Add a signifiance threshold for `plot_fold_change()`

[LuisHeinzlmeier]

Description of feature

While updating the DGE chapter in single cell best practices, I noticed that plot_fold_change() is plotting the top n log fold changes of all results, even those with high p-values. So, I filtered the results like that:

pds2.plot_fold_change(res_df[res_df["adj_p_value"] < 0.01].copy(), n_top_vars=15)

I was wondering if there shouldn't be a parameter in plot_fold_change() that filters the significant results. The parameter could be called sig_thres or alpha, for example. I would even say sig_thres should have a default (e.g., sig_thres=0.01).

I'd be happy to make this small improvement myself. I welcome any feedback.

[Zethson]

Not a fan of alpha as its a plotting function.
sig_thres is kinda meh because sig could be signature or a misspelling of sign. Also, libraries like torch use thresh for threshold I think.

What about padj_threshold? Could you maybe please also search the codebase for other parameters that do something similar or have a similar name? I wouldn't be surprised if we had some of these.