Migration Parameter

[GoogleCodeExporter]

Although the format of the input for MaCS is similar to that for ms, there are 
some differences (e.g. in the scaling of the mutation and recombination rates). 
The documentation does not state in what format the migration parameter M for 
an n-island model should be given. In ms, M=4Nm where N is effective population 
size of one island and m is the proportion of inhabitants of an island that 
come from other islands each generation. Is this also the case in MaCS?

Original issue reported on code.google.com by jackson....@gmail.com on 1 Aug 2014 at 1:11

[GoogleCodeExporter]

Yes the migration parameters are consistent with ms

Original comment by gche...@gmail.com on 2 Aug 2014 at 1:28

• Changed state: Accepted

[GoogleCodeExporter]

When I ran the commands:

ms 2 10000 -t 1200.0 -r 239.999992 30000000
macs 2 30000000 -t 4e-05 -r 8e-06 -h 10000

for 10000 simulations, I obtained the results shown in histograms ms1.png and 
macs1.png. As expected, the distributions of number of segregating sites in the 
generated data are almost identical.

However, upon running the commands:

ms 2 30 -t 4000.0 -r 799.999992 100000000 -I 9 2 0 0 0 0 0 0 0 0 400.0 
macs 2 100000000 -t 4e-05 -r 8e-06 -h 10000 -I 9 2 0 0 0 0 0 0 0 0 400.0

(in this case for 30 simulations), I obtained the histograms shown in ms2.png 
and macs2.png. As you can see, the distributions are very different.

What am I doing wrong? 

Original comment by jackson....@gmail.com on 8 Aug 2014 at 3:33

Attachments:

• ms1.png
• macs1.png
• ms2.png
• macs2.png

[GoogleCodeExporter]

Original comment by gche...@gmail.com on 28 Jan 2015 at 10:16

• Changed state: WontFix