Steps to add preliminary PRECAST results to https://libd.shinyapps.io/Spatial_NAc

[lcolladotor]

This is something we left pending from February 2024. Here are more detailed steps Nick. Sign up for a DSgs with me if you have any questions.

• Add the preliminary PRECAST clustering results for k = 2 to 28 to the SPE object used for the spatialLIBD app. This involves updating https://github.com/LieberInstitute/spatial_NAc/blob/main/code/06_deploy_app/prep_spe.R with code you have at https://github.com/LieberInstitute/spatial_NAc/blob/main/code/10_precast/02-explore_results.R.
• Choose a preliminary PRECAST cluster k in coordination with others (likely Svitlana). I recommend a small value of k (under 10); see last bullet point. We know that ultimately these results will be replaced by others computed by Prashanthi et al. Aka, the scripts you write for resolving this issue will be further adapted by Prashanthi et al.
• Using code similar to https://github.com/LieberInstitute/spatial_NAc/blob/main/code/06_deploy_app/prep_spe.R (but with the actual counts), create a pseudobulked object using spatialLIBD::registration_pseudobulk() (also save it under the 06_deploy_app dir since we'll use it for spatialLIBD). As noted on Slack at https://jhu-genomics.slack.com/archives/C02LG5XSANT/p1707847605977579, you'll want to adapt code from https://github.com/LieberInstitute/Visium_SPG_AD/blob/master/code/11_grey_matter_only/01_create_pseudobulk_data.R (like computing PCs, editing the rowData(sce_pseudo)$gene_search and other things there.
• Using the pseudobulked object, adapt https://github.com/LieberInstitute/Visium_SPG_AD/blob/master/code/11_grey_matter_only/03_model_pathology.R.
• Adapt https://github.com/LieberInstitute/spatial_NAc/blob/main/code/06_deploy_app/app.R with code from https://github.com/LieberInstitute/Visium_SPG_AD/blob/e13e1a3604aa2dc43d76fa6923541ef403b0b13d/code/05_deploy_app_wholegenome/app.R#L35-L81. If you chose a large k, then you'll need to adapt code from https://github.com/LieberInstitute/spatialDLPFC/blob/main/code/deploy_app_k16/subset.R instead.
• Check https://github.com/LieberInstitute/spatial_NAc/blob/main/code/06_deploy_app/app.R locally (note that the working directory has to be set to the location of app.R to simulate the behavior from shinyapps.io; aka, do not use here::here() on that script!).

Once you are done, let me know and we can update the app by running https://github.com/LieberInstitute/spatial_NAc/blob/main/code/06_deploy_app/deploy.R