Update deconvolution plots

[lcolladotor]

Hi Nick,

From our meeting with Sang Ho and Kristen today, we have a few requests for the spot deconvolution plots.

• Subset to only the AD samples in the gray matter spots https://github.com/LieberInstitute/Visium_SPG_AD/tree/master/plots/21_spot_deconvo/explore_results. That is, use

  Visium_SPG_AD/code/22_gene_pairs_path/01_gene_pairs_path.R

  Lines 30 to 31 in 69108f2

  	## Subset to just gray matter spots only in the AD samples
  	spe <- spe[, spe$BayesSpace_harmony_k02 != 2 & spe$diagnosis == "AD"]

  .
• For https://github.com/LieberInstitute/Visium_SPG_AD/blob/master/plots/21_spot_deconvo/explore_results/pathology_barplots_inverted.pdf and https://github.com/LieberInstitute/Visium_SPG_AD/blob/master/plots/21_spot_deconvo/explore_results/pathology_boxplots.pdf use the pathology colors we have. You can re-use

  Visium_SPG_AD/code/22_gene_pairs_path/01_gene_pairs_path.R

  Lines 205 to 209 in 69108f2

  	## Identify the colors for the pathology groups
  	cols <- unique(co_expr_se$path_groups_colors)
  	names(cols) <- sapply(cols, function(x) {
  	names(co_expr_se$path_groups_colors[co_expr_se$path_groups_colors == x])[1]
  	})

  + scale_fill_manual() like at

  Visium_SPG_AD/code/22_gene_pairs_path/01_gene_pairs_path.R

  Line 177 in 69108f2

  scale_fill_manual(values = discrete_cell_palette) +

  .
• If you can remake all 3 plots for all spots (and/or just the white matter ones), those extra plots could be useful for just internal purposes. This item is of lower priority though.

Thanks,
Leo