_ontology_edit.obo        -> GO file with the ontology used for CAFA3.
│   │ 
│   ├── intermediate -> Intermediate files generated during training and benchmark datasets preparation. 
│   │   │
│   │   ├── LevDist_negProteins.tab               -> Edit distance of negative and positive proteins of the training set.
│   │   │
│   │   ├── LevDist_negProteinsBenchTrain.tab     -> Edit distance of negative proteins of the benchmark set with positive
│   │   │                                            proteins of the training set.
│   │   ├── LevDist_posProteins.tab               -> Edit distance between positive proteins of the training set.
│   │   │
│   │   ├── LevDist_posProteinsBenchTrain.tab     ->  Edit distance of prositive proteins of the benchmark set with positive
│   │   │                                            proteins of the training set.
│   │   ├── propAnnot_Bench_Euka_BP.tab           -> Propagated GO annotations gained between T_0 and T_1 for benchmark 
│   │   │                                            proteins.   
│   │   └── propAnnot_Train_Euka_BP.tab           -> Propagated GO annotations gained between T_{-1} and T_0 for training 
│   │                                                proteins. 
│   ├── processed    -> Datasets used for training and testing DeeProtGO when predicting annotations for NK proteins. 
│   │   │
│   │   ├── Benchmark  -> Data for NK proteins from eukarya organisms in CAFA3 benchmark for BP.
│   │   │       │
│   │   │       ├── Emb_BP_Euka.h5          -> SeqVec embeddings.
│   │   │       │
│   │   │       ├── LevSim_BP_Euka.h5       -> Proteins similarity based on edit distance.
│   │   │       │
│   │   │       ├── NegEntries_Euka_BP.tab  -> UniProt entry names of negative proteins.
│   │   │       │
│   │   │       ├── netOut_BP_Euka.h5       -> One-hot enconding matrix representing GO BP terms of benchmark proteins.
│   │   │       │
│   │   │       ├── PosEntries_Euka_BP.tab  -> UniProt entry names of positive proteins.
│   │   │       │
│   │   │       └── Taxon_BP_Euka.h5        -> One-hot enconding matrix representing proteins taxon.
│   │   │       
│   │   └── Training   -> Data used for training DeeProtGO to predict GO BP terms of  NK proteins from eukarya organisms.
│   │           │
│   │           ├── Emb_BP_Euka.h5                      -> SeqVec embeddings.
│   │           │
│   │           ├── GOTermsNetPropagatedNK_Euka_BP.tab  -> Proteins sequences obtained from .
│   │           │
│   │           ├── GOTermsPropRel_Euka_BP_train.tab    -> One-hot enconding of relationships between GO terms, used for scores │   │           │                                          propagation.
│   │           ├── LevSim_BP_Euka.h5                   -> Proteins similarity based on edit distance.
│   │           │
│   │           ├── NegEntries_Euka_BP.tab              -> UniProt entries of negative proteins.
│   │           │
│   │           ├── netOut_BP_Euka.h5                   -> One-hot enconding matrix representing GO BP terms of benchmark
│   │           │                                           proteins.
│   │           ├── PosEntries_Euka_BP.tab              -> UniProt entries of positive proteins.
│   │           │
│   │           └── Taxon_BP_Euka.h5                    -> One-hot enconding matrix representing proteins taxon.
│   │
│   └── raw  -> Metadata of training and benchmark proteins
│       │
│       ├── benchmarkNKEukaBPInfo.tab     -> Data of benchmark proteins
│       │
│       └── trainingNKEukaBPInfo.tab      -> Data of training proteins
|                          
│
├── examples        -> Files obtained when running the DeeProtGO example provided here
│   │
│   └── train_NK_EUKA_BP       -> Files obtained during training of DeeProtGO for predicting BP terms of NK proteins
│       │                                           
│       ├── DeeProtGO_PSD_Emb_Taxon_Euka_BP_NK.pt    -> Trained DeeProtGO model.
│       │                                           
│       ├── PSD_Emb_Taxoncheckpoint.pt               -> Last DeeProtGO model obtained during training.
│       │                                           
│       ├── test_13052021.log                        -> DeeProtGO performance on test partition of the training dataset.
│       │                                            
│       ├── train_13052021.log                       -> DeeProtGO loss on train partition of training the dataset.
│       │
│       └── valid_13052021.log                       -> DeeProtGO performance on validation partition of the training dataset.
|                        
│
├── scripts         -> Files obtained when running the DeeProtGO example provided here
│   │
│   ├── dataPreparation.ipynb      -> Python notebook for preparing the data for training and testing DeeProtGO
│   │
│   ├── DeeProtGOTesting.ipynb     -> Python notebook for training DeeProtGO
│   │
│   └── DeeProtGOTraining.ipynb    -> Python notebook for testing DeeProtGO
│
├── src             -> Source files of DeeProtGO
    │
    ├── dataloader.py      -> Class and methods for loading the data
    │
    ├── DNN.py             -> Deep neural network behind DeeProtGO
    │
    ├── DNNModel.py        -> DeeProtGO model and its functions
    │
    ├── earlyStop.py       -> Method for early stopping
    │
    ├── logger.py          -> Class and methods for saving performance measures and model loss during training and testing.
    │
    └── sampler.py         -> Class and methods for generating random batchs for being used during model training.

2. Data preparation

Note: All the results generated in this section are provided in data/intermediate and data/processed directories.

The data preparation notebook is provided with instructions to build all input and output data required for training DeeProtGO for predicting BP terms for NK proteins and for testing it with CAFA3 benchmark proteins. Doing so may take several hours.

3. Model training

Note: All the results generated in this section are provided in the examples/train_NK_EUKA_BP directory.

The DeeProtGO training notebook is provided with all the steps required to train DeeProtGO for predicting BP terms for NK proteins.

4. Model testing on CAFA3 benchmark

The DeeProtGO testing notebook is provided with all the steps required to evaluate DeeProtGO when predicting BP terms for NK proteins of eukarya organisms in the CAFA3 benchmark dataset.