Entering edit mode
3.3 years ago
linhzye
▴
10
Dear all,
please may I ask for a suggestion :
I have a scRNA-seq dataset, which has 2 groups (Control and Model).
I would like to process all the samples into cluster by tSNE or UMAP , then classification or differentiation the cell types.
Which method should I choose?
- Method one: The model group and the control group were analyzed respectively.
- Method two: All the samples were analyzed together.
In my opinion, as there may be differences in gene expression between the two groups, it is necessary to analyze and identify cells separately. But many articles only provide one figure about tsne or cell classification, and I suspect that the method two is adopted.
Thank you for your reply.
According to link you provide:
Chapter 9 Dimensionality reduction: For example, clustering aims to identify cells with similar transcriptomic profiles by computing Euclidean distances across genes.
Chapter 10 Clustering: Clustering is an unsupervised learning procedure that is used in scRNA-seq data analysis to empirically define groups of cells with similar expression profiles.
Will the difference of expression/transcriptomic profile between the treatment group and the control group lead to different result when clustering separately or together?
Or do we infer that the differences in gene expression between the two groups are too small to affect the cluster analysis?
See chapters 7 and 13 in that book.
Thank you!
I think I get it.
If the datasets generated in the same batche, it was no need to eliminate batch effect. Otherwise, did it.
Thank you all of you.