This is a pre-release version of a website accompanying the article "The genetic landscape of Ethiopia: diversity, intermixing and the association with culture" which is currently in preparation. It describes an analysis of the genomes of 1,268 Ethiopian individuals from 68 different ethnic groups, along with other individuals from outside Ethiopia.
A good place to start is the help page.
See the Help page for more information on using this site.close this window
Each point on the map represents a single person in our study. The location reflects where this person is from (based on birthplace information). In total there are 1,268 individuals on the map. The colour scheme can be seen in the 'Selection population(s)' dropdown, and it reflects the classification of first language.
To compare genetic distance between individuals, we use a measure called total variation distance (TVD). The total variation distance is computed by modelling each individual's genome as a 'mosaic' of segments from other genomes - either those of other Ethiopians in the study ('within-Ethiopia' analysis) or those from other countries outsied Ethiopia (outside-Ethiopia analysis). Technically, these models are constructed using CHROMOPAINTER, and TVD is computed from the output vectors of copying proportions. The more similar two individual are, the more similar their copying vectors will be, and the lower their TVD.
TVD ranges from 0 (meaning completely similar DNA, depicted with large circles on the map) to 1 (meaning very different DNA, depicted as small circles on the map).
The value of TVD depends on the set of individuals that are used as source donors in the analysis. We provide two analyses here - a 'Ethiopia-internal' analysis that includes all individuals in our study (the Ethiopian individuals shown and further non-Ethiopian individuals detailed in our paper) as donors, and an 'Ethiopian-external' analysis that only includes non-Ethiopians as donors. Copying patterns from non-Ethiopians reflects DNA similarity over a longer timescale, evidenced by the generally shorter length of segments copied. Some populations, like the Shabo, are genetically quite different from other Ethiopian populations on recent timescales, but appear more similar under the Ethiopian-external analysis.close this window
This website accompanies the article "The genetic landscape of Ethiopia: diversity, intermixing and the association with culture" which is current in preparation.
Each point on the map represents a person in our study, located based on their birthplace. The initial colouring shows the classification of their first language.
If you hover or click on a person (for example, the individual labelled EAF4676a), the points will be replaced with blue and red circles of different sizes. This is a way of showing how similar the DNA of these individuals is: larger circles represent people that are more genetically similar to the chosen individual, and smaller circles those that are less similar. The circles are coloured to show individuals in the same ethnic group (red) or different ethnic groups (blue). The 'Selected populations' box at the top of the screen shows the label of the selected individual. Clicking on an individual allows you to also hover over a second population (drawn in orange) for comparison. (Click elsewhere on the map to reset the selection.)
Technically, we measure genetic similarity using a metric called the total variation distance (TVD). TVD ranges from zero to one and measures how different the genomes of two individuals are, relative to other individuals in the study. We think of this as measuring 'genetic distance'; the 'genetic similarity' is then 1 - TVD so that larger circles represent people with more similar genomes.
The box on the lower left of the screen shows a different visualisation of the genetic distances from all other individuals to the chosen individual: it shows the distribution of the genetic distances from the chosen individual to all other study individuals. The height of each bar represents the proportion of individuals study with the given genetic distance (as shown on the x axis) to the chosen individual. Again, the plot is split into red bars (which include individuals in the same ethnic group), orange bars for any second group selected, and blue bars for everyone else.
The box on the lower right compares genetic and geographical distances between whole groups at a time. These are computed by averaging over individuals in each group. You can sort this plot by average genetic distance, or by average geographical distance using the sliding switches.
Finally, the 'Selected population(s)' box at the top of the screen allows you to select populations or language groups for comparison in the same way; the map and boxes then show genetic distnaces averaged over the individuals in the group. You can also use the menu bar to select the way that the genetic distances are computed - either by comparing each individual to all Ethiopian individuals in the study, or by comparing to a further set of non-Ethiopian individuals. Because genes tend to spread slowly, these two analyses can be thought of as capturing genetic similarity over different timescales - the 'Ethiopia-internal' analysis captures more recent similarity while the 'Ethiopia-external' analysis captures older timescales. Our paper describes some of the ways these differ.
For more details, see the technical details.close this window