heterogeneity in cluster when using Seurat Doheatmap
0
0
Entering edit mode
5 months ago
Jie • 0

Hi, I am trying to identify clusters from my data (scRNA-seq for ~10k heart immune cells ) using Seurat4 in R server.

When I attempted to find clusters, I firstly tested resolution=0.4 in FindClusters() and got . ( min.pct = 0.2, logfc.threshold = 0.25 for FindAllMarkers(), I keep these parameters unchanged)

Then I used Doheatmap() to show top20 markers, and I found obvious subclusters in several top clusters. So I increased my resolution to 0.8, 1.0, 1.5, but still it seem weird.

The figure.1~3 refer to res=0.4, 0.8, 1.5 , we can see subclusters.

SO I would like to have your advice on the causes of this situation and what to do next.

My exploration:

  1. I have used clustree to find the proper resolution, we consider 1.5 too high and unnecessary, so increasing resolution might not be a good way to solve the problem.
  2. I subset the cluster0 and re-run the pipeline (from FindVariableFeatures to Doheatmap), the heatmap looks better, however the marker genes are different from last step. And is this a good way?

Any help or discussion will be appreciated.

res=0.4 res=0.8 res=1.5

scRNAseq cluster r Seurat seurat • 576 views
ADD COMMENT
0
Entering edit mode

Having multiple bars doesn't mean you have sub-clusters. If you change the order of the genes you will obtain different bars. What is important in your heatmap, is that each clusters have a unique pattern which indicate different expression profiles. Furthermore, if your cluster contains similar cell types (e.g. fibroblast and myo-fibroblast or TCD8 and NK), you expect to see both shared (faith in colour) and not shared (intense colour) bars.

ADD REPLY
0
Entering edit mode

Thanks for your reply! :)

ADD REPLY

Login before adding your answer.

Traffic: 1530 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6