How can I IntegrateData and save all features?

[ShuqingL]

I have 2 SpaceRanger outs.

First, I transform them to Seurat objects:

library(Seurat);
library(ggplot2);
library(ggpubr);
library(scales);

setwd('/path/to/To_Seurat');
sample_dirs <- list(
sample1 = '/path/to/SpaceRanger/sample1/sample1/outs/',
sample2 = '/path/to/SpaceRanger/sample2/sample2/outs/');
hb_genes <- unlist(strsplit('HBA1,HBA2,HBB,HBD,HBE1,HBG1,HBG2,HBZ', split=','));
#To_Seurat
process_sample <- function(data_dir, sample_id, mito_pre, hb_genes) {
obj <- Load10X_Spatial(data.dir=data_dir);
[email protected]$orig.ident <- sample_id;
Idents(obj) <- sample_id;
names(obj@images) <- sample_id;
obj[['percent.mito']] <- PercentageFeatureSet(obj, pattern=paste0('^',mito_pre));
obj[['percent.mito']][is.na(obj[['percent.mito']])] <- 0;
exists_hb <- hb_genes[hb_genes %in% rownames(obj)];
obj[['percent.HB']] <- if (length(exists_hb) > 0) { PercentageFeatureSet(obj, features=exists_hb) } else { 0 }
obj[['percent.HB']][is.na(obj[['percent.HB']])] <- 0;
saveRDS(obj, file=paste0(sample_id, '.rds'));
return(obj);
}
for (sample_id in names(sample_dirs)) { process_sample(sample_dirs[[sample_id]], sample_id, 'MT-', hb_genes); }

Second, I do some filter and scale them:

setwd('/path/to/Filter_Scale');
sample_objs <- list(
sample1 = '/path/to/To_Seurat/sample1.rds',
sample2 = '/path/to/To_Seurat/sample2.rds');
#Filter
process_sample <- function(sample_id, Seurat_object_path) {
data <- readRDS(Seurat_object_path);
data <- subset(data, subset=nFeature_Spatial > 500 & nFeature_Spatial < 10000 & nCount_Spatial > 500 & nCount_Spatial < Inf & percent.mito > -Inf & percent.mito < 10 & percent.HB > -Inf & percent.HB < 0.5);
temp_data <- data;
Idents(temp_data) <- 'temp';
temp_data_avgExpr <- AverageExpression(temp_data, assays='Spatial', layer='counts');
data <- subset(data, features=row.names(data)[temp_data_avgExpr[[1]][,1]>0]);
rm(temp_data,temp_data_avgExpr);
return(data);
}
samples <- list();
for (sample_id in names(sample_objs)) {
samples[[sample_id]] <- process_sample(sample_id, sample_objs[[sample_id]]);
}
#Scale
options(future.globals.maxSize = 10000 * 1024^2);
process_sample <- function(sample_id, data) {
regress_item <- c('nCount_Spatial');
if(summary([email protected]$percent.mito)[['Max.']]>0) {regress_item <- c(regress_item, 'percent.mito');}
data <- SCTransform(data, assay='Spatial', vars.to.regress=regress_item, verbose=FALSE, return.only.var.genes=FALSE);
saveRDS(data, file=paste0(sample_id, '.rds'));
}
for (sample_id in names(samples)) {
process_sample(sample_id, samples[[sample_id]]);
}

Third, I want to integrate these tow object and save all features:

sample1 <- readRDS('/path/to/Filter_Scale/sample1.rds');
sample2 <- readRDS('/path/to/Filter_Scale/sample2.rds');
seu.list <- list(sample1,sample2);
names(seu.list) <- c('sample1','sample2');
for (i in 1:length(seu.list)) { seu.list[[i]] <- RenameCells(seu.list[[i]], add.cell.id = names(seu.list)[i]); }
common_features <- Reduce(intersect, lapply(seu.list, rownames));
seu.list <- lapply(seu.list, function(x) { x <- x[common_features,]; return(x); })
#Integrate
features <- SelectIntegrationFeatures(object.list=seu.list);
options(future.globals.maxSize = 10000 * 1024^2);
seu.list <- PrepSCTIntegration(object.list=seu.list, anchor.features=features, verbose=FALSE);
anchors <- FindIntegrationAnchors(object.list=seu.list, anchor.features=features, dims=1:30, normalization.method='SCT');
combined <- IntegrateData(anchorset=anchors, dims=1:30, normalization.method='SCT', features.to.integrate=common_features);

But when I run "seu.list <- lapply(seu.list, function(x) { x <- x[common_features,]; return(x); })", it tells me the error:
Error in .subscript.2ary(x, i, j, drop = TRUE) : subscript out of bounds

If I do not run "seu.list <- lapply(seu.list, function(x) { x <- x[common_features,]; return(x); })", the follow 4 lines(show befor) run correct. But when running to the "combined <- IntegrateData(anchorset=anchors, dims=1:30, normalization.method='SCT', features.to.integrate=common_features)", I will get other error messages:
[1] 1
Warning: Layer counts isn't present in the assay object; returning NULL
[1] 2
Warning: Layer counts isn't present in the assay object; returning NULL
Merging dataset 2 into 1
Extracting anchors for merged samples
Finding integration vectors
Error in (function (cond) :
error in evaluating the argument 'x' in selecting a method for function 't': subscript out of bounds
In addition: Warning messages:
1: Different cells and/or features from existing assay SCT
2: Different cells and/or features from existing assay SCT
3: Adding image data that isn't associated with any assays
4: Adding image data that isn't associated with any assays
5: Key ‘slice1_’ taken, using ‘sample2_’ instead
6: Adding image data that isn't associated with any assays
7: Adding image data that isn't associated with any assays
8: Key ‘slice1_’ taken, using ‘sample2_’ instead

Can someone tell me why?