normalized score for SpatialFeaturePlot

Question

[Omics data mining]

Hello everyone

I am working on spatial transcriptome data. I am following some tutorial (https://nbisweden.github.io/workshop-scRNAseq/labs/compiled/seurat/seurat_07_spatial.html#Subset_ST_for_cortex) for the QC, normalization and integration of multiple samples. After all analysis, I am interested in generating the feature-plot of desired set of genes. I used following two approaches :

brain1 <- LoadData("stxBrain", type = "anterior1") brain2 <- LoadData("stxBrain", type = "posterior1")

Approach 1 (merge) and featureplot brain <- merge(brain1, brain2)

brain <- SCTransform(brain, assay = "Spatial", verbose = TRUE, method = "poisson")

SpatialFeaturePlot(brain, features = c("Hpca", "Ttr"))

It shows comparable gene expression scale across two slides of spatial transcriptome samples

Approach 2 Integration and featureplot

st.list = list(anterior1 = brain1, posterior1 = brain2)

run SCT on both datasets

st.list = lapply(st.list, SCTransform, assay = "Spatial", method = "poisson")

st.features = SelectIntegrationFeatures(st.list, nfeatures = 3000, verbose = FALSE) st.list <- PrepSCTIntegration(object.list = st.list, anchor.features = st.features, verbose = FALSE)

int.anchors <- FindIntegrationAnchors(object.list = st.list, normalization.method = "SCT", verbose = FALSE, anchor.features = st.features) brain.integrated <- IntegrateData(anchorset = int.anchors, normalization.method = "SCT", verbose = FALSE)

SpatialFeaturePlot(brain.integrated, features = c("Hpca", "Ttr"))

I am facing same issue with my custom data. I have four samples grouped into two conditions (control and treatment). Using integrated approach, I find difference in the normalization gene expression score (because each sample was normalized separately using SCT) and scales are not comparable across the samples as shown in above example dataset. Ttr anterior1 samples min and max value is different from posterior1 samples. Its same case for the Hpca gene.

Just to be sure, I would like to know right approach to plot feature plot of desired set of genes.

I would appreciate all the suggestion

[jv]

not sure that this is the issue, but are you sure that SpatialFeaturePlot is pulling from the SCT assay in each case? Another way to ask this: what is the default assay of the brain.integrated and brain objects?