Analysis notebooks for the manuscript “DeepSCENIC: transfer learning from sequence-to-function models enables causal gene regulatory network inference”
DeepSCENIC combines representations learned by sequence-to-function models with single-cell RNA and ATAC measurements to infer transcription factor-to-regulatory element (TF–RE), regulatory element-to-target gene (RE–TG), and transcription factor-to-target gene (TF–TG) relationships. This repository contains the manuscript and the annotated analysis notebooks used for Figures 2–4. It does not contain the DeepSCENIC training library; that code is maintained in the DeepSCENIC repository.
| Path | Description |
|---|---|
notebooks/Figure2.ipynb |
Figure 2 - ENCODE cell-line validation. |
notebooks/Figure3.ipynb |
Figure 3 — Melanoma perturbations. |
notebooks/Figure4.ipynb |
Figure 4 — Cross-species cortical regulation. |
The input data and trained model outputs to reproduce the analysis are not included in this repository. They can be downloaded from DOI: 10.5281/zenodo.20560125, or from their original publications.