Overview

tajima can be used to calculate Tajima's D across a sliding window or using bins. For an explanation of what Tajima's D is, see this excellent video by Mohamad Noor.

In order for a SNP to be incorporated in the calculation, it must:

  • Have an allele frequency greater than 0 and less than 1.
  • Be biallelic.
  • Be a SNP.
  • Be a diploid site.

Usage

Parameters:

  • window-size - Size of window from in which to calculate Tajima's D.
  • step-size - Size of step taken.
  • --sliding - Fluidly slide along genome, capturing every window of a given window-size. Equivelent to step-size = 1;
  • --no-header - Outputs results without a header.
  • --extra - Adds on filename, window-size and step-size as additional columns. Useful for comparing different files / parameters.

Tip

You can specify window-size and step-size using commas or scientific notation (e.g. 1,000,000 or 1E7).

Output

tajima will output the following columns:

  • CHROM
  • BIN_START - Starting interval position inclusive.
  • BIN_END - Ending interval position not inclusive.
  • N_Sites - Number of sites used to calculate Tajima's D.
  • N_SNPs - Number of SNPs present in interval. Certain sites are excluded.
  • TajimaD - Tajima's D calculation.

If you optionally specify --extra, the following columns will also be included in output:

  • filename
  • window_size
  • step_size

Examples

Tajima

The figure above illustrates the types of windows over which Tajima's D can be calculated.

Bin calculation

If you set the window-size and step-size as the same value, the bins will not overlap. This is depicted in the figure above as 'Bin'. 

vk tajima 1,000,000 1,000,000 <vcf>

The code above will calculate Tajima's D using 100,000 bp bins across the genome.

CHROM BIN_START BIN_END N_Sites N_SNPs TajimaD
I 0 1000000 24 8 -0.344142
I 1000000 2000000 47 20 0.666153
I 2000000 3000000 34 18 0.418091
I 3000000 4000000 22 10 -0.676877
I 4000000 5000000 11 4 -0.652344
I 5000000 6000000 8 2 -0.498306
I 7000000 8000000 8 4 -0.537028

Sliding window

When run, the code below will calculate Tajima's D across a 100,000 bp sliding window that moves 1,000 bp with each iteratino.

vk tajima 100,000 1,000 <vcf>
CHROM BIN_START BIN_END N_Sites N_SNPs TajimaD
I 6000 106000 2 1 -0.740994
I 7000 107000 2 1 -0.740994
I 8000 108000 2 1 -0.740994
I 9000 109000 2 1 -0.740994
I 10000 110000 2 1 -0.740994
I 11000 111000 2 1 -0.740994

Continous sliding window

When run, the code below will calculate Tajima's D across a 100,000 bp sliding window that captures every unique bin of variants that fall within 100,000 bp of one another.

vk tajima 1E5 --sliding <vcf>
CHROM BIN_START BIN_END N_Sites N_SNPs TajimaD
I 0 100000 2 1 -0.740994
I 90777 190777 2 2 -0.0333856
I 154576 254576 2 1 0.690099
I 207871 307871 2 1 -0.740994
I 263709 363709 2 1 -0.740994
I 321321 421321 2 1 -0.740994
I 294407 394407 3 1 -0.740994
I 391250 491250 3 2 -0.110617