Name | SMILES | Activation Status |
---|---|---|

Methanethiol | CS | 0 |

Ethanethiol | C(C)S | 0 |

Benzenethiol | C1(=CC=CC=C1)S | 1 |

S.No. | Parameter | Type | What it means |
---|---|---|---|

1 | n_tasks | Integer | Number of tasks |

2 | graph_conv_layers | List of integer | Width of channels for GCN layers. graph_conv_layers[i] gives the width of channel for the i-th GCN layer |

3 | activation | String | The activation function to apply to the output of each GCN layer |

4 | residual | Boolean | Whether to add a residual connection within each GCN layer |

5 | batchnorm | Boolean | Whether to apply batch normalization to the output of each GCN layer |

6 | dropout | Float | The dropout probability for the output of each GCN layer |

8 | predictor_hidden_feats | Integer | The size for hidden representations in the output MLP predictor |

9 | predictor_dropout | Integer | The dropout probability in the output MLP predictor |

10 | mode | String | The model type, ‘classification’ or ‘regression’ |

11 | number_atom_features | Integer | The length of the initial atom feature vectors |

12 | n_classes | Integer | The number of classes to predict per task (only used when mode is ‘classification’) |

13 | self_loop | Boolean | Whether to add self loops for the nodes, i.e. edges from nodes to themselves. When input graphs have isolated nodes, self loops allow preserving the original feature of them in message passing. |

S.No. | Parameter | Type | What it means |
---|---|---|---|

1 | n_tasks | Integer | Number of tasks |

2 | graph_attention_layers | List of integer | Width of channels per attention head for GAT layers. graph_attention_layers[i] gives the width of channel for each attention head for the i-th GAT layer. If both graph_attention_layers and agg_modes are specified, they should have equal length |

3 | n_attention_heads | Integer | Number of attention heads in each GAT layer |

4 | residual | Boolean | Whether to add a residual connection within each GAT layer |

5 | aggregation_modes | List of String | The way to aggregate multi-head attention results for each GAT layer, which can be either ‘flatten’ for concatenating all-head results or ‘mean’ for averaging all-head results. agg_modes[i] gives the way to aggregate multi-head attention results for the i-th GAT layer |

6 | dropout | Float | The dropout probability for the output of each GAT layer |

7 | alpha | Float | A hyperparameter in LeakyReLU, which is the slope for negative values |

8 | predictor_hidden_feats | Integer | The size for hidden representations in the output MLP predictor |

9 | predictor_dropout | Integer | The dropout probability in the output MLP predictor |

10 | mode | String | The model type, ‘classification’ or ‘regression’ |

11 | number_atom_features | Integer | The length of the initial atom feature vectors |

12 | n_classes | Integer | The number of classes to predict per task (only used when mode is ‘classification’) |

13 | self_loop | Boolean | Whether to add self loops for the nodes, i.e. edges from nodes to themselves. When input graphs have isolated nodes, self loops allow preserving the original feature of them in message passing. |

S.No. | Parameter | Type | What it means |
---|---|---|---|

1 | n_tasks | Integer | Number of tasks |

2 | n_atom_feat | Integer | Number of features per atom |

3 | n_graph_feat | Integer | Number of features for atom in the graph |

4 | n_outputs | Integer | Number of features for each molecule |

5 | layer_sizes | List of Integer | List of hidden layer size(s) in the propagation step: length of this list represents the number of hidden layers, and each element is the width of corresponding hidden layer |

6 | dropout | Float | Dropout probability, applied after each propagation step and gather step |

7 | layer_sizes_gather | List of Integer | List of hidden layer size(s) in the gather step |

8 | n_classes | Integer | The number of classes to predict (only used in classification mode) |

9 | uncertainty | Boolean | If True, include extra outputs and loss terms to enable the uncertainty in outputs to be predicted |

10 | mode | String | The model type, ‘classification’ or ‘regression’ |

S.No. | Parameter | Type | What it means |
---|---|---|---|

1 | n_tasks | Integer | Number of tasks |

2 | num_layers | Integer | Number of graph neural network layers, i.e. number of rounds of message passing |

3 | num_timesteps | Integer | Number of time steps for updating graph representations with a GRU |

4 | graph_feat_size | Int | Size for graph representations |

5 | number_atom_features | Integer | The length of the initial atom feature vectors |

6 | dropout | Float | Dropout probability |

7 | n_classes | Integer | The number of classes to predict per task (only used when mode is ‘classification’) |

8 | self_loop | Boolean | Whether to add self loops for the nodes, i.e. edges from nodes to themselves. When input graphs have isolated nodes, self loops allow preserving the original feature of them in message passing. |

SMILES | Activation Status | Probability of Class 0 | Probability of Class 1 |
---|---|---|---|

CC(CCC=C(C)C)CC=O | 1 | 0.1934605 | 0.8065395 |

C1CCCCCC(=O)CCCC1 | 0 | 0.9994111 | 0.0005889 |

CCC(=O)C1=CC=CC=C1 | 1 | 0.4913514 | 0.5086485 |

How to run pre-trained models

There are two ways users can use saved models to get the predictions on query data:

1. Users can use **GetPrediction_From_Checkpoint_deepGraphh.ipynb ** in the Get Prediction folder on our GitHub.

2. Users can use the Google Collaboratory Notebook

The prediction is based on Graph-based Deep Learning models.

No.

We understand that it can be a little time-consuming, considering the high computations. Please try to be patient, and do not close the tabs, however, you can change the tabs.

deepGraphh has provided example datasets for user’s reference.