In the article that referenced the data that I am currently processing, it used the computeSumFactors method to normalize them.
And with my studies, I realized that this method is better for normalizing sparse count data
But I don’t know if the tool introduced in the tutorial on Clustering 3K PBMCs with Scanpy for normalization (Normalize with scanpy) uses the same normalization method or not?
And if this is not the case, due to the fact that my data is a type of sparse data and normalization by the computeSumFactors method is mandatory for them.
Does Galaxy have a tool to normalize my data in this way?
hi @maryam-gh99 we probably might have discussed about it but I am still replying here so that other users might benefit.
The default normalizations by Scanpy and Scran are different. Scanpy uses a simple shifted logarithm whereas scran estimates size factors by pooling cells. Please refer to the single-cell best practices guide to learn some basic details about those two normalization methods. There is always a debate that the choice of normalization method should be based on the post-processing steps. However, this recent benchmark of scRNA-seq normalization methods showed that shifted logarithm performed quite well. We have both the Scran and Scanpy normalizations available on Galaxy.
1 Like