sumstats_summary when populations all have n=1

[Peter G]

Hello,

Ive run populations whereby i assign each individual sample (which is an individual organism) a unique population. this was primarily to construct a phylogeny (Which ive done with the help of the forum - thanks!), but i would like to know about weather the statistics STACKS produces for this kind of analysis are credible or if population sizes of 1 mean its nonsense? heres the first couple of lines. can this be interpreted that individual 252 has an  a nucleotide diversity (Pi) value of 0.1019 and is therefore more diverse  than sample 83 that has a Pi of 0.584??  (of course there are standard errors to consider etc, but as a trend.. ?

Would it therefore be acceptable to look for correlations between these values and environmental data i have per sample? 

# Pop ID	Private	Num Indv	Pi	Var	StdErr	Fis	Var	StdErr
83	42	1	0.0584	0.055	0.0019	0	46.1405	0
252	127	1	0.1019	0.0915	0.0028	0	44.0094	0
117	35	1	0.0629	0.0589	0.0019	0	45.9214	0

Finally, the file gives several statistics. some are likely to be of no value with population sizes of zero such as Fis. Which values are going to be useful out of 

Obs Het,

Obs Hom,

Exp Het,

Exp Hom,

Pi.??

Also, is it possible to get D values per sample/population? 

thank you. it would be a massive help to know more about this!

[Nonno Hasegawa]

Hi, has this been answered at all? 
I also had the same question - I ran populations with three different spatial scales of population maps with the same individuals. Im now wondering whether these sumstats calculated when n=1 for a population is meaningful. 

thanks!

[Catchen, Julian]

No, they would not be valid, similar to any standard statistical analysis with a sample size of 1.