Assign cell types for Gawad data

[allyhawkins]

In order to better evaluate the integration methods on scpca datasets, we want to incorporate cell type information where possible. We want to use the Gawad data which has matching CITE-seq data so that we can use the CITE-seq data to aid in cell type annotation.

To tackle this we should start with the 5 Gawad libraries that are already included in sample-info/scpca-processed-libraries.tsv. These libraries have already been processed through downstream analyses so we can start with those SCE objects to annotate cell types.

In doing this we will also need to normalize the CITE-seq data before using it for validation. We might consider trying automated cell type methods like Azimuth and then using the CITE-seq markers to help validate or identify tumor vs. non-tumor cells. This might take a little bit of manual exploration at first so we might further break up this issue once we get started.

[allyhawkins]

After our discussion, we've decided to explore using SingleR to annotate cell types in the Gawad data. This should be done through an exploratory notebook where we read in the normalized Gawad data (objects that have both normalized RNA and CITE-seq data) and then use SingleR for cell type annotation.