binchenlab/InsilicoCell

Data file description:

Two different scenarios were considered for prediction performance evaluation: sample-level holdout validation and entity-level holdout validation.

Sample-level holdout validation:

This scenario was used for evaluating model performance based on sample-level holdout validation. In this scenario, all the input samples in the test set were never seen during training. One sample corresponds to one label.

Data files include:
"drug-induced_gene_expression_change_sample-level_holdout_test_set.csv"
"drug-protein_binding_sample-level_holdout_test_set.csv"
"TF-gene_association_sample-level_holdout_test_set.csv"
"drug_sensitivity_sample-level_holdout_test_set.csv"
"gene_effect_score_sample-level_holdout_test_set.csv"
"gene_mutation_sample-level_holdout_test_set.csv"
"CNV_sample-level_holdout_test_set.csv"

Entity-level holdout validation:

This scenario was used for evaluating model prediction performance on input samples containing unseen entities, such as unseen cell lines and unseen compounds, which did not appear in the training set and were viewed by the model as new cell lines and new compounds in the test set to predict on.

Data files include:
"drug-induced_gene_expression_change_entity-level_holdout_test_set.csv"
"drug-protein_binding_entity-level_holdout_test_set.csv"
"TF-gene_association_entity-level_holdout_test_set.csv"
"drug_sensitivity_entity-level_holdout_test_set.csv"
"gene_effect_score_entity-level_holdout_test_set.csv"
"gene_mutation_entity-level_holdout_test_set.csv"
"CNV_entity-level_holdout_test_set.csv"

Model prediction performance is generally better under sample-level validation setting compared to the more stringent entity-level validation setting. We will release the training set after the acceptance of the paper.

Dataset field discription:

Column	Description	Used by
SMILES	Drug structure in SMILES notation	Drug-induced expression, drug-protein binding, drug sensitivity
cell_iname	Cell line name	Drug-induced expression, drug sensitivity, gene effect, mutation, CNV
gene_name	Gene name	Drug-induced expression, TF-gene, gene effect, mutation, CNV
target_sequence	Protein amino acid sequence	Drug-protein binding, TF-gene
time_h	Treatment duration in hours	Drug-induced expression
dose_uM	Treatment concentration in μM	Drug-induced expression
task_id	Name of the task	All seven tasks
label	Ground truth label corresponding to each input sample (continuous values for regression tasks, binary values for classification tasks)	All seven tasks

Downloading data:

Check our github repo on how to download the data, run InsilicoCell for prediction and evaluation.