INDIGENA:

Dependencies

• Groovy 4.0.26
• Java 8+

Python KGE Methods

• Python 3.10
• mowl
• pykeen
• torch
• wandb
• pandas
• tqdm
• click

Installation

git clone https://github.com/bio-ontology-research-group/indigena.git
cd indigena/
conda env create -f environment.yml
conda activate indigena

Usage

1. Uncompress UPheno file

First, extract the data archive:

cd data
gunzip upheno.owl.gz

2. Semantic Similarity Baselines (Groovy)

These baseline methods compute ontology-based semantic similarity between genes and diseases using the SLIB library.

Run with default parameters:

groovy semantic_similarity.groovy -r data -fold 0

Run with custom parameters:

groovy semantic_similarity.groovy -r data -ic resnik -pw resnik -gw bma -fold 0

Run SimGIC variant:

groovy semantic_similarity_simgic.groovy -r data -ic resnik -fold 0

Parameters:

• -r, --root_dir: Data directory (default: data)
• -ic, --ic_measure: Information content measure (resnik, sanchez)
• -pw, --pairwise_measure: Pairwise measure (resnik, lin)
• -gw, --groupwise_measure: Groupwise measure (bma, bmm)
• -fold: Cross-validation fold number (default: 0)

Output: Results saved to data/baseline_results/

Evaluate results:

python evaluate_sem_sim.py data/baseline_results/<results_file>

3. Knowledge Graph Embeddings (Python)

This approach uses mOWL to project ontology into triples and PyKEEN to train KGE models. We use W&B to track experiments. Therefore, before running kge.py, change the entity name in wandb.init to you W&B username.

Run basic KGE model:

python kge_transe.py --fold 0 --mode inductive --graph2 --no_sweep

Run with hyperparameters:

python kge_transd.py --fold 0 --mode inductive \
  --embedding_dim 100 --batch_size 128 --learning_rate 0.001 \
  --graph2 --no_sweep

Parameters:

You can look at the hyperparameters in each script. They usually look like this:

• --fold: Cross-validation fold number
• --model_name: KGE model (transe, transd, convkb)
• --mode: Evaluation mode (inductive, transductive)
• --graph2: Add gene-phenotype edges
• --graph3: Add disease-phenotype edges
• --graph4: Add gene-disease training edges
• --embedding_dim: Embedding dimensions
• --batch_size: Training batch size
• --learning_rate: Learning rate
• --only_test: Only test existing model (skip training)
• --description: Weights & Biases run description
• --no_sweep: Disable W&B sweep mode
• --pretrained_model: Features to initialize ConvKB embeddings (transe, transd).

Problems running the models?

Please create a Github issue or PR!