Issues with very small number of spots

[lmweber]

Add some guidelines in vignette / documentation mentioning minimum suggested numbers of spots, i.e. the model tends not to work well at very small numbers (e.g. 100 spots). From feedback from @lahuuki (thank you!)

Possibly also trying different random seeds can help in these situations.

[lmweber]

Also consider this issue in the context of models with spatial domains: do we need to recommend a minimum number of spots per domain? (e.g. >= 100 spots per domain, and otherwise only running models across the whole slide, i.e. without covariates for spatial domains)

[lmweber]

Or simply a warning message notifying the user that we have a small number of spots and models can be unstable in this case

[lmweber]

Addressed in tutorial / vignette in this section.