spatial shrunken centroids for high resolution data

[Miao YU (yufree)]

Hi Community, 

Thanks for this useful software!

I am dealing with a high resolution MS imaging with high resolution MS. After the pretreatment, I got around 2000 reference peaks or features across 1 million pixels. For PCA, it's still fine. However, it seems the spatial shrunken centroids analysis is very slow for those data. 

I tried to subset the pixels and found it will take more than one hour when process ~90000 pixels. However, when the pixels went to 1 million pixels, spatial shrunken centroids became very slow and one day is not enough. For the same number of pixels, when I increase the r, the speed is also slow. I wonder if such SSC analysis could be optimized for larger scale data size? Will the memory usage be the major limitation for this algorithm? I am running those analysis on hpc and I wonder if anyone can suggest a computational resource to process those data.

Thanks,

Miao