scSELpy (Single-cell selection in Python) is a manual cell selection tool to support Scanpy-based pipelines. It calls Scanpy generated plots and allows to user to draw polygons on top of them, in order to select cells. The cells located within these drawn polygons are assigned an identity, which will be stored in the Anndata object as an observation. Cells located within multiple Polygons will have multiple identities for in the same observation, separated by a comma.
Python3-7+
pip install scselpy
Installation and import are fully in lower-case.
When running scSELpy on Jupyter Notebook, the backend will temprary change to an interactive backend. The default interactive backend is Qt5Agg. If you are getting this error: ImportError: Failed to import any of the following Qt binding modules: PyQt6, PySide6, PyQt5, PySide2 please install PyQt5 with pip install PyQt5.
If the backend does not work on your computer, try using to a different one by running scselpy.pl.umap(adata,interactive_backend="TkAgg"). All matplotlib supported backends can be found here.
While running scSELpy in a Python shell such as ipython the default backend is usually interactive and therefore a switch will not be conducted, however, if you are experiencing troubles, it is possible to temporary switch the interactive backend to e.g. Qt5Agg or TkAgg by using the command above.
In order to get started with scSELpy, please refer to the documentation. The Tutorial can also be found as a notebook on this github/folder.