error outputting PDF outfile

[Elizabeth Strasko]

Hello,

I've been running some test runs on MIGRATE (the MPI version on a supercomputer) for a few months now, with relatively few issues. Recently, I've been sending bigger runs, and the results have not included a PDF file. My other test runs thus far have not had this issue. I'll attach my parmfile. Thank you for any advice you may have on this - I know there are ways to read the results without it, but as a graduate student who is new to this, I've really liked referencing the PDF outfiles. 

Thanks,

Liz

[Peter Beerli]

Dear Liz,

sorry for the slow response,

I guess the failure of the PDF printing may be due to a buffer/memory overrun. How many loci and populations do you run?

Your parmfile contains a few small things that I would change:

without replication, these options:

gelman-convergence=YES:SUM

 is not doing anything; the use of the convergence measure is brittle in that it often does not really work. I suggest running the program a few times and comparing the parameter estimates. I usually increase the long-inc by a factor of 5x or 10x if the results are stable; I assume convergence was fine.

change 

long-chains=3

to 

long-chains=1  #3 will not do anything; I need to remove this option as it is not functional anymore

experiment with 100, 500, 1000 (see above)

long-inc=100

geo=YES:geofile.txt

often this does not seem to be needed, as you want the geneflow between locations and not geneflow scaled by distance per location

mutation=DATA

if you have sequence data, I would set this CONSTANT, as this will make it easier to compare your parameter values per locus, as they are all scaled by the same mutation rate

mig-histogram=ALL:0.001:mighistfile

skyline=YES:0.001:skylinefile

Without dated tips, these are often poorly estimated; I suggest not using them in early runs (or not at all)

pdf-terse=NO

if you have 1000s of loci, consider using 

pdf-terse=YES

this reduces the size of the report data table, etc

best

Peter

--
You received this message because you are subscribed to the Google Groups "migrate-support" group.
To unsubscribe from this group and stop receiving emails from it, send an email to migrate-suppo...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/migrate-support/1469afe8-4495-4cb9-98d4-5eed47b35572n%40googlegroups.com.