Combining counts of 2 genes into 1 gene - spatial

[LukeJ89]

Hi,
I have some spatial data (Visium) that I have merged and am comparing gene expression between two inflammatory conditions. There is a particular gene that is significantly expressed in one of the conditions that I am interested in, let's call it Gene1. However, there is another gene that codes for the exact same protein, call it Gene1B, and this is significantly expressed in a the other condition, although not as greatly, which could impact the overall differential expression for that protein.
I was wondering if it were possible to combine the counts of Gene1 and Gene1B to make a new count called GF1, and run DGE analysis as this will give me a better understanding of the gene expression level for that one protein?

`

KP001B_spatial <-Load10X_Spatial("Z:/Spatial_Transcriptomics/KP001/KP001_B/outs/",
filename = "filtered_feature_bc_matrix.h5",
assay = "Spatial",
slice = "KP001B",
filter.matrix = TRUE,
to.upper = FALSE)

AD1_spatial <-Load10X_Spatial("Z:/Spatial_Transcriptomics/AD/AD_1/outs/",
filename = "filtered_feature_bc_matrix.h5",
assay = "Spatial",
slice = "AD_1",
filter.matrix = TRUE,
to.upper = FALSE)

Healthy1_spatial <-Load10X_Spatial("Z:/Spatial_Transcriptomics/Healthy/H_1/outs/",
filename = "filtered_feature_bc_matrix.h5",
assay = "Spatial",
slice = "H_1",
filter.matrix = TRUE,
to.upper = FALSE)

All_skin <- c(KP001B_spatial, AD1_spatial, Healthy1_spatial)
merged.data <- merge(x = All_skin[[1]], y = All_skin[-1], merge.data = T)
merged.data <- SCTransform(merged.data, assay = "Spatial", vars.to.regress = "sample.id")
merged.data <- RunPCA(merged.data, assay = "SCT")
merged.data.hm <- RunHarmony(merged.data, assay.use = "SCT", project.dim = F, group.by.vars = "sample.id")
merged.data.hm <- merged.data.hm %>%
RunUMAP(reduction = "harmony", dims = 1:40) %>%
FindNeighbors(reduction = "harmony", dims = 1:40) %>%
FindClusters(resolution = c(0.4, 0.5, 0.6, 0.7)) %>%
identity()
Idents(merged.data.hm ) <- 'SCT_snn_res.0.5'`

[dcollins15]

The easiest way to do this is probably to read in the counts matrices via Read10X_h5, sum the relevant columns in the matrices, instantiate a Seurat object using CreateSeuratObject and then add the images in manually after reading them in via Read10X_Image - something along the lines of:

counts <- Read10X_h5("Z:/Spatial_Transcriptomics/KP001/KP001_B/outs/filtered_feature_bc_matrix.h5")
counts_GF1 <- counts["Gene1", ] + counts["Gene1B", ]
counts <- rbind(counts, GF1 = counts_GF1)
counts <- counts[!rownames(counts) %in% c("Gene1", "Gene1B"), ]

KP001B_spatial <- CreateSeuratObject(counts, assay = "Spatial")

image <- Read10X_Image(
  "Z:/Spatial_Transcriptomics/KP001/KP001_B/outs/spatial",
  assay = "Spatial",
  slice = "KP001B",
  filter.matrix = TRUE
)

image <- image[Cells(KP001B_spatial)]
KP001B_spatial[["KP001B"]] <- image