Hello, I am looking to identify specific cells with high expression of marker genes using single-cell transcriptomics. Should I find cell populations expressing these marker genes at high levels through standard clustering methods, or should I instead directly select cells with higher levels of expression?Which is better?
Personally I recommend you to preform the standard clustering pipeline and visualized under UMAP embeddings first. Then you can featureplot genes of interest to check if they were specifically expressed in certain clusters. Try to identify the major cell types in your data based on co-expression of at least two reported marker genes (e.g. neuron, epi, blood cell). If genes of interest were uniquely detected in one population, then you can focus on them. Otherwise you need to distinguish if the 'marker genes' you mentioned are the real marker genes or not.
Sometimes the experimentally used marker genes did not equal to the specificity in scRNA-seq data.
- Badges
- Science topic
- Similar topics
- Microbiology
- Assays