Seurat -Find conserved cluster markers and DE genes in two samples
For a given cluster this tool gives you
Conditions are determined in the Setup tool with the Sample or Group name parameter.
- the cell type markers that are conserved across conditions (conserved_markers.tsv)
- differentially expressed genes between the conditions (de-list.tsv)
- Name of the cluster 
- Return only positive marker genes [TRUE]
- Fold change threshold for conserved markers in ln scale [0.25]
- Fold change threshold for differentially expressed genes in ln scale [0.25]
- Adjusted p-value cutoff for conserved markers [0.05]
- Adjusted p-value cutoff for differentially expressed genes [0.05]
As inputs, give the combined Seurat object.
Select the cluster you want to inspect by setting its name in the parameter field. As an example, the parameter
is set to cluster "3".
You can filter both the conserved marker genes and the differentially expressed genes using the parameters. By default, only genes with
adjusted p-value < 0.05 are listed in the result table and only positive markers are kept. Note that Seurat adjusts p-values using the rather conserved Bonferroni method for multiple testing correction.
You can choose to include in the analysis only genes whose average fold change is higher than a selected value. This prefiltering speeds up the analysis
and narrows down the result list. Note that Seurat calculates fold change in ln scale (log base e), so if you are interested in two-fold expression changes in linear scale,
you need to enter 0.693 in the parameter.
The result table de-list.tsv contains the genes which are differentially expressed between the two conditions in the selected cluster. The columns include
- p_val = p-values for the differentially expressed genes (larger the p-value -> higher the
likelihood that the gene is in the list just be chance)
- p_val_adj = adjusted p-value. This value is corrected for multiple testing:
when we test thousands of genes, we can get some statistically significantly
differentially expressed genes just by chance. There are different methods to correct for this,
Seurat uses the Bonferroni method. When filtering the table and reporting your results,
use this value.
- avg_logFC = average ln fold change (how much higher or lower the expression of this gene is
in the chosen cluster between the conditions
- pct.1 = what percentage of the cells in the condition 1 show some expression for this gene
- pct.2 = what percentage of the cells in the condition 2 show some expression for this gene
conserved_markers.tsv contains a ranked list of conserved marker genes for the chosen cluster
(= genes that show specific expression in this particular cell type/cluster, regardless of the treatment).
In addition to the columns listed above, this table contains also the following columns:
- max_pval = maximum p-value of the two samples (note: not looking at the adjusted p-values)
- min_pval = minimum p-value of the two samples (note: not looking at the adjusted p-values)
For more details, please check the Seurat tutorials
for multiple sample analysis.
- conserved_markers.tsv: Conserved cell type markers across the conditions
- de-list.tsv: Differentially expressed genes between the conditions