Multi-subject Single Cell deconvolution(MuSiC) with python code.
More information about MuSiC here xuranw's github
Fix some bugs in original MuSiC when calculate the weight by co-variance of multi-subject.
Data formats(TXT) examples can be found in data folder.
Use the function like :
sc_Meta = Get_SC_Meta_Info('sc_meta.txt')
sc_Count = Get_SC_Count_Info('sc_count.txt')
bulk_Meta = Get_bulk_Meta_Info('bulk_meta.txt')
bulk_Count = Get_bulk_Count_Info('bulk_Count.txt')
select_ct = ['alpha', 'beta', 'delta', 'gamma', 'acinar', 'ductal']
#weight gene by variance of cross-subject
Results = music_prop(bulk_Meta, bulk_Count, sc_Meta, sc_Count,ct_cov=False, select_ct=select_ct)
#weight gene by co-variance of cross-subject
Results = music_prop(bulk_Meta, bulk_Count, sc_Meta, sc_Count,ct_cov=True, select_ct=select_ct)
results :
Results.initial_p #NNLS results
Results.weight_p #Weighted-NNLS results
Results.R2 #R-squared for weighted-NNLS
...
...