After performing the IBD analysis using the "R" Package adegenet 1.4-0, the program obtains a histogram with X : simulated values , Y: permuted values , and a dot which is the correlation between geographical coordinates matrix and a matrix of genetic distances ...
In the figure I can observe that there is an isolation by distance since the correlation value is higher than most of the sim values ... could any one please give a brief distribution of what the sim values and the permuted values exactly are ??
The dot represents your actual correlation between the two distance matrixes (genetic distances and geographical distances). The distribution represents the correlation values you obtain when data is permuted, that is, when you expect no geographic structure. You can find more information in the tutorial vignettes of the adegenet package.
Hi,
What you see here is the distribution of simulated (spearman or pearson) based correlations of your genetic distance matrix (e.g pairwise Fst or pairwise individual distance) and pairwise distance matrix. So what the algorithm does is, take the euclidean distance matrix and shuffle it (permutate it) then a simple pearson correlation between your geneitic distance matrix is calculated. This is done 999 (default) times,then the distribution of these values is plotted as histogramm= Sim by frequency and your value is plotted in between. If your value is extreme (above 97.5 or below 2.5 percentile normally) you have a higher ( or lower ) r with your original data then expected by chance and conclude you have isolation by distance if above 97.5 of all values. Theoretically you could have familiarity by distance (
To add to the above comment, the sim values are the simulated values (i.e. the distribution of simulated values according to a model with no IBD) and the y axis shows the number of times (frequency) that the simulated values were obtained during the analysis. Thus, the histogram is showing you a distribution of correlated values that you expect in the absence of IBD, while your dot shows you the correlation for your data. Hope that helps :-)
In my mind, "simulated values" and "permuted values" are one and the same thing, and are represented on the X-axis. Personally, I prefer using the expression "expected values".
This figure provides the distribution of expected (or simulated or permuted) Mantel correlations (X-axis) under the null hypothesis of no IBD, that is, "simulated" by random permutations of pairwise data (probably 10000 permutations in your case).
The Y-axis stands for the frequency of each class of expected (or simulated or permuted) correlation values : for instance, more than 2500 permutations (out of 10000) led here to an expected correlation ranging from -0.1 to 0.
The dot stands for the observed correlation between your genetic and euclidean distances (here around 0.45). Actually, you can see that the observed correlation is rather high when compared to the expected distribution of correlation values under the null hypothesis, although the shape of the distribution is not "classic". The Mantel test (associated with this figure) may give you a proper information regarding the significance of the IBD in your data.
Hope it helped !
Good luck =)
Hi all, I think that the different comments above are really helpful! You can also have a look on adegenet webpages where there are plenty of tutorials, explanations, help, and a friendly forum! Here is the webpage adress: http://adegenet.r-forge.r-project.org/
Here, is the link to the vignettes that Dr Perfectti mentioned to you,
http://cran.r-project.org/web/packages/adegenet/vignettes/adegenet-basics.pdf
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Statistical Software
- R Statistical Package