Change in eRegulons filtering for Heatmap_dotplot

[Goultard59]

Hello,

I noticed that the refiltering based on the TF names allows the persistence of eRegulons that are not in the eRegulons list, like extended eRegulons or +_- regulons.

For example if my subset_eRegulons contains :
'CREB1_++(1176g)'

Then in my heatmap I might also find:
'CREB1_++(1176g)',
'CREB1_+-(187g)',
'CREB1_extended_++(1176g)',
'CREB1_extended_+-(189g)'

[SeppeDeWinter]

Hi @Goultard59

You're right! This is a problem.

The reason it is done like this is to be able to intersect with actual TF names (for which we have expression values (i.e. not the AUC values)).

I'm currently making changes to the code so that we don't have to rely at all anymore on the actual name of the eRegulon and do string operations (which are a bit dirty).

So I won't merge this code yet. But thank you for the contribution.

Best,

Seppe

[Goultard59]

Hi,

I added a few fixes when no repressor was identified :
A fix on x text label orientation
A check for the availability of repressor that lead to an error in plotnine if split_repressor_activator is set to True

Best regards,
Adrien Dufour.