ValueError: 85 value(s) are negative.

[mayowaakinwale]

Hi,

Just trying to run through the tutorial using my own data but have run into this issue when predicting terminal states.

`terminal_states = [
"PNEC_2",
"Multiciliated"
]
vk = cr.kernels.VelocityKernel(adata_target).compute_transition_matrix()
estimator = cr.estimators.GPCCA(vk)

estimator.compute_macrostates(n_states=7,cluster_key="predicted_celltype")
estimator.set_terminal_states(terminal_states)
estimator.compute_fate_probabilities()
estimator.plot_fate_probabilities(same_plot=False)`

Any suggestions for how to resolve this?

[WWXkenmo]

Hey,

This issue is due to cellrank step, I have three ways you could have try and see which one will work out.

1. replace vk -> 0.8vk+0.2ck estimator = cr.estimators.GPCCA(vk)
2. Try solver = 'direct'
3. Try to install petsc4py and slepc4y ensure you can use method=brandts

Let me know if above suggestions will help!

Best,
Ken

[WWXkenmo]

Also please check the cellrank repo (issues) to see solution, for example: theislab/cellrank#1022