Difficulties 'poppr' function

[Hilaire KABORE]

I have genotyping-sequencing data from 151 isolates containing 102,464 SNPs. When I run poppr function, it runs for hours without giving anything at the end. Thinking it was related to the capacity of my computer, I ran it on a server and got this error message.

*** caught segfault ***
address 0x2b81d5e6b000, cause 'memory not mapped'

Traceback:
 1: .Ia.Rd(popx, missing)
 2: .ia(poplist[[x]], sample = sample, method = method, quiet = quiet,     missing = missing, hist = FALSE, namelist = namelist)
 3: FUN(X[[i]], ...)
 4: lapply(sublist, function(x) {    namelist <- list(file = namelist$File, population = x)    .ia(poplist[[x]], sample = sample, method = method, quiet = quiet,         missing = missing, hist = FALSE, namelist = namelist)})
 5: poppr(genclone_exsero)
An irrecoverable exception occurred. R is aborting now ...
/var/spool/slurm/slurmd/job21271156/slurm_script: line 10:   985 Segmentation fault      (core dumped) Rscript stratification_genind.r

[Zhian Kamvar]

Hello, 

You are running out of memory because the covariance matrix needs nearly 40GB of memory for that operation alone:

cells <- choose(102464, 2) # number of cells needed for covariance matrix
rows <- ceiling(sqrt(cells)) # number of rows in an equivalent square matrix
memuse::howbig(nrow = rows, ncol = rows)
#> 39.111 GiB

If you want an index of association estimate, you should convert your data to snpclone (assuming you came from vcfR): https://knausb.github.io/vcfR_documentation/export_genlight_snpclone.html and then use either `samp.ia()` or `win.ia()` to calculate the index from subsamples of loci.

All the best,
Zhian

--
You received this message because you are subscribed to the Google Groups "poppr" group.
To unsubscribe from this group and stop receiving emails from it, send an email to poppr+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/poppr/7180cbb3-3b98-4389-9d7b-5e07a87558c8n%40googlegroups.com.