So I have some mitochondrial control region (CR) sequences from several populations, and I'm calculating Tajima's D for each population in DnaSP and Arlequin. Each program agrees with respect to the values for Tajima's D for each population.
Interpretation of p-values in DnaSP seems rather easy: it tells you whether the p-value is less than 0.05, or between 0.05 and 0.10 (anything greater than 0.10 is simply given as p > 0.10).
Arlequin presents p-values differently: it presents the proportion of simulated values of D that are less than the observed value (P(D simul < D obs)). So, for example, in one population, DnaSP returns a result of p < 0.05, and in Arlequin, the same value for D is associated with P(D simul < D obs) = 0.99230, which makes some sense, since D > 0.
But there's one population that has me a bit perplexed:
D = 2.34103
DnaSP says that 0.05 < P < 0.10
Arlequin, on the other hand, gives me: P(D simul < D obs) = 0.97645
I'm not asking which result to present; rather, I want to know how to interpret the p-value given by Arlequin, and whether the result presented is significant at alpha = 0.05. Does it means that P(D simul > D obs) = 1 - 0.97645 = 0.02355?
I'm sure this is an elementary question, so thanks for your time and patience!
It would help if you could either copy/paste the relevant part of your result file OR link the file itself. But yes, since the simulation creates an expected distribution of your data under no selection, if your estimated value of D exists in the tail of that distribution, it is significant at the level equivalent to the proportion of values that were smaller in the simulation.
My guess about the difference in P between DNAsp and Arlequin is the number of simulations performed.
Hello Everyone,
I need a little help I analyzed sequences by tajima's D and fu's Fs neutrality tests and I found negative values for Tajima's D and a not significant p value as well as for fu's fs I found some positive values and the p value is not significant, what does that mean?
Othmane Daoui it either means your sample was too small to detect selection OR that your sequences are not under detectable selection.
Joel D Anderson what did you mean by your sequences are not under detectable selection
Both of those tests (D and Fs) are meant to test for deviation from the null model of neutral evolution. A significant P-value for D or Fs indicates that the sequences you are analyzing have demonstrated significant deviation from the neutral model. The fact that none of your values are significant indicates EITHER that your sequence set is not subject to natural selection OR you have obtained a false negative. In other words, P-values that indicate no significance tell you that you have detected no evidence for selection, population expansion, hitchhiking, or any other process that might cause those statistics to deviate from what the data would look like under a neutral evolution model.
- Badges
- Science topic
- Similar topics
- Biological Science
- Zoology
- Ornithology