Hi guys, I need some insights here. I have a microsatellite data for two populations of O. rufipogon and O. sativa for comparison. I just want to compare their genetic diversity population structure of the two wild rice populations. I initially used 128 SSR markers but only used 48 after removing loci that are monomorphoc, non-amplifying, and with more than 5% missing bands. I computed the Nei‘s genetic distance (1973) and constructed from it a neighbor-joining tree. I am actually happy with the result as it confirms that the two populations of rufipogon are unique indeed based on the used loci. However, I saw one strain from one population (nap13) which was grouped closely together with the O. sativa genotypes (see attached photo). I am wondering what causes this phenomenon and how to deal with it. Your answers will be highly appreciated.
Dear Sandy,
As i see in the figure, the given strain (nap13) was not grouped in any of two subpopulations. I suggest the use of another software for clustering such as STRUCTURE.
Asadollah Ahmadikhah Yes, I am considering using STRUCTURE. What settings can you recommend for STRUCTURE based on this parameters: 48 loci for 48 individuals? Thank you so much :D
I would also recommend using dapc analyses through the R environment.
Matthijs P. van den Burg I already tried DAPC but I don't really understand it (please find 4 photos). Can you help explain these results? Also, I tried using R but it will only run properly if the burn-in period and MCMC is set to 500. I found out that in Structure nap 13 (19) is closely grouped to the rufi population than to sativa.
I constructed another NJ tree using Nei's (1973) genetic distance with 999 bootstraps. Here is the result.
In neither tree do you have support values. I am guessing that the first tree is not bootstrapped and is instead simply the neighbor joining tree with no measures of confidence in each branch. The second is presumably the same except that you have bootstrapped it so you have 999 trees. The bootstrap values show what proportion of trees have branch in the tree which is a consensus of all the trees. If this is so, then the second tree should have bootstrap values on each branch. Here you are especially interested in the bootstrap support for the branch connecting nap13 to the other nap samples. What is that bootstrap value?
Patrice Showers Corneli How can I display the bootstrapped values? I used powermarker to generate the tree file and used Mega to visualize it. I am also thinking of removing all the loci with missing data. If I do that I will be left with only 10 ssr markers. Is it ok?
In STRUCTURE set up burning period on 100000 and MCMC simulations on 100000 too. And Run program with 10 repeats.
I see no need to remove missing data as phylogenetic analysis will just ignore them.
The bootstrap values are extremely important because otherwise you can not know whether or not each branch is well supported. In your case, the confidence in position of the nap13 gene cannot really be assessed without the bootstrap.
The bootstrap will tell you precisely how well your data supports the monophyly of the nap clade given the model you choose.
With respect to Powermark and the bootstrap, the manual suggests that you can get bootstrap values by inputting the output of multiple trees into Phylip and using consense. This is sensible, But you might try using the Distance option with bootstrapping in Seaview (Gouy, M. Guindon, S. & Gascuel., O. (2010) SeaView version 4 : a multiplatform graphical user interface for sequence alignment and phylogenetic tree building. Molecular Biology and Evolution 27(2):221-224).
Seaview is free and very good. When you do the bootstrapping with Distance and tree will pop up with the bootstrap values and the branch lengths. You may either read the bootstrap values from that tree or export the tree file as an unrooted tree and open it in Figtree (also free and very good) and choose to look at the bootstrap value.
Powermark suggests that the output is such that it can be exported to other programs for further analysis (eg Principal Components Analysis).
You may find a slightly different tree depending on the models you used for Powermark and for Seaview . If you run into any difficulty email me please.
Patrice Showers Corneli Thank you so much. I will now work on it based on your recommendation :)
I attach a majority-rule bootstrap consensus tree from the tree files you sent. I did this one in PAUP and as you can see much of the tree has little support which probably means that the phylogenetic information in the data are quite sparse.
I cannot tell from your data which number corresponds to nap13 so I do not know whether it is in one of the few clades that is supported by good bootstrap support. Here the clade including 43-47 is clearly monophyletic (all more closely related to one another than to any other sample in the tree)with 95% of the bootstrap replicates having this clade. Also 26 and 15 are well supported as monophyletic. If either of these contain all nap specimens and contain nap13 then you are pretty safe saying that nap13 should be included in the larger nap clade.
If these clades contain nap13 and non-nap specimens, then you have some sense that nap13 is more closely related to other specimens than to the nap clade.
Finally if any 13 is not is either of the clades with 90% or higher support then you cannot really tell where it belongs.
I also note that your DAPC output is a principle components analysis. I do a l to of PCA so would be willing to look at your output if you would like. I just need to know what the list of variables is.I assume it is the SSR data.
Patrice Showers Corneli Thank you so much, Nap13 is number 19. The Apo Clade is number 1-14, while the Nap Clade is from 15 to 41, and the Osa Clade is from 42-48. Both Apo and Nap are wild rice (O. rufipogon) accessions while the Osa are cultivated rice accession. For this study, I actually used microsatellite markers for cultivated rice maybe that's why the Osa clade has strong bootstrapped values. My question is what does it imply if most of the tree has little support values?
- Badges
- Science topic
- Similar topics
- Agricultural Science
- Agronomy
- Crop Science
- Crop