detection of positive selection in population using Tajima's n Fu Li test

Question

[evo_genomics]

Kindly help me

I am doing the evolutionary analysis of gene in primates. Initially I estimated ka/ks ratio. Then I applied Tajim's and Fu Li test on human population to check positive selection in human population. My dataset size is 42 and I have out grouped the population with bonobo sequence.

the problem I have faced is that Tajima's d statistics is giving significant results (-2.4) at p value<0.01 but Fu Li result is not statistically significant (-1.6 at 0.1>p value>0.05).

Could I conclude on the basis of single test (Tajima's) that gene is under positive selection in human population.

[evo_genomics]

thanks everyone.

[David W]

Evo genomics -- I removed your answers below and made them into comments to your original question. QA and sites like biostars work best when the answer section is reserved for posts and answer the question. Followups, clarification and discussion should go into comments

Answer 1

[zeeefa]

As far as I know, negative Tajima's D (<0) suggests that negative (purifying) selection is operating. It indicates an excess of low frequency variation/polymorphisms.

[evo_genomics]

but my concept is negative Tajima's D shows positive selection and positive value shows purifying selection based on paper mentioned below.

As followed is mentioned in "Genomic insights into positive selection"

Positive values of Tajima's D arise from an excess of intermediate frequency alleles and can result from population bottlenecks, structure and/or balancing selection. Negative values of Tajima's D indicate an excess of low frequency alleles and can result from population expansions or positive selection.

Answer 2

[David W]

You have a negative estimates for D with and D* (Fu and LI's version), which is evidence of negative selection as your quote shows.

The two estimates differ in what they measure, D is about the difference between estimators for population size based on segregating sites and nucleotide diversity. D* is based on the number of singleton sites. That the estimates differ (slightly), might tell you something about the genes you are interested in.

Answer 3

[Siva]

I don't think you can conclude that the gene is under positive selection. There could be other evolutionary processes (e.g. changes in population structure) which can give significant D value. You might want to check this discussion.