methylVI is a generative model designed to facilitate analyses of single-cell methylomic data generated from bisulfite sequencing platforms (e.g. snmC-seq, sciMET, etc.). methylVI assumes that the measurements from individual cytosines have already been aggregated into a cell by genomic region (e.g. gene bodies, promoter regions, etc.) matrix, and the model accepts CpG and/or CpH methylation measurements as input.

To install the latest version of methylVI via pip

pip install methyl-vi

Installation should take no more than 5 minutes.

• Embed high-dimensional methylation profiles into a lower-dimensional latent space.
• Produced denoised methylation profiles (demo notebook).
• Integrate data collected in different experimental conditions (e.g. with different sequencing platforms, at different time points, etc.) (demo notebook).
• Transfer cell type labels between datasets using a semi-supervised learning (demo notebook)
• Map newly collected query datasets to previously assembled reference atlases (demo notebook)

If you find methylVI useful for your work, please consider citing

@inproceedings{
    methylVI,
    title={A deep generative model of single-cell methylomic data},
    author={Ethan Weinberger and Su-In Lee},
    booktitle={NeurIPS 2023 Generative AI and Biology (GenBio) Workshop},
    year={2023},
    url={https://openreview.net/forum?id=Mg2DM0F3AY}
}