Reproducinility of FOXN1 activity in Figure 6b

[kmakino14]

I input 10x PBMC 10K multisource data into LINGER to calculate the TF activity of the bulk data (reference 72) based on the LINGER-inferred cell population GRNs. Unfortunately, the activity of FOXN1 did not reach the values shown in the paper and was not lower in AML patients (regulon activity was about 10e-4 order as in the violin plot of the PBMC tutorial). Could I perhaps request the code you used? Thank you in advance for your help.

[kmakino14]

The problem seems partly due to the fact that the "pseudo_bulk" function return values are not always the same. np.random.seed(42) does not work and should probably be random.seed(42). In any case, I do not get the FOXN1 activity values shown in Figure 6b of the paper.