ValueError: No macrostates have been selected.

[meaksu]

Report

When running predict_terminal_states(), I am getting the following error. Does this mean it wasn't able to determine any terminal state? predict_initial_states() on the other hand is able to run and predict a state.

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
Cell In[16], line 1
----> 1 g.predict_terminal_states()
      2 g.plot_macrostates(which="terminal", legend_loc="right", s=100)

File ~/miniconda3/envs/cellrank2/lib/python3.11/site-packages/cellrank/estimators/terminal_states/_gpcca.py:318, in GPCCA.predict_terminal_states(self, method, n_cells, alpha, stability_threshold, n_states, allow_overlap)
    316     stability = pd.Series(np.diag(coarse_T), index=coarse_T.columns)
    317     names = stability[stability.values >= stability_threshold].index
--> 318     return self.set_terminal_states(
    319         names,
    320         n_cells=n_cells,
    321         allow_overlap=allow_overlap,
    322         params=self._create_params()
    323     )
    324 else:
    325     raise NotImplementedError(f"Method `{method}` is not yet implemented.")

File ~/miniconda3/envs/cellrank2/lib/python3.11/site-packages/cellrank/estimators/terminal_states/_gpcca.py:433, in GPCCA.set_terminal_states(self, states, n_cells, allow_overlap, cluster_key, **kwargs)
    431     states = [states]
    432 if not len(states):  # unset the states
--> 433     raise ValueError("No macrostates have been selected.")
    435 is_singleton = memberships.shape[1] == 1
    436 memberships = memberships[list(states)].copy()

ValueError: No macrostates have been selected.

Version information

Package	Version
cellrank	2.0.7
numpy	2.4.3
pandas	2.3.3
scanpy	1.11.5
matplotlib	3.10.8
anndata	0.12.10
cellmapper	0.2.5
Component	Info
Python	3.11.15 (main, Mar 11 2026, 17:20:07) [GCC 14.3.0]
OS	Linux-5.14.0-611.38.1.el9_7.x86_64-x86_64-with-glibc2.34
Updated	2026-03-28 01:46
Dependencies

[Marius1311]

Hi @meaksu, could you please update CellRank to the most recent version and then try again? Also, the high-level interface is deprecated and will be removed in v3, see our tutorials for the recommended workflow, which gives you more control.

[meaksu]

Hi, thanks for your reply. I am still getting the same error after upgrading to version 2.2.0. I listed the dependencies below. Also, I am using the tutorial here: https://cellrank.readthedocs.io/en/latest/notebooks/tutorials/estimators/600_initial_terminal.html
I'm not sure which is the high-level interface, would the above link be the correct tutorial to follow?

| Package    | Version |
| ---------- | ------- |
| cellrank   | 2.2.0   |
| numpy      | 2.4.4   |
| pandas     | 2.3.3   |
| scanpy     | 1.12    |
| matplotlib | 3.10.8  |
| anndata    | 0.12.10 |
| cellmapper | 0.2.5   |

| Dependency        | Version      |
| ----------------- | ------------ |
| python-dateutil   | 2.9.0.post0  |
| wrapt             | 2.1.2        |
| platformdirs      | 4.9.4        |
| pure_eval         | 0.2.3        |
| pyparsing         | 3.3.2        |
| pygam             | 0.12.0       |
| threadpoolctl     | 3.6.0        |
| legacy-api-wrap   | 1.5          |
| session-info2     | 0.4          |
| slepc4py          | 3.24.3       |
| setuptools        | 80.10.2      |
| jupyter_client    | 8.8.0        |
| python-utils      | 3.9.1        |
| ipykernel         | 7.2.0        |
| natsort           | 8.4.0        |
| pygpcca           | 1.0.4        |
| comm              | 0.2.3        |
| jedi              | 0.19.2       |
| llvmlite          | 0.46.0       |
| tornado           | 6.5.5        |
| ipython           | 9.12.0       |
| matplotlib-inline | 0.2.1        |
| Pygments          | 2.20.0       |
| pyzmq             | 27.1.0       |
| pytz              | 2026.1.post1 |
| cycler            | 0.12.1       |
| scipy             | 1.16.3       |
| h5py              | 3.16.0       |
| patsy             | 1.0.2        |
| typing_extensions | 4.15.0       |
| fast-array-utils  | 1.4          |
| stack-data        | 0.6.3        |
| pynndescent       | 0.6.0        |
| numcodecs         | 0.16.5       |
| jupyter_core      | 5.9.1        |
| executing         | 2.2.1        |
| psutil            | 7.2.2        |
| networkx          | 3.6.1        |
| six               | 1.17.0       |
| kiwisolver        | 1.5.0        |
| decorator         | 5.2.1        |
| petsc4py          | 3.24.6       |
| PyYAML            | 6.0.3        |
| packaging         | 25.0         |
| seaborn           | 0.13.2       |
| umap-learn        | 0.5.11       |
| scikit-learn      | 1.8.0        |
| parso             | 0.8.6        |
| prompt_toolkit    | 3.0.52       |
| numba             | 0.64.0       |
| pillow            | 12.1.1       |
| statsmodels       | 0.14.6       |
| donfig            | 0.8.1.post1  |
| zarr              | 3.1.6        |
| docrep            | 0.3.2        |
| traitlets         | 5.14.3       |
| scvelo            | 0.3.4        |
| tqdm              | 4.67.3       |
| asttokens         | 3.0.1        |
| debugpy           | 1.8.20       |
| google-crc32c     | 1.8.0        |
| progressbar2      | 4.5.0        |
| wcwidth           | 0.6.0        |
| rich              | 14.3.3       |
| joblib            | 1.5.3        |

| Component | Info                                                                              |
| --------- | --------------------------------------------------------------------------------- |
| Python    | 3.12.13 | packaged by Anaconda, Inc. | (main, Mar 19 2026, 20:20:58) [GCC 14.3.0] |
| OS        | Linux-5.14.0-611.38.1.el9_7.x86_64-x86_64-with-glibc2.34                          |
| Updated   | 2026-03-31 20:40   

[KejingDong]

I also encountered this problem. Have you solved it?