R session crashes when fitting betta model about breakaway HOT 6 CLOSED

Hi @bkmontero -- thanks for opening this issue. Can I please confirm that you have restarted R and repeatedly seen this same crash/error message? It looks like breakaway is not installing properly, or that R's pointers are off, which can usually be addressed with a quit and restart. I also strongly encourage the use of R Projects to avoid pointer issues.

Could you please confirm that you see the same error after a restart, and that you are working in a R Project? Once we hear from you we will dive in deeper.

from breakaway.

Hi Amy,

Thanks for your quick reply!

Yes, I am working in a R project, I consistently get the same error after restarting R, (and after restarting my computer) and also after re-installing breakaway. I am not using pointers.

I am happy to share the data with you if that helps!

from breakaway.

Hi @bkmontero! If you're able to provide a minimal working example that allows me to reproduce the error I can help you troubleshoot this issue. :)

from breakaway.

Hi @paulinetrinh!

Thanks so much for your help! :)

Files:
ps_asv_tab.txt
ps_metadata.txt

# data
asv_tab <- read.table("~/ps_asv_tab.txt"), header = TRUE)
metadata <- read.table("~/ps_metadata.txt"))

# phyloseq object
OTU <- otu_table(asv_tab, taxa_are_rows = FALSE)
samples <- sample_data(metadata) 
ps_coel <- phyloseq(OTU, samples)

# estimate richness using breakaway
rich_coel <- breakaway(ps_coel)
rich_coel

meta <- data %>%  sample_data %>% as_tibble %>% 
          mutate("sample_names" = ps_coel %>%  sample_names)
meta %>% names
combined_rich_coel <- meta %>% left_join(summary(rich_coel), by = "sample_names")

# fit betta model to breakaway estimates
estimates <- combined_rich_coel$estimate
ses <- combined_rich_coel$estimate
X <- breakaway::make_design_matrix(ps_coel, "Island_Mainland")
betta(estimates, ses, X)$table

Hope this works out, I'm a newbie at posting issues! 😬

from breakaway.

@bkmontero Thanks for this! I've been looking into this and I realized that you had said you filtered out the "spurious" ASVs. breakaway actually works best when you do not filter out these less abundant ASVs as we need those for more accurate estimation of diversity. If you try running breakaway on an unfiltered ASV table do you still get an error?

If you do get an error still and are able to share that table I can help troubleshoot from there! :)

from breakaway.

Hi @paulinetrinh

I just realised there might be something wrong with my R installation/computer. When I run the same analysis on my laptop everything works fine. So now I have to figure out why it's not working on my desktop 😕
Any suggestions are much appreciated :)

PS. Thanks for your comment about keeping low abundant ASVs- I'll keep it in mind!

from breakaway.