DeepPBMC#
The DeepPBMC model is a peak regression model trained to predict genomic region accessibility over seven cell types from human PBMC data.
The model is pre-trained on a set of 278K consensus peaks, followed by fine-tuning on 51K cell type-specific peaks.
The model is a CNN multiclass regression model using the dilated_cnn() architecture.
Details of the data and the model can be found in the original publication.
Citation
Kempynck, N., De Winter, S. et al. CREsted: modeling genomic and synthetic cell-type-specific enhancers across tissues and species. Nature Methods (2026). https://doi.org/10.1038/s41592-026-03057-2
Data source
De Rop, F.V. et al. Systematic benchmarking of single-cell ATAC-sequencing protocols. Nature Biotechnology (2024). https://doi.org/10.1038/s41587-023-01881-x
Usage#
1import crested
2import keras
3
4# download model
5model_path, output_names = crested.get_model("DeepPBMC")
6
7# load model
8model = crested.utils.load_model(model_path)
9
10# make predictions
11sequence = "A" * 2114
12predictions = crested.tl.predict(sequence, model)
13print(predictions.shape)