Batch Similarity

Compute similarity between two sets of molecules: either centroid (pharmacophore) or batch (enrichment) 'centroid' mode: computes mean embedding (centroid) for the first and second set of molecules and returns its distance 'batch' mode: means it computes similarities of max 10K randomly sampled pairs and averages the result

Parameters:

• smiles1 and smiles2 (List[str]): Lists of smiles.
• similarity_metric (EmbNames): The model and similarity metric to use for the calculation.
• distance_type (DistanceType): The type of vector distance to use for the similarity calculation.
• sim_mode (SearchModes): The mode of similarity (default is "centroid", another is "batch").

Returns:

• A dictionary containing the molecules of the nearest cluster, possibly with a message pointing to the path where to find embeddings

Query Parameters

[  "CC(=O)Oc1ccccc1C(=O)O",  "CC(=O)Oc1ccncc1C(=O)O",  "CC(=O)Oc1cnccc1C(=O)O"]

[  "CC=O",  "CCC",  "CCCC=O"]

Response Body