Get aData object to process MELD graph parameters

[Baboon61]

Super interesting work and congratulations for the Nature Biotech !

I am trying to run your in depth tutorial on my own dataset and as recommended in your notebook, one should reprocess the graph parameters for the MELD algorithm because the best beta and knn parameters may vary according to dataset size and structure.

Running the meld.Benchmarker() I came across this function benchmarker.fit_graph(adata.X, knn=knn) where aData seems to be an agraph object to pass as input to the fit_graph function.

I did not find any information about how to get this object in your notebooks or in the manuscript. Can you help me get this object ? Thanks !

[dburkhardt]

Oh that's an AnnData object: https://anndata.readthedocs.io, so adata.X is just the normalized counts matrix (n_cells by n_genes)

[Baboon61]

benchmarker.fit_graph(data_libnorm, knn=knn) should do it then ,thanks !