change readme

R/FastIntegration.R

@@ -225,12 +225,12 @@ FastFindWeights = function(
   )
 
   distances = Seurat::Distances(object = knn_2_2)
-  a = as.numeric(distances[,ncol(x = distances)]/distances[,2])
-  b = distances/distances[,2]
-  b = which(b > qnorm(0.99, mean(a), sd(a)))
-
-  distances = 1 - (distances / distances[, ncol(x = distances)])
-  distances[b] = 0
+  # a = as.numeric(distances[,ncol(x = distances)]/distances[,2])
+  # b = distances/distances[,2]
+  # b = which(b > qnorm(0.99, mean(a), sd(a)))
+  #
+  # distances = 1 - (distances / distances[, ncol(x = distances)])
+  # distances[b] = 0
   cell.index = Seurat::Indices(object = knn_2_2)
   FindWeightsC = getFromNamespace("FindWeightsC", "Seurat")
   weights = FindWeightsC(

README.md

@@ -56,6 +56,8 @@ for (i in 1:length(rna.list)) {
 
 
 ### Onestop function
+
+For large sample size (> 200 samples), we recommend to use step by step integration.
 ```R
 library(FastIntegration)
 # rna.list is the list of seurat object
@@ -77,6 +79,13 @@ library(FastIntegration)
 BuildIntegrationFile(rna.list = rna.list, tmp.dir = "./", nCores = 50)
 FastFindAnchors(tmp.dir = "./", nCores = 50)
 
+# close current R session and open a new one to clean the memory (This is important for large data integration)
+# In the new session, please just set work directory and do not load any data. Then run the following codes:
+
+library(Seurat)
+library(pbmcapply)
+library(FastIntegration)
+
 genes = readRDS("FastIntegrationTmp/raw/1.rds")
 genes = rownames(genes)
 idx = split(1:length(genes), cut(1:length(genes), 20, labels = FALSE))