search

BIMSBbioinfo / bavaria

Batch-adversarial variational auto-encoder (BAVARIA) for simultaneous dimensionality reduction and integration of single-cell ATAC-seq datasets
Other
12 stars 6 forks source link

Error in tutorial/00_preparation.ipynb #1

Closed prete closed 3 years ago

prete commented 3 years ago

Hi, tried the preparation notebook and this cell fails:

adatas = []
for batchname in batches:
    barcodes = pd.read_csv(batches[batchname]['cells'])
    barcodes = barcodes[barcodes.is__cell_barcode==1]
    barcodes.loc[:,"batch"] = batchname
    barcodes.set_index('barcode', inplace=True)

    frags = get_fragment_bedtool(batches[batchname]['frag'], barcodes)

    peakcounts = peak.intersect(frags,
                             wa=True,
                             wb=True).to_dataframe()
    sparse_data = np.asarray([np.ones(peakcounts.shape[0]),
                             peakcounts.name, peakcounts.itemRgb]).T
    sparse_data = np.unique(sparse_data, axis=0)
    mat = coo_matrix((sparse_data[:,0], (sparse_data[:,1], sparse_data[:,2])),
                     shape=(len(peak), len(barcodes)))
    adata = anndata.AnnData(mat.T.tocsr(), obs=barcodes, var=peakdf)
    adatas.append(adata)

the error is in this line:

    adata = anndata.AnnData(mat.T.tocsr(), obs=barcodes, var=peakdf)

because peakdf it's not defined in that notebook, was it meant to be dfpeak ?

wkopp commented 3 years ago

Hi @prete . Yes, that is correct. peakdf should be dfpeak. I pushed a corrected version of the notebook. Thank you very much for reporting!