Input and output

[ShellyCoder]

Dear author,

Thanks for developing the software! I have two questions but did not find a clear answer from your paper:

1. In the first figure below, I found that the file auto/Zeisel/Zeisel.h5 is a matrix of 3005 cells and 19972 genes, is it an unnormalized raw count matrix? Has it been normalized by library size, or log2-transformed?

2. I tried to perform the example code bash zeisel_exp.sh, and it succeed. But I am not clear about the result file, I guess the result.h5 file is the matrix after imputation. But which is the hidden layer file used for clustering as mentioned in the article?

3. If the input is the raw count, then the result file is also the missing count value. I don't know if I understand it correctly.

4. Will the tool filter the cell and gene, i.e. will the dimension of the input matrix and the dimension of the output matrix be inconsistent, and how to set the parameter to avoid this situation?

Thanks!