problems with likelihood analisys in POY 5.1.1

21 views
Skip to first unread message

lourd...@gmail.com

unread,
Aug 15, 2017, 12:48:27 PM8/15/17
to POY - Phylogenetic Analysis Software
Hello, 

I am submitting the following commands:

read(prealigned:("16S.fas",tcm:(1,1)))
search(max_time:00:00:30)
select()
transform(likelihood:(gtr,rates:gamma:(4),priors:estimate))
search(max_time:00:04:00)
report("arbol.tre", trees:(branches), "reporte.txt", lkmodel)
exit()


I am running this small analysis just to test the performance of POY (in a parallel environment: 40 processors) when doing a likelihood analysis. 

I am getting the following error (right after the Wagner is finished):

poy: likelihood.c:2544: readjust_brents_gtr: Assertion `b >= x' failed.
[amazonia02:61548] *** Process received signal ***
[amazonia02:61548] Signal: Aborted (6)
[amazonia02:61548] Signal code:  (-6)
poy: newkkonen.c:1599: backtrace: Assertion `len1<=len2' failed.
poy: likelihood.c:2544: readjust_brents_gtr: Assertion `b >= x' failed.
[amazonia02:61542] *** Process received signal ***
[amazonia02:61542] Signal: Aborted (6)
[amazonia02:61542] Signal code:  (-6)
poy: newkkonen.c:1633: backtrace: Assertion `(i== -1)&&(j==-1)' failed.
poy: likelihood.c:2544: readjust_brents_gtr: Assertion `b >= x' failed.
[amazonia02:61543] *** Process received signal ***
[amazonia02:61543] Signal: Aborted (6)
[amazonia02:61543] Signal code:  (-6)
poy: newkkonen.c:1599: backtrace: Assertion `len1<=len2' failed. 


Please I would really appreciate help tot interpret this error message. 

Lourdes 

Nicholas Lucaroni

unread,
Aug 15, 2017, 3:00:02 PM8/15/17
to POY - Phylogenetic Analysis Software
Shucks. I had run those functions (newkonnen and brents method) very regularly with large/small datasets and no issues ever came up like this in a long time. And since both functions failed, i'm guessing something else is going on -- well, hoping.

- What is the distribution of lengths for your sequence data? Minimum / maximum length is probably fine.
- What C compiler/toolchain are you using or are you using a distributed binary? (gcc/clang/...)
- Is this reproducible? (does every run of the script fail?)
- Does the analysis work sequentially? 

I cannot make guarantees that i'll be able to dig into this, those are both fairly subtle complicated sections of the code.
Reply all
Reply to author
Forward
0 new messages