Adding result to the seurat object

[idkaavan]

Thank you for the interesting tool! can you please kindly guide me how I can save the result or add it to the seurat object? Thank you in advance.

[renjun0324]

https://github.com/renjun0324/MURP/issues/1#issuecomment-1399207365

here you are

[idkaavan]

Thank you, but is the link an empty issue search or am I wrong?

[renjun0324]

Sorry, I pointed to an empty link. Here is the correct code.

library(Seurat)
library(MURP)
library(dplyr)
data(sdata)

# 1. seurat object for test
obj_1 = CreateSeuratObject(t(sdata))
obj_2 = CreateSeuratObject(t(sdata))

# 2. normalize data
obj_1 = NormalizeData(obj_1)
obj_2 = NormalizeData(obj_2)

# 3. murp
murp_1 = MURP(Data = as.matrix(obj_1@assays$RNA@data) %>% t, fast = TRUE)
murp_2 = MURP(Data = as.matrix(obj_2@assays$RNA@data) %>% t, fast = TRUE)

# 4. create new seurat object
new_obj_1 = CreateSeuratObject(murp_1$Recommended_K_cl$centers)
new_obj_2 = CreateSeuratObject(murp_2$Recommended_K_cl$centers)

# 5. combine...

Please note that MURP does not extract a subset of points from the original data as the result, but rather generates new points, which are then used to create a new Seurat object. The mapping relationship between MURP and the original data can be achieved through the murp_1$Recommended_K_cl$cluster operation.

[markpharm]

您好，我尝依据您提供的方法试运行了MURP，但是这一方法对内存的需求似乎很大，运行速度也显得不够快，不知道是否有可以修改的方法。以下是我的代码
murp_1 = MURP(Data = as.matrix(PV.Tel@assays$RNA@layers$data) %>% t, omega = 1/6,iter = 20,cores = 6,seed = 723,max_murp = round(ncol(PV.Tel)*0.375))

[renjun0324]

您好，我尝依据您提供的方法试运行了MURP，但是这一方法对内存的需求似乎很大，运行速度也显得不够快，不知道是否有可以修改的方法。以下是我的代码 murp_1 = MURP(Data = as.matrix(PV.Tel@assays$RNA@layers$data) %>% t, omega = 1/6,iter = 20,cores = 6,seed = 723,max_murp = round(ncol(PV.Tel)*0.375))

对的，比较吃内存，可以试试加这个参数，fast=TRUE，不过FALSE和TRUE的结果会不一样。另外我最近看到有一个MetaQ 的方法应该也是整类似的活，可以试试。链接：https://www.nature.com/articles/s41467-025-56424-6