Hi to everybody,
  In this opportunity, I let to know the answers that I received about the problem of 
  interpretation mismatch distribution and raggedness index. The problem was:Results
  that the curve is unimodal but the signicance of raggedness index is
  not significant under the null hipothesis of neutral evolution. The
  analysis was performed in DNAsp with coalescent simulations considered
  no-recombination and observed theta values.  The question is: The
  populations suffer or not an expansion? 
Almost responses coincide in that the first step is realizing a neutrality test (e.g., Fu 1997), 
  if is this statistic is significant negative the next step is perform a mismatch distribution, to 
  characterize the expansion. In addition, the program Arlequin 3.00 (Excoffier et al 2005) permits 
  test the demographic and spatial models of expansion with bootstrap approach.
   
  This message is also for Allegra Briggs and Arnaud Bataille (I hope that this responses are 
  useful for you).
   
  Thanks very much to everyone who replied, your answer were very useful.
   
  Ahh!! I forgoted, excuse me for my poor english.
 
Best regards,
  Juan Jose Martinez
Cat. Genetica de Poblaciones y Evolucion
FCEF y N - UNC
Cordoba, Argentina
juan_jmart@yahoo.com.ar
   
  >
  Si el test no es significativo no puedes rechazar la Ho (pob. 
estacionaria). 
Sin embargo los analisis basados en la mismatch distribution son poco 
potentes, 
(ver Ramos-Onsins y Rozas 2002) y sobre todo para regiones nucleares 
(con 
recombinacion), aunque creo que trabajas con regiones mitocondriales.
Te aconsejaria que miraras el comportamiento de los estadisticos R2 
(Ramos-
Onsins y Rozas 2002) y el Fs de Fu (Fu 1997).
  Saludos,
  Julio Rozas
   
  >
  Yes, mismatch distributions are difficult to interpret. I have never  
done the analysis in DnaSP but it is probably similar to Arlequin,  
which I use.
  It is ok to have significant fit to a unimodal curve but non- 
significant raggedness index because they are measuring slightly  
different things.    Mismatch is describing pairwise differences  
between haplotypes but raggedness is the variation around the curve.  
An empirical mismatch distribution that does not deviate from a  
unimodal distribution of pairwise differences among haplotypes and  
has a smooth distribution (Harpending, 1994) suggests recent  
population expansion (Slatkin and Hudson, 1991; Rogers and  
Harpending, 1992). In other words, a mismatch distribution, P > 0.05  
means you can’t reject the null hypothesis of population expansion. A  
non-significant raggedness index means you have a relatively good fit  
of your data to a model of population expansion. At any rate, this is  
how I understand it.
  You might find some helpful information at http://www.rannala.org/ 
gsf/, especially under Arlequin re. mismatch distributions.
Also see  Harpending, H. C. (1994). Signature of ancient population  
growth in a low-resolution mitochondrial DNA mismatch distribution.  
Human Biology, 66, 591-600. for discussion of raggedness index.
 Rogers, A. R., & Harpending, H. (1992). Population growth makes  
waves in the distribution of pairwise genetic differences. Molecular  
Biology and Evolution, 9(3), 552-69.
 Slatkin, M., & Hudson, R. R. (1991). Pairwise comparisons of  
mitochondrial DNA sequences in stable and exponentially growing  
populations. Genetics, 129, 555-62.
  for some discussion.
  Good luck,
Kathryn
   
  >
  Supongo que la razón or la que tus resultados no son significativos 
bajo 
la hipótesis de  evolución neutral, es precisamente porque la población 
está se está expandiendo. En DNAsp puedes realizar diferentes tests de 
neutralidad.
  Si tu población se está expandiendo, tus resultados deben ser negativos 
y no significativos en el test D de Tajima (lo cual indica desviación 
de 
neutralidad debido a expansión, selección o "hitchhiking"), resulados 
negativos y no significativos de los tests D* y F* de Fu & Li (lo cual 
indicaría que no hay selección o hitchhiking) y resultados negativos y 
significativos del test Fs de Fu (lo que indicaría crecimiento 
poblacional).
  Has tratado de usar el programa ARLEQUIN? En el puedes hacer análisis 
de 
mismatch distribution y probar si tus datos se desvían de dos modelos 
diferentes: uno de expansión puramente demográfica y el segundo de 
expansión geográfica.
  Espero que esto sea de ayuda. Y perdón por mi español, estoy un poco 
fuera de práctica.
  Saludos... Rodrigo
   
  >
  The test is evaluating a null hypothesis of a population expansion.   
The failure to reject the null hypothesis (the non-significant  
raggedness index) indicates that you don't have any support for a  
stable (non-expanding) population - which is thought to result in a  
highly multi-modal, and therefore ragged, signal.  Failure to reject  
the null hypothesis of expansion doesn't absolutely mean that your  
population has undergone one, but doesn't support any alternative  
hypotheses of stability.
  Hope this helps,
Hayley
   
  >
  No estoy muy familiarizado con el programa DnaSP pero me parece que 
estas 
interpretando mal los resultados. Los test para rechazar el modelo de 
neutralidad que debes hacer son Fu & Li 93, Fu 97 y/o Tajima 89, creo 
que 
todos ellos los hace DnaSP. Existen otros test similares y puedes ver 
una 
revision sobre ellos en: http://dx.doi.org/10.1007/s00239-003-0027-y. 
Creo 
que el test sobre el índice de "raggedness" construye la distribucion 
de 
este índice bajo la hipótesis nula de una expansión demográphica con 
los 
parámetros estimados por la analisis de la "mismatch distribution". Es 
decir 
que el "valor-p" te indica, cuanto mayor es, una concordancia mayor 
entre 
tus datos y los parámetros estimados. Al menos asi es como funciona en 
el 
software Arlequin.
  Mi consejo es que: 1) hagas un test de neutralidad, éste es el que te 
va a 
indicar si ha habido o no una expansión demográfica. 2) si tienes 
evidencia 
de expansión demográfica (test de neutralidad significativo) entonces 
puedes 
hacer el analisis de la "mismatch distribution" para caracterizar esa 
exansion (es decir estimar la edad de la expansión, y los tamaños 
poblacionales antes y después).
  Espero que esto te sea útil, suerte
  Miguel
  
Miguel de Navascués, BSc(Hons), PhD
http://m.navascues.googlepages.com/
   
  >
  Perhaps you might try an alternative analysis that could also point  
out if your population expanded. In DnaSP there is an option for  
Tajima's D and Fu's Fs - if the values are significantly negative,  
the population might have expanded in the past. Alternatively, try to  
run the mismatch analysis in Arlequin, where you also have an option  
to bootstrap the analysis. Also in Arlequin you can run the indeces  
of selective neutrality that are sensitive to demographic population  
expansion (Fu's Fs).
  These result together might help you to understand the nature of your  
data.
  Best regards,
Natalia Martinkova
  
 

juan_jmart@yahoo.com.ar