aristoteleo / scribe-python-notebooks

Tutorials of Scribe Python package usage cases

Scribe-Python-notebooks

Tutorials of Scribe Python package usage cases

scribe-python-notebooks's Issues

Include benchmark with the neural simulation network in Scribe paper

Currently the notebook includes an application on the simulation system from scanpy. The simulation system used in the Scribe paper should also included, ideally full types of data should be included (real time, RNA velocity, pseudotime and live-image).

problem running Scribe(adata, t0_key='spliced', t1_key='unspliced')

Hi Xiaojie,
A really cool tool for building the network, thanks for putting it together.
When running NASC_seq_Jurkat_T_cell_demo.ipynb with the demo data I was not able to calculate causality score running Scribe(adata, t0_key='spliced', t1_key='unspliced').

Full error text:

IndexError Traceback (most recent call last)
in
----> 1 Scribe(adata, t0_key='spliced', t1_key='unspliced')

/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/Scribe/Scribe.py in causal_net_dynamics_coupling(adata, TFs, Targets, guide_keys, t0_key, t1_key, normalize, drop_zero_cells, copy)
127 z_orig = [[i] for i in z_orig]
128
--> 129 causal_net.loc[g_a, g_b] = cmi(x_orig, y_orig, z_orig)
130
131 adata.uns['causal_net'] = {"RDI": causal_net.fillna(0)}

/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/Scribe/information_estimators.py in cmi(x_orig, y_orig, z_orig, normalization, k)
144 N = len(x)
145
--> 146 dx = len(x[0])
147 dy = len(y[0])
148 dz = len(z[0])

IndexError: list index out of range

Fixed with your advise by setting
drop_zero_cells=False in the function causal_net_dynamics_coupling

Scribe(adata, t0_key='spliced', t1_key='unspliced', drop_zero_cells=False)

Hope this report maybe useful for others
Thanks!

Is there a demo integrated with dynamo?

Thank you!