stereo.core.ms_data.MSData.to_integrate

MSData.to_integrate(scope, res_key, _from, type='obs', item=None, fill=nan, cluster=True)[source]

Integrate an obs column or a var column from some single-samples spcified by _from to the merged sample.

Parameters:

  • scope (slice) – Which integrate mss group to save result.

  • res_key (str) – New column name in merged sample obs or var.

  • _from (slice) – Where to get the single-sample target infomation.

  • type (Literal['obs', 'var']) – obs or var level, defaults to ‘obs’.

  • item (Union[list, ndarray, str, None]) – The column names in single-sample obs or var, defaults to the value of res_key.

  • fill – Default value when the merged sample has no conrresponding item, defaults to np.NaN.

  • cluster (bool) – Whether it is a clustering result, defaults to True.

Note

The length of scope must be equal to _from.

The type just only supports ‘obs’ currently.

Examples

Constructing MSData from 5 single-samples.

>>> import stereo as st
>>> data1 = st.io.read_h5ad('../data/10.h5ad')
>>> data2 = st.io.read_h5ad('../data/11.h5ad')
>>> data3 = st.io.read_h5ad('../data/12.h5ad')
>>> data4 = st.io.read_h5ad('../data/13.h5ad')
>>> data5 = st.io.read_h5ad('../data/14.h5ad')
>>> ms_data = data1 + data2 + data3 + data4 + data5
>>> ms_data
ms_data: {'0': (493, 30254), '1': (285, 30254), '2': (753, 30254), '3': (731, 30254), '4': (412, 30254)}
num_slice: 5
names: ['0', '1', '2', '3', '4']
obs: []
var: []
relationship: other
var_type: intersect to 0
mss: []

Integrating all samples to a merged one.

>>> ms_data.integrate()

Integrating an obs column named as ‘celltype’ from first three samples to the merged sample, to name as ‘celltype’

>>> from stereo.core.ms_pipeline import slice_generator
>>> ms_data.to_integrate(res_key='celltype', scope=slice_generator[0:3], _from=slice_generator[0:3], type='obs', item=['celltype'] * 3)

Integrating an obs column named as ‘celltype’ from all samples to the merged sample, to name as ‘celltype’

>>> from stereo.core.ms_pipeline import slice_generator
>>> ms_data.to_integrate(res_key='celltype', scope=slice_generator[:], _from=slice_generator[:], type='obs', item=['celltype'] * ms_data.num_slice)