Ethan Weinberger

Reference mapping with MuData-based models

2

(Addresses #2393) While certain multimodal models (e.g. totalVI) can be used with either `AnnData` or `MuData` objects, currently the `ArchesMixin` class only supports `AnnData`-based models and so reference mapping cannot...

Multimodal reference mapping with MuData objects

3

In the current (1.0.4) version of scvi-tools, multimodal models (e.g. totalVI) can be set up and trained with either AnnData or MuData objects. Thus, it would be nice to be...

Performance without histology better than with histology

1

Hi Jian, First of all, thanks for the awesome tool! I (along with @symoon) been investigating using SpaGCN for one of our projects. We're successfully able to reproduce the results...

Save per-modality feature names for MuData models

3

Current MuData-based models save their `var_names` as a single list, with no information on which feature came from which modality. This is problematic for certain downstream features (e.g. reference mapping...

SemiSupervisedDataSplitter not properly separating unlabelled cells into train/val/test

1

When splitting datasets into train/validation/test sets, I would expect that the union of train/validation/test samples should be equal to the full dataset. Moreover, data points should only be present once...

Add notebook for downloading McFarland 2020 Figure 1 data

5

This PR adds a Jupyter notebook to download the data from McFarland et al., 2020 used to produce Figure 1 (i.e., response to idasanutlin and control DMSO for different cell...

Implement methylVI

1

This PR implements methylation variational inference (methylVI) for modeling methylation data from single-cell bisulfite sequencing (scBS-seq) experiments, described originally in [Weinberger and Lee (2024).](https://openreview.net/forum?id=Mg2DM0F3AY). In short, methylVI assumes that counts...