Tutorial

1. What is iMKT?
2. General help
3. About this tutorial
4. How can I upload and analyze my own data?
   • Polymorphism and divergence data
   • multi-FASTA
5. Criteria to select an MKT method depending on your data and question(s) at hand
6. Are the selective regimes of the X chromosome different when comparing ancestral vs colonizing Drosophila populations?
7. Do genes located in regions of high recombination undergo adaptive evolution?
8. Can I perform these anayses locally? Continuing your analyses with the iMKT R package
9. Need more help?
10. Cite us!

What is iMKT?

General help

×

• MKT analysis: it allows analyzing the user's own population genomics data and estimating four selective regimes by applying four different MKT methodologies.
  
• PopFly data analysis: it allows analyzing the selective regimes of 13,753 protein-coding genes in 16 populations of D. melanogaster by using the data collected in PopFly Also, it allows performing statistical analyses by comparing different lists of genes to answer complex population genetics hypotheses.
  
• PopHuman analysis: it allows analyzing the selective regimes of 20,643 protein-coding genes in 26 human populations by using the data collected in PopHuman. Also, it allows performing statistical analyses by comparing different lists of genes to answer complex population genetics hypotheses.
  
• Help and tutorial: this resource will guide you through the iMKT site through a step-by-step worked example.
  

About this tutorial

How can I upload and analyze my own data?

1. Polymorphism and divergence data

×

×

×

×

2. Multi-FASTA analysis

The iMKT server also accepts multi-FASTA alignments as input. In this case, click the multi-FASTA switch as the input data type. Polymorphism and divergence data will be estimated on putatively neutral (4-fold degenerated) and putatively selected (0-fold degenerated) sites. Then, choose any MKT methodology.

×

These are the requirements of the multi-FASTA files:

1. Sequences need to be protein-coding sequences (CDS) and aligned.
2. The first sequence in the file is the reference sequence, which will be used to compute the degeneracy of each nucleotide position.
3. The last sequence in the file is the outgroup sequence, which will be used to compute divergence data and infer the ancestral alleles.
4. Any other sequences between the first and the last sequence are considered polymorphic sequences, which will be used to compute the Derived Allele Frequency of polymorphic sites. At least two sequences are needed.

You can see an example of a multi-FASTA file by clicking the option Use multi-FASTA example as input.

×

More than one multi-FASTA files can be uploaded. In this case, you can use some advanced options to perform more complex analyses. For example, the option Concatenate CDS files will perform an MKT for each file separately, and an additional MKT for the concatenated data of all uploaded files. Alternatively, if you already have a multi-FASTA file containing several concatenated loci, you can select the option Multi-locus file to input a single line with the coordinates of all loci in the alignment.

If you don't know how to extract CDS data from genomic alignments or process the DAF, we provide this Python script as well as public data corresponding to the Drosophila Genome Nexus and FlyBase reference sequence. Check the help option () to learn about the required inputs.

The Derived Allele Frequency (DAF) polymorphism data and divergence estimations computed by iMKT server will be available for download from the final report.

Criteria to select an MKT method depending on your data and question(s) at hand

Here are some basic guidelines that can help you choose an MKT methodology depending on your data and questions at hand:

• If single-gene data is used, the asymptotic method (aMKT) is usually not applicable, since aMKT requires a high volume of polymorphic data to fit the asymptotic function. In this case, you can either use the FWW method, the extended MKT (eMKT) or standard MKT.
• aMKT is usually appropriate when multiple genes are concatenated.
• The choice among standard MKT, FWW method or eMKT will depend on whether you want to remove and/or quantify negative selection. For data unaffected by negative selection, the standard MKT can be used. Both FWW MKT and eMKT remove negative selection, but only the latter allows its quantification. Furthermore, the FWW method discards more data than the eMKT.

Are the selective regimes of the X chromosome different when comparing ancestral vs colonizing Drosophila populations?

1. First, we need to select the genes to be analyzed. Because we are interested in the evolution of the X chromosome, we can use the Filter by chromosome option and select all the genes located in the X chromosome. We can select all the rows (a total of 2,169 protein-coding genes of the X chromosome) by checking the first checkbox of the table.

×
2. Next, we select the different populations. Because we are interested in the comparison of ancestral and colonizing populations, we select the Raleigh (RAL) population from North America, and the Zambia (ZI) population from Africa, as a colonizing and an ancestral population, respectively.
3. Finally, we need to choose an MKT test. Given that we are going to analyze a large number of genes altogether, we can use the asymptotic MKT, which will give a less underestimated α. For the purpose of this analysis, we concatenate genes, i.e., analyze all the selected genes altogether, as we are not interested in the selection estimates of individual genes but in the global pattern of X-specific genes.


×
4. After submitting the query, results are displayed in a complete report, which is divided into different sections.
5. First, details of the analysis show the selected test(s), whether the data was concatenated or not, the parameters used for the test(s) (if any), the populations selected, and the number of genes analyzed.
   

×
6. The next section displays a summary of the MKT tables and descriptive statistics for each population analyzed. We can see that the African population (ZI) has much more nucleotide variation in both polymorphic and divergent sites than the North-American population (RAL) (compare the values in Tables 1 vs 3, and 2 vs 4) as expected, since the African population is the ancestral one.


×



7. Finally, the next section shows the results of the selected MKT test.


×

×
8. In the table, values of the exponential model are shown (a, b, c). Then, the asymptotic α is shown with the 95% confidence interval (CI) around it. The α estimated with the standard MKT is also shown for comparison.
9. The results show that the African (ZI) population has a higher proportion of adaptive divergence in the X-linked genes than the North-American population (RAL) (see alpha asymptotic in RAL vs ZI). However, the difference is only very slight in the α estimated with the standard MKT (α original). These results are in agreement with what the population genetics theory predicts: large populations (i.e., ancestral populations) harbor more genetic variation and a more effective natural selection. Also, the standard MKT underestimates the estimated proportion of adaptive divergence (i.e., α original).

Do genes located in regions of high recombination undergo adaptive evolution?

1. First, we need to select the genes to be analyzed. Because we are interested in the evolution of genes located in regions of high recombination, we can use the Min recombination rate option. We set this option to 7cM/Mb and select all the rows (a total of 475 protein-coding genes located in regions that harbor a recombination rate higher than 7cM/Mb, according to Comeron et al., 2012) by checking the first checkbox of the table.


×
2. Then we select an MKT test. In this case, the number of genes selected is lesser than in the previous example and it might be not enough for an asymptotic MKT, so we select the Extended MKT method.
3. Finally, we check the option Compare against whole-genome distribution, which will compare the distribution of the selected 475 protein-coding genes against the distribution of all D. melanogaster protein-coding genes.


×
4. In addition to all the sections described in the previous example, the resulting report provides the distribution of α estimates in both the group of genes located in regions of high recombination and all the genes in the Drosophila genome. The significance of the differences among groups is assessed by comparing the confidence intervals (CI).


×
5. The results of this analysis show that genes located in regions of high recombination harbor a much higher proportion of adaptive divergence (α) than the general genome-wide distribution. Natural selection is more effective in regions of the genome of high recombination, while it is impeded by the Hill-Robertson effect in regions with low levels of recombination.

Can I perform these anayses locally? Continuing your analyses with the iMKT R package

Need more help?

Cite us!