Recovering Dynamics and Differential Kinetics

[ACastanza]

Hello,

I have a couple questions about the Differential Kinetics tutorial: https://colab.research.google.com/github/theislab/scvelo_notebooks/blob/master/DifferentialKinetics.ipynb#scrollTo=GuzC3L3AUKtT

Lets say that instead of the "standard" velocity estimation, you used the dynamical model in the first "Basic Velocity Estimation" section, do you then need to change anything in specific in running the "Differential Kinetic Test" section (for example, do you need to run the dynamics recovery again after the differential kinetic test?

Additionally, I noticed that in the "Differential Kinetic Test", you run scv.tl.recover_dynamics after scv.tl.differential_kinetic_test (is this actually needed? it appears based on the output text from the function that this is already running "recovering dynamics") F

Finally, then when running scv.tl.velocity it's run with just specifying diff_kinetics=True but not mode='dynamical' Should mode='dynamical be passed at this step or is just assumed when diff_kinetics=True is passed?

Thanks!

[WeilerP]

@VolkerBergen, any thoughts?

[ACastanza]

I think if you run scv.tl.velocity(adata, mode='dynamical', diff_kinetics=True) after running scvelo with the stochastic model instead of the dynamical model in the very first step you get the following error even if you recompute the dynamics:

Traceback (most recent call last):
  File "/opt/conda/lib/python3.8/site-packages/pandas/core/indexes/base.py", line 3080, in get_loc
    return self._engine.get_loc(casted_key)
  File "pandas/_libs/index.pyx", line 70, in pandas._libs.index.IndexEngine.get_loc
  File "pandas/_libs/index.pyx", line 101, in pandas._libs.index.IndexEngine.get_loc
  File "pandas/_libs/hashtable_class_helper.pxi", line 4554, in pandas._libs.hashtable.PyObjectHashTable.get_item
  File "pandas/_libs/hashtable_class_helper.pxi", line 4562, in pandas._libs.hashtable.PyObjectHashTable.get_item
KeyError: 'clusters'

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "/opt/gpbeta_2/gp_home/taskLib/scVelo.ComputeVelocity.1.1978/compute_scvelo.py", line 254, in <module>
    main()
  File "/opt/gpbeta_2/gp_home/taskLib/scVelo.ComputeVelocity.1.1978/compute_scvelo.py", line 195, in main
    scv.tl.velocity(adata, mode='dynamical', diff_kinetics=True)
  File "/opt/conda/lib/python3.8/site-packages/scvelo/tools/velocity.py", line 445, in velocity
    clusters = adata.obs[groupby]
  File "/opt/conda/lib/python3.8/site-packages/pandas/core/frame.py", line 3024, in __getitem__
    indexer = self.columns.get_loc(key)
  File "/opt/conda/lib/python3.8/site-packages/pandas/core/indexes/base.py", line 3082, in get_loc
    raise KeyError(key) from err
KeyError: 'clusters'

I'm pretty sure this is getting caused by something not being detected properly here: https://github.com/theislab/scvelo/blob/master/scvelo/tools/velocity.py#L448

But I'm still not entirely sure it's not something else I did wrong instead.
My data has leiden clusters not louvian so they're not stored under the "clusters" key, but it shouldn't be looking for them at that step anyway since it should be doing the groupby = adata.uns["recover_dynamics"]["fit_diff_kinetics"] step not the else: groupby = "clusters" I think

[WeilerP]

@ACastanza, could you please open an issue for this with a reproducible code example, the traceback and package versions used. I'll look into it once I find the time.