Can someone help me in understanding the relation between Tajima's D test and ZnS test for linkage disequilibrium? And how does recombination influence the statistics? Basically, the Tajima's D test reveals the deviation of neutral model at intron 1, however, ZnS test shows that the exon 1 and 2 show significant LD. These observations seem to be not consistent to each other. How can I explain this? Thank you!
Hi Tânia
Why do you think your observations are not consistent? What is the sign of the Tajima's test?
Exon 1 and 2 could be at disequilibrium just for proximity (usually, LD decays within a few Kb), or because they are linked to a marker under selection. Could this be some part of intron 1?
Hope this helps. But do not hesitate to ask further
Pablo
Hi Pablo, thank you for your answer and your availability. I am really having difficulties in understanding this one on my own… The sign of TajD is negative and significant, being their 95% confidence intervals result from coalescent simulations conducted considering all segregating sites and an intermediate level of recombination. And here is also where I think one of my issues lies. Let me try to explain:
I have sequenced a specific gene of several populations of an outcrossing plant species. Using DnaSP one can calculate TajD and ZnS indexes without recombination, considering full recombination or a given (estimated) value of recombination per gene. I would think that the most consistent would be using this last value, estimated from the fragment that I am using for the analyses. However, this for me gets into a circular reasoning (as I use the data to estimate a value that is crucial to the significance of that data for a given index). Plus, because of the separation between exons and introns, the R/gene estimated is extremely variable depending on the amplicon (from more than 200 on exon 1, to 0 at the intron, and 50 on exon 2). Considering the all gene, I would be talking of an estimated R/gene of around 20. Should I use this value for the simulations?
I am really uncertain of the best approach here.
Thank you so much for the tips.
Dear Tânia:
You have a couple of issues here. First, TajD. As you probably know, a significant negative value of TajD means either a recent selective sweep (or linkage to it) or a recent population expansion following a bottleneck. In order to distinguish between the two possibilities, you need to compare "many" genes (as population expansion should be detected on all sequences, whereas a selective sweep would be exclusive to some selected gene).
On the other hand,the DnaSp manual (the help included in the program) indicates that for Coalescent Simulations the intermediate level of recombination is the case for most nuclear genes (except if you hve a hotspot) and that you should use the value of per gene recombination parameter R. Therefore, do not separate exons and introns. Use the whole gene sequence for your estimates.
Finally, please note that you can use ZZ=Za-ZnS to detect intragenic recombination.
Hope this helps
Pablo.
Dear Pablo,
Thank you so much.
Yes it helps! I will look at it and read a bit more, now after your insights (maybe I will still get back with even more doubts!).
Best wishes,
Tânia
- Badges
- Science topic
- Similar topics
- Biological Science
- Genetics
- Genetic Analysis