Reproducibility of human tonsil data analysis

[Irisapo]

Hi,

Can you guys help me in reproducing the clustering and gene score results on the human tonsil data?
There might be something I did wrong, as I am only able to get 2 clusters, and half of the genes predicted non-zero scores with the parameter settings you guys have in the script/paper.
Here is what I did:

I filtered out 12 spots as you guys did, 
I then used the 2488 remaining spots to do dimensionality reduction (addIterativeLSI) and clustering (addClusters). 
The gene scores were then calculated using addImputeWeights. 
All the parameters I used are the same from the manuscript/your script. 

As I am noticing that the script you guys provided here are specific for mouse datasets. I wondering whether I have missed anything.

Thank you for your help in advance!
Shiyi