cfDecon is an efficient deep-learning framework for cell-free DNA deconvolution analysis that operates at the resolution of individual reads. The tool uses a multichannel autoencoder core module and iterative refinement to estimate cell-type proportions and generate condition-aware cell-type-specific methylation profiles.
Example script can be found in ./scripts/run.sh
This paper is currently under review. For questions about the code, please:
-
Open an issue on GitHub
-
Propose an issue through emails: lijiayi03531@gmail.com and yixuan000518@gmail.com