'AnnData' object error

Hi there, Facing the following error:

scv.utils.show_proportions(sdata)
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-33-6e397acd5750> in <module>()
----> 1 scv.utils.show_proportions(sdata)

/Users/awilk/anaconda3/lib/python3.6/site-packages/scvelo/preprocessing/utils.py in show_proportions(adata)
     20     """
     21     layers_keys = [key for key in ['spliced', 'unspliced', 'ambiguous'] if key in adata.layers.keys()]
---> 22     counts_per_cell_layers = [adata.layers[key].sum(1) for key in layers_keys]
     23 
     24     counts_per_cell_sum = np.sum(counts_per_cell_layers, 0)

/Users/awilk/anaconda3/lib/python3.6/site-packages/scvelo/preprocessing/utils.py in <listcomp>(.0)
     20     """
     21     layers_keys = [key for key in ['spliced', 'unspliced', 'ambiguous'] if key in adata.layers.keys()]
---> 22     counts_per_cell_layers = [adata.layers[key].sum(1) for key in layers_keys]
     23 
     24     counts_per_cell_sum = np.sum(counts_per_cell_layers, 0)

AttributeError: 'AnnData' object has no attribute 'sum'

Same error is coming up with scv.pp.filter_genes(), scv.pp.normalize_per_cell(), etc. Some more info:

sdata
AnnData object with n_obs × n_vars = 17829 × 14152 
    obs: 'orig.ident', 'clusters'
    obsm: 'UMAP_1', 'UMAP_2'
    layers: 'spliced', 'unspliced', 'ambiguous'

sdata.layers['spliced'], sdata.layers['unspliced']
(<17829x14152 sparse matrix of type '<class 'numpy.float32'>'
 	with 25134338 stored elements in Compressed Sparse Row format>,
 AnnData object with n_obs × n_vars = 17829 × 14152 )

sdata.layers['spliced'].sum(1), sdata.layers['unspliced'].sum(1)
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-38-2dd9862f1041> in <module>()
----> 1 sdata.layers['spliced'].sum(1), sdata.layers['unspliced'].sum(1)

AttributeError: 'AnnData' object has no attribute 'sum'

The raw data are there, but it seems there must be something wrong about the way I’m storing it? (probably a simple fix I’m just missing). Thanks!

[ajwilk]

You’re right, the way I was writing the AnnData object only wrote into the spliced layer. I do all of my pre-processing in R, and so I’m trying to create the AnnData object using .mtx files written from R. This odd-looking workaround allowed me to write into all layers of the AnnData object, but still appears to throw errors downstream:

sdata = ad.read_mtx("~/combined.emat.mtx")
sdata.layers['spliced'] = sdata.X
ndata = ad.read_mtx("~/combined.nmat.mtx")
sdata.layers['unspliced'] = ndata.X
mdata = ad.read_mtx("~/combined.smat.mtx")
sdata.layers['ambiguous'] = mdata.X
sdata
AnnData object with n_obs × n_vars = 17829 × 14152 
    layers: 'spliced', 'unspliced', 'ambiguous'
pd_obs = pd.read_csv("~/combined.obs.csv")
pd_obsm = pd.read_csv("~/combined.obsm.csv")
sdata.obs = pd_obs
sdata.obsm = pd_obsm

sdata
AnnData expects string indices for some functionality, but your first two indices are: RangeIndex(start=0, stop=2, step=1). 
/Users/awilk/anaconda3/lib/python3.6/site-packages/anndata/core/alignedmapping.py:162: UserWarning: AnnData does not currently support writing or reading of 'Series' objects in obsm for either hdf5 or zarr formats.
  self._validate_value(value, key)
AnnData object with n_obs × n_vars = 17829 × 14152 
    obs: 'orig.ident', 'clusters'
    obsm: 'UMAP_1', 'UMAP_2'
    layers: 'spliced', 'unspliced', 'ambiguous'

sdata.layers['spliced'], sdata.layers['unspliced'], sdata.layers['ambiguous']
(<17829x14152 sparse matrix of type '<class 'numpy.float32'>'
 	with 25134338 stored elements in Compressed Sparse Row format>,
 <17829x14152 sparse matrix of type '<class 'numpy.float32'>'
 	with 11552274 stored elements in Compressed Sparse Row format>,
 <17829x14152 sparse matrix of type '<class 'numpy.float32'>'
 	with 3400639 stored elements in Compressed Sparse Row format>)

scv.utils.show_proportions(sdata)
Abundance of ['spliced', 'unspliced', 'ambiguous']: [0.63 0.31 0.06]

scv.pp.normalize_per_cell(sdata)
scv.pp.filter_genes_dispersion(sdata, n_top_genes=3000)
scv.pp.log1p(sdata)
scv.pp.filter_and_normalize(sdata, min_shared_counts=10, n_top_genes=3000)
scv.pp.moments(sdata, n_pcs=30, n_neighbors=30)
Normalized count data: X, unspliced.
/Users/awilk/anaconda3/lib/python3.6/site-packages/scanpy/preprocessing/_deprecated/highly_variable_genes.py:176: RuntimeWarning: invalid value encountered in greater_equal
  gene_subset = df['dispersion_norm'].values >= disp_cut_off
---------------------------------------------------------------------------
NotImplementedError                       Traceback (most recent call last)
<ipython-input-29-0feaa876caed> in <module>()
      1 scv.pp.normalize_per_cell(sdata)
----> 2 scv.pp.filter_genes_dispersion(sdata, n_top_genes=3000)
      3 scv.pp.log1p(sdata)
      4 scv.pp.filter_and_normalize(sdata, min_shared_counts=10, n_top_genes=3000)
      5 scv.pp.moments(sdata, n_pcs=30, n_neighbors=30)

/Users/awilk/anaconda3/lib/python3.6/site-packages/scvelo/preprocessing/utils.py in filter_genes_dispersion(data, flavor, min_disp, max_disp, min_mean, max_mean, n_bins, n_top_genes, log, copy)
    251             from scanpy.api.pp import filter_genes_dispersion
    252             filter_genes_dispersion(adata, flavor=flavor, min_disp=min_disp, max_disp=max_disp, min_mean=min_mean,
--> 253                                     max_mean=max_mean, n_bins=n_bins, n_top_genes=n_top_genes, log=log)
    254     return adata if copy else None
    255 

/Users/awilk/anaconda3/lib/python3.6/site-packages/scanpy/preprocessing/_deprecated/highly_variable_genes.py in filter_genes_dispersion(data, flavor, min_disp, max_disp, min_mean, max_mean, n_bins, n_top_genes, log, subset, copy)
    110         adata.var['dispersions_norm'] = result['dispersions_norm']
    111         if subset:
--> 112             adata._inplace_subset_var(result['gene_subset'])
    113         else:
    114             adata.var['highly_variable'] = result['gene_subset']

/Users/awilk/anaconda3/lib/python3.6/site-packages/anndata/core/anndata.py in _inplace_subset_var(self, index)
   1362         Same as ``adata = adata[:, index]``, but inplace.
   1363         """
-> 1364         adata_subset = self[:, index].copy()
   1365         self._init_as_actual(adata_subset, dtype=self._X.dtype)
   1366 

/Users/awilk/anaconda3/lib/python3.6/site-packages/anndata/core/anndata.py in copy(self, filename)
   1555                            # as uns was copied already before
   1556                            self._uns.copy() if isinstance(self._uns, DictView) else deepcopy(self._uns),
-> 1557                            self._obsm.copy(), self._varm.copy(),
   1558                            raw=None if self._raw is None else self._raw.copy(),
   1559                            layers=dict(self.layers),

/Users/awilk/anaconda3/lib/python3.6/site-packages/anndata/core/alignedmapping.py in copy(self)
     89     def copy(self):
     90         d = self._actual_class(self.parent, self._axis)
---> 91         for k, v in self.items():
     92             d[k] = v.copy()
     93         return d

/Users/awilk/anaconda3/lib/python3.6/_collections_abc.py in __iter__(self)
    742     def __iter__(self):
    743         for key in self._mapping:
--> 744             yield (key, self._mapping[key])
    745 
    746 ItemsView.register(dict_items)

/Users/awilk/anaconda3/lib/python3.6/site-packages/anndata/core/alignedmapping.py in __getitem__(self, key)
    119             asview(
    120                 self.parent_mapping[key],
--> 121                 ViewArgs(self.parent, self.attrname, (key,))
    122             ),
    123             self.subset_idx

/Users/awilk/anaconda3/lib/python3.6/functools.py in wrapper(*args, **kw)
    805                             '1 positional argument')
    806 
--> 807         return dispatch(args[0].__class__)(*args, **kw)
    808 
    809     funcname = getattr(func, '__name__', 'singledispatch function')

/Users/awilk/anaconda3/lib/python3.6/site-packages/anndata/core/views.py in asview(obj, view_args)
     95 def asview(obj, view_args):
     96     raise NotImplementedError(
---> 97         "No view type has been registered for {}".format(type(obj))
     98     )
     99 

NotImplementedError: No view type has been registered for <class 'pandas.core.series.Series'>

[VolkerBergen]

Looks like a whole anndata object is stored in sdata.layers['unspliced'], which does not make sense… The raw data - if needed later on - is to be stored in sdata.raw. Did you write that to unspliced layers?