Skip to main content

Meta2Vec, an embedding for Metabolomics study

Project description


meta2vec is a Python package for metabolite embedding, which allows for the representation of metabolites in a vector space.

meta2vec package contains three modules:

  • distance: provides functions for calculating the similarity distance between two metabolites using their embeddings.
  • utils: provides helper functions for working with HMDB (Human Metabolome Database) dataset.
  • visualize: provides functions for visualizing the embeddings using UMAP.


meta2vec package can be installed via pip:

pip install meta2vec


Here is a brief overview of how to use meta2vec package:

Load Pre-trained Embeddings

from meta2vec import from_pretrained

# Load the TransE pre-trained embeddings
hmdb_embeddings = from_pretrained(model_type="TransE", embedding_dir="embedding")

The above code loads the pre-trained HMDB embeddings from the embedding directory. If the embeddings are not already present in the directory, the function will download them from the GitHub repository.

Calculate Distance between Two Metabolites

from meta2vec import hmdb_id_cosine_distance, hmdb_id_euclidean_distance

# Calculate the cosine distance between two HMDB IDs
cosine_distance = hmdb_id_cosine_distance(hmdb_embeddings, "HMDB0000001", "HMDB0000002")

# Calculate the Euclidean distance between two HMDB IDs
euclidean_distance = hmdb_id_euclidean_distance(hmdb_embeddings, "HMDB0000001", "HMDB0000002")

Find Most Similar Metabolites

from meta2vec import most_similar

# Find the most similar HMDB IDs to a given HMDB ID
similar_compounds = most_similar(hmdb_embeddings, "HMDB0000001")

Visualize Embeddings

from meta2vec import visualize_umap

# Visualize the embeddings using UMAP


If you would like to contribute to this project, please contact

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

Meta2Vec- (13.6 kB view hashes)

Uploaded source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page