Confusion about data formatting

76 views
Skip to first unread message

bbarker505

unread,
Apr 13, 2011, 3:51:27 PM4/13/11
to popABC
Hi all,

I am completely new to using approximate Bayesian computation analyses
programs and discovered popABC a couple of days ago. There is probably
a very simple solution to my issue, but I was hoping somebody could
help me. I am confused with regards to the input files. I have 5
sequence datasets (1 mtDNA + 4nDNA). I converted them all into .nexus
formats that could be read by popABC and then made .len tables from
them. It seems that by doing it this way though that only one locus
can be analyzed at a time (i.e. 5 different analyses). When I was
going through the example files provided by Dr. Lopes it appears that
he used a multi-locus ".pop" format datafile that had all loci
together in the same file. Does such a multi-locus .pop file have to
be made by hand or is there a program that can do this? I saw that
there is a tool for converting .len files to .pop files, but again it
only appears to do one locus at a time. Also, how do you make a multi-
locus .pop file when there is 2 nDNA sequences per individual but
only 1 mtDNA sequence per individual? Does the second mtDNA sequences
have to be "?" for missing data (as with programs such as Structure or
Migrate when you analyze mtDNA together with nDNA sequences)?

Thank you for any clarification on how to best format my input files.
I really appreciate any help.

Best,
Brittany



Joao Sollari Lopes

unread,
Apr 14, 2011, 7:22:17 AM4/14/11
to popABC
Hi Brittany,

You can actually get away with editing the .len file directly.
Just change the number of loci used and put 5 's's in the following
line. Afterwards just add the "tables". Just make sure that the order
of the populations (if you're using more that 1) is consistent in all
5 loci.

As for the difference between n and 2n loci in the .len file, this
will only affect the number of samples you have. Specifically you will
have double the amount of samples in the 2n loci.
Don't forget, however, to add the correct scalars in the .prs file.
And make sure to be consistent regarding the order you add the loci
"tables" in the .len file.

Hope that helps,
Joao

bbarker505

unread,
Apr 15, 2011, 9:15:05 AM4/15/11
to popABC
Thank you Joao! That worked. I knew it was probably something really
simple.

On Apr 14, 5:22 am, Joao Sollari Lopes <j.sollari.lo...@gmail.com>
wrote:
Reply all
Reply to author
Forward
0 new messages