DeepDDS

class DeepDDS(*, context_channels, context_hidden_dims=None, drug_channels=66, drug_gcn_hidden_dims=None, drug_mlp_hidden_dims=None, context_output_size=32, fc_hidden_dims=None, dropout=0.5)[source]

Bases: chemicalx.models.base.Model

An implementation of the DeepDDS model from [wang2021].

This implementation follows the code on github where the paper and the code diverge. https://github.com/Sinwang404/DeepDDs/tree/master

See also

This model was suggested in https://github.com/AstraZeneca/chemicalx/issues/19

wang2021

Wang, J., et al. (2021). DeepDDS: deep graph neural network with attention mechanism to predict synergistic drug combinations. arXiv, 2107.02467.

Methods Summary

forward(context_features, molecules_left, …)

Run a forward pass of the DeeDDS model.

unpack(batch)

Return the context features, left drug features and right drug features.

Methods Documentation

forward(context_features, molecules_left, molecules_right)[source]

Run a forward pass of the DeeDDS model.

Parameters
  • context_features (FloatTensor) – A matrix of cell line features

  • molecules_left (PackedGraph) – A matrix of left drug features

  • molecules_right (PackedGraph) – A matrix of right drug features

Return type

FloatTensor

Returns

A vector of predicted synergy scores

unpack(batch)[source]

Return the context features, left drug features and right drug features.