Hi everybody,
I am wondering if anyone has tried to impute the protein expression values based on single nuclei / cell RNA-seq data. Something like transferring the knowledge from CITE-seq or other experiments where the protein data has not been measured, but the information could be transferred from available other datasets to predict the protein expression value.
Hi Anatoly,
This is a very interesting use case of single-cell modelling which I am interested in as well! I As far as my knowledge goes there is no module in the scvi-tools library that is specifically designed for this task. However, during my research I came across the scMM model (GitHub - kodaim1115/scMM) which performs cross-modality generation. The model was also explicitly tested on CITE-seq data so this might be in the direction you were thinking?
1 Like