Entering edit mode

10 months ago

synat.keam
▴
100

Dear all fellows,

What is the standard way for reproducible Seurat clustering in single cell? I have been trying to use `set.seed`

function like

```
library(tidyverse)
library(Seurat)
library(patchwork)
set.seed(198752)
```

However, once it comes to running the following script, I get new shape of cluster all the time I rerun it. How to make it reproducible because I am planning to use it for publication.

```
All_arthritis_integration_UMAP<- RunHarmony(All_arthritis_normal, group.by.vars = "orig.ident")
All_arthritis_integration_UMAP <- FindNeighbors(All_arthritis_integration_UMAP, reduction= "harmony", dims = 1:40)
All_arthritis_integration_UMAP <- FindClusters(All_arthritis_integration_UMAP, resolution = 0.2)
All_arthritis_integration_UMAP <- RunUMAP(All_arthritis_integration_UMAP, reduction = "harmony", dims = 1:40,
n.neighbors = 30, min.dist = 0.4, spread =3)
`
```

Also, I would like to ask whether I could adjust some parameters in `RunUmap`

to make cluster look better by trying some parameters like `n.neighbors= XX`

, `min.dist= XX`

, `spread= XXX`

because if I run the default function, the clusters in my data looks messy.

Kind Regards,

Synat

Is that your birthday? Be careful on the web, my friend!

not my birthday. thanks

Nice. Glad to know. I'll delete my comment in a bit.