CREsted paper

Scripts and notebooks for analyses in "CREsted: modeling genomic and synthetic cell type-specific enhancers across tissues and species". For CREsted source code & documentation please visit: the CREsted repository and documentation. The manuscript's Data Availability section contains a detailed description of all public and new data used for all analyses.

The analyis data required for running these notebooks can be downloaded at: https://resources.aertslab.org/CREsted/manuscript_data/

Figures

Notebooks and scripts are subdivided in separate folders.

Figure 1

CREsted package overview.

No analyses were performed for this figure.

Figure 2

CREsted provides detailed insights into enhancer codes of mouse cortical cell types.

Model training and evaluation and enhancer code analysis notebooks for this dataset are found as tutorial notebooks in the CREsted repository. Additional notebooks for main figure reproducability are found in this folder.

Figure 3

A CREsted human PBMC model captures validated TFBS.

Figure 4

CREsted identifies high similarity between mesenchymal-like enhancer codes in cancer.

Figure 5

CREsted-trained models outperform large, pre-trained models on cell type-specific chromatin accessibility predictions.

A general model training notebook can be found for this dataset in the CREsted repository.

Figure 6

CREsted designs enhancers specifically active in targeted cell types of a developing zebrafish.