Hi Mahmoud,

First of all, the three modalities, i.e. IAM, TPM and SMM are not methods but mutation models. As you use microsatellites, the IAM mutation model should not be very informative. Second, you only have five markers, which seems quite low to detect for bottleneck at a genome wide level. However, if you carefully read at the documentation and associated papers, you might find some interesting clues on how to discuss your results. Maybe in your case the Wilcoxon test is the most adapted one, as it has been demonstrated to be the most accurate one in case of low number of polymorphic loci. It is based on the shape of the distribution of allele frequencies, which should be L-shaped in case of no bottleneck and skewed if there is bottleneck. In rou case it's seems that the only significant test was the one for heterozigosity excess, which should be the signature of a recent expansion of the population.More, this test does not rely on any mutation model assumption (by the way if you look at your result file, the Wilcoxon tests results are the same wichever mutation model considered)

See : Piry, S., Luikart, G., Cornuet, J.-M., 1999. Computer note. BOTTLENECK: a computer program for detecting recent reductions in the effective size using allele frequency data. J Hered 90, 502–503. doi:10.1093/jhered/90.4.502

and especially this one : Luikart, G., Allendorf, F.W., Cornuet, J.-M., Sherwin, W.B., 1998. Distortion of allele frequency distributions provides a test for recent population bottlenecks. J Hered 89, 238–247. doi:10.1093/jhered/89.3.238

for an example of use, see : Chaïr, H., Duroy, P.O., Cubry, P., Sinsin, B., Pham, J.L., 2011. Impact of past climatic and recent anthropogenic factors on wild yam genetic diversity. Molecular Ecology 20, 1612–1623. doi:10.1111/j.1365-294X.2011.05041.x

In all the case, these tests are interesting, but you rather go for a coalescent-based one which should be more accurate in detecting recent change in population effective size, like for instance msvar from M.A. Beaumont.

Storz, J.F., Beaumont, M.A., 2002. Testing for Genetic Evidence of Population Expansion and Contraction: An Empirical Analysis of Microsatellite Dna Variation Using a Hierarchical Bayesian Model. Evolution 56, 154–166. doi:10.1111/j.0014-3820.2002.tb00857.x

Girod, C., Vitalis, R., Leblois, R., Fréville, H., 2011. Inferring Population Decline and Expansion From Microsatellite Data: A Simulation-Based Evaluation of the Msvar Method. Genetics 188, 165–179. doi:10.1534/genetics.110.121764

Good luck with your analysis !

Hi Mahmoud,

First of all, the three modalities, i.e. IAM, TPM and SMM are not methods but mutation models. As you use microsatellites, the IAM mutation model should not be very informative. Second, you only have five markers, which seems quite low to detect for bottleneck at a genome wide level. However, if you carefully read at the documentation and associated papers, you might find some interesting clues on how to discuss your results. Maybe in your case the Wilcoxon test is the most adapted one, as it has been demonstrated to be the most accurate one in case of low number of polymorphic loci. It is based on the shape of the distribution of allele frequencies, which should be L-shaped in case of no bottleneck and skewed if there is bottleneck. In rou case it's seems that the only significant test was the one for heterozigosity excess, which should be the signature of a recent expansion of the population.More, this test does not rely on any mutation model assumption (by the way if you look at your result file, the Wilcoxon tests results are the same wichever mutation model considered)

See : Piry, S., Luikart, G., Cornuet, J.-M., 1999. Computer note. BOTTLENECK: a computer program for detecting recent reductions in the effective size using allele frequency data. J Hered 90, 502–503. doi:10.1093/jhered/90.4.502

and especially this one : Luikart, G., Allendorf, F.W., Cornuet, J.-M., Sherwin, W.B., 1998. Distortion of allele frequency distributions provides a test for recent population bottlenecks. J Hered 89, 238–247. doi:10.1093/jhered/89.3.238

for an example of use, see : Chaïr, H., Duroy, P.O., Cubry, P., Sinsin, B., Pham, J.L., 2011. Impact of past climatic and recent anthropogenic factors on wild yam genetic diversity. Molecular Ecology 20, 1612–1623. doi:10.1111/j.1365-294X.2011.05041.x

In all the case, these tests are interesting, but you rather go for a coalescent-based one which should be more accurate in detecting recent change in population effective size, like for instance msvar from M.A. Beaumont.

Storz, J.F., Beaumont, M.A., 2002. Testing for Genetic Evidence of Population Expansion and Contraction: An Empirical Analysis of Microsatellite Dna Variation Using a Hierarchical Bayesian Model. Evolution 56, 154–166. doi:10.1111/j.0014-3820.2002.tb00857.x

Girod, C., Vitalis, R., Leblois, R., Fréville, H., 2011. Inferring Population Decline and Expansion From Microsatellite Data: A Simulation-Based Evaluation of the Msvar Method. Genetics 188, 165–179. doi:10.1534/genetics.110.121764

Good luck with your analysis !

Philippe! Fantastic helpful comment include valuable papers frankly! Actually, we provide 12 microsatellite primers for this study. now, we finished the first part (5 loci) of mentioned investigation. I find the L-shape among all of the allele patterns via  P>0.05 for all.

 lots of hug

Mahmoud

Dear Seyed,

Philippe's comments are valuable. I attach an article on bottleneck with allozyme markers that could be useful for you as practical example. Pay attention to the mutation model! I used the IAM (suitable for allozymes), you should use TPM or SMM.

regards

Dear Mahmoud,

I agree with Ferruccio, you have to pay more attention for the mutation model used!!The SMM and TPM models are more appropriate for microsatellite data. In particular, the TPM model provides a more realistic picture of mutational events in microsatellite loci. 

Good luck!

Dear Mahmoud,

As Philippe said, the significant heterozigozity deficiencies that you found with the wilcoxon tests may suggest that your population has experienced a recent demographic expansion instead of a population bottleneck. However, this type of signals may also arise in other situations (i) they may correspond to false signal induced by isolation by distance (see Leblois et al 2006 Molecular Ecology), (ii) false signals of demographic change arising from particular gene flow patterns (i.e. asymmetric gene flow, e.g. for freshwater organisms, see Paz-Vinas et al 2013, Molecular Ecology), or (iii) by the recent arrival of rare alleles by immigration (see Luikart and Cornuet 1998; Piry et al 1998).

Such a deviation can also arise in the earliest stages of a bottleneck, when a transient excess of alleles may exists (especially for markers evolving under a SMM mutational model; see Cornuet and Luikart 1996).

You should therefore take into account the type of organism you are studying, and the deviations of your population from a Wright-Fisher model to interpret your result.

As Philippe suggested, it will be useful to complement your Bottleneck analyses with other complementary methods such as the one implemented in MSVAR (see Philippe’s answer) and/or the Garza & Williamson test (Garza & Williamson 2001, Molecular Ecology).

Further, I suggest you to take a look to the following articles that deal with the robustness of this family of methods to many deviations from the Wright-Fisher model (e.g. asymmetric gene flow, population sub-structure, isolation by distance…) or to mutational model a mis-specifications:

Chikhi et al 2010 Genetics

Paz-Vinas et al 2013 Molecular Ecology

Leblois et al 2006 Molecular Ecology

Faurby and Pertoldi 2012 Evolutionary Ecology Research

Peery et al 2012 Molecular Ecology

Best regards,

Ivan

When reading the output from Bottleneck, is the probability the significance (p) value?

Dear Michelle,

Yes, the probability reported in the output corresponds to the p-value computed by BOTTLENECK. Note that these p-values should be adjusted afterwards for multiple comparisons (in the case that you analyze & compare many populations).

Best regards,

Ivan

I am trying to inquire about bottleneck computation but fortunately I have got a great list of information from you all. Firstly, thank you all for your wonderful explanations.

In the same way could you please clarify me one thing. I am estimating genetic diversity of two isolated populations of same species. In one of the population the observed heterozygosities are more at few loci than that of the expected. I am not able to figure it out what could be the true cause. Bollteneck software had computed a L-shaped allellic pattern. Please suggest. 

Howdy Folks. Trying to interpret Bottleneck output reported in a manuscript (I've not worked with the software myself). 

How should the p-values output by BOTTLENECK be interpreted? I am assuming that, as with any typical statistical test, a LOW value means significance, correct? So, in the case of the example that Mahmoud posted above, there is (marginally) significant heterozygote deficiency in each population, but none of the populations are significant using a two-tailed test. 

Is that correct?

I have a similar question as this....we used AFLP,and this is one of my results

SIGN TEST

Assumptions: all loci fit I.A.M., mutation-drift equilibrium.

Expected number of loci with heterozygosity excess: 131.26

163 loci with heterozygosity deficiency and 152 loci with heterozygosity excess.

Probability: 0.01064

Assumptions: all loci fit S.M.M., mutation-drift equilibrium.

Expected number of loci with heterozygosity excess: 152.89

170 loci with heterozygosity deficiency and 145 loci with heterozygosity excess.

Probability: 0.20243

therefore whether this population bottleneck or significiant heterozygosity deficiency

Hello ! Thanks a lot for these very useful comments !

I used Genepop format data file to run my analyses. I am surprised because the mean number of individuals sampled per locus resulted in the summary file (mean_N) corresponds to the double number of sampled individuals in my real dataset !

Am I misinterpreting the results or the program is misinterpreting my data file ?

Thank you very much !

Hi everyone,

Is it possible to analyse data from dominant markers in Bottleneck ?

I am asking because I had this comment by a Reviewer :

"The Bottleneck program is inappropriate for analyzes with dominant markers, since it is a program that is based on heterozygosity. In addition, the authors are performing population analyzes in a single population that is likely comprised of highly related individuals, and this also interferes with the analyzes. Remove the Bottleneck analysis, and focused on analysis of patterns of relationship among individuals ".

Is that correct?

Thanks