max_range_size Fatal Error

[Sébastien Mirolo]

Hi Nick and everyone,

I'm new to BioGeoBEARS and ancestral area reconstruction (ancestral range estimation, I did my research) in general. I'm trying to analyse a small dataset of 16 species on 10 different areas. As the analysis takes ages to run with that many areas, I would like to set the max_range_size to 3 or 4 but it gives me this error message when I check the parameters with check_BioGeoBEARS_run(BioGeoBEARS_run_object) :

FATAL ERROR in inputs: max_tipsize=6 > inputs$max_range_size=3

I understand that it's because one of my species have a present range of 6 areas but I don't get why it's a problem to constrain the reconstruction. I'm working on bryophytes, a large geographic range is almost the rule. Is there any way to run the analysis without decreasing the number of areas?

Very best,

Sébastien

[Julia Dupin]

Hi Sébastien,

with max_range = 3 you can only estimate ranges of up to 3 areas. If one of your species has a range of 6, then your max range has to be at least 6, otherwise the model cannot even estimate that input range.

I'll also add that 10 areas for a 16-tip tree is too much. At times, we might want to be as precise as possible and define many areas to represent in detail the distribution of a set of species, but this results in a large number of areas that does not necessarily help answering a question.

Julia

[Nick Matzke]

Yep that's it Julia -- if max_range_size=3, then all tips must be 3 areas or less.  

(You could, if you want, use a manual_states_list and time-stratify it to allow 6-area ranges from e.g. 0-1 million years ago, and then only 3-area ranges before that, but then you'd have to 

(a) justify this decision, this time-point cutoff, etc.

(b) do the setup, which is complex, but based on: http://phylo.wikidot.com/example-biogeobears-scripts#modified_states_stratified

Thanks for the answer, I've been travelling!

Cheers,

Nick

(PS 10 areas for a 16-tip tree is physically possible on the computer, 

--
You received this message because you are subscribed to the Google Groups "BioGeoBEARS" group.
To unsubscribe from this group and stop receiving emails from it, send an email to biogeobears...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/biogeobears/a08d6b04-afdb-4787-b833-56b836fc49ben%40googlegroups.com.

[Nick Matzke]

Meant to add as a PS:

(PS 10 areas for a 16-tip tree is physically possible on the computer, the issue is whether one has a strong justification for postulating 2^10 = 1024 possible different geographic ranges, when you have only observed a maximum of 16 different ranges. It may well be that collapsing some areas, or limiting the maximum range size, makes more sense in this situation.)

[Viktor Baranov]

Hi Nick and everyone!

I am getting a similar error message,

> check_BioGeoBEARS_run(BioGeoBEARS_run_object) Read 8 items Read 372 items Error in if (max_tipsize > inputs$max_range_size) { : missing value where TRUE/FALSE needed

I am fairly sure though that max_tipsize is always one in geogfn file, and I played with different settings for max_range_size, but none, solved the problem.

Any Idea what might be causing the error?

Thanks a lot!

Viktor