Reply
Highlighted

Cluster-specific genes/marker-gene-based classification for immune cells?

Posted By: cykao, on Dec 6, 2017 at 12:50 AM

We have used Loupe Cell browser to view the clusters generated from our scRNA-seq for CD45+ mouse cells. How do further we characterize these distinct populations by cluster-specific genes/marker-gene-based classification? Is there any suggested bioinformatic package? I will try Seurat first.  Will CIBERSORT work for scRNA-seq data? Thanks for advice.

4 Replies

Re: Cluster-specific genes/marker-gene-based classification for immune cells?

Posted By: Leah, on Dec 8, 2017 at 2:21 PM

In Loupe Cell Browser, you can visualize the putative cell sub-populations identified based on either graph-based clustering or k-means clustering, as you have already done.

 

You can also export the table of cluster-specific genes based on the “Significant Genes Analysis”.

 

By default, Cell Ranger also produces a file indicating which genes are differentially expressed in each cluster relative to all other clusters. 

 

analysis/diffexp/*_clusters/differential_expression.csv

https://support.10xgenomics.com/single-cell-gene-expression/software/pipelines/latest/output/analysi...

 

Significant genes with log2 fold-change greater than 0 for each cluster represent candidate cell-type enriched/specific markers.

 

One relatively simple approach would be to apply gene set or pathway enrichment analysis tools to the lists of cluster-specific/cluster-enriched genes to further characterize each putative sub-population identified as a cluster.

 

Hope this helps.

Re: Cluster-specific genes/marker-gene-based classification for immune cells?

Posted By: freshcalendar1, on Dec 15, 2017 at 2:55 PM

thank you so much for your advice Smiley Happy

Re: Cluster-specific genes/marker-gene-based classification for immune cells?

Posted By: cykao, on Dec 18, 2017 at 6:05 AM

Thaks for the reply. I tried but pathway enrichmet aalysis on the lists exported from cluster-specific/cluster-enriched up-regulated genes did not show much hint in my case. Basically I have CD45+ cells isolated from WT and KO mice. We used 10x genomics to get the scRNA-seq done on ~ 4000 cells from each genotype. The key question to address is to predict which kind of immune cells are more abundant or less abundant in the KO mice. I did not see a very separated clustering (as the PBMC sample dataset from 10x Gennomics) from our Loupe Cell Brower but I did see a clear clustering difference when the WT and KO data were put together in Loupe Cell Browser. With the split view, I could see some clustering are dominatley eenriched by the WT and some are enriched by the KO. So I'd like to know what kind of immune cells these clusters could be. For example, for ILC2, in Loupe Cell Brower, is there a way that I could find and label all the cells with "Il2ra, Thy1, Kit, Il7r, Ly6a, Il1rl1, Il17rb, Rora, Gata3" co-expression? Thanks.

Re: Cluster-specific genes/marker-gene-based classification for immune cells?

Posted By: shauna-10x, on Dec 19, 2017 at 10:58 AM

Hi,

 

Thanks for your follow up question.  I consulted with the support team and this is an issue they can better handle through their ticketing system.  I've opened a ticket for you and support will be following up to help answer your question.

 

Thanks,

Shauna