Re: [dadi-user] Projection Choice - 2D models

[Ryan Gutenkunst]

Hello Kaden,

There is no reason to set projection values equal among populations. (I regret doing it for the first dadi paper.)

Best,

Ryan

On Jan 16, 2026, at 7:36 PM, Kaden Winspear <kadenw...@gmail.com> wrote:

Hello Dr. Gutenkunst,

I have a question regarding projection choice in dadi. I am running 1D and 2D models for two populations with unequal sample sizes (20 and 30 individuals), which results in unequal projection values. For my initial analyses, I chose projections that maximized the number of segregating sites (34 and 46). The 1D models appear reasonable.

However, in the 2D analyses I consistently see long residual tails in the dadi plots. This made me wonder whether differences in projection values alone can bias 2D inference.

Is there any known issue or recommendation regarding unequal projection values in 2D dadi analyses? Is it better to set projection values equal for these sort of models? 

Thank you for your time.

Best,
Kaden Winspear

--
You received this message because you are subscribed to the Google Groups "dadi-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dadi-user+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/dadi-user/13f28d57-b553-4648-9c96-dbb87c91fc67n%40googlegroups.com.

[Kaden Winspear]

Thank you for the help!

To view this discussion visit https://groups.google.com/d/msgid/dadi-user/77A117F5-3325-4D56-BEB3-C2DAE75DEA29%40gmail.com.