Your ancestry composition in a worldwide context

One of ADNTRO’s main features is describing your ancestral makeup of every individual as well as to leverage their genetic information with the results from a curated selection of Genome Wide Association studies that are available (GWAS), which help you to unlock the stories your genetic code has to tell..

This methodology is based on a Nature publication (one of the most prestigious scientific journals in the world) and the 1000 Genome database which does not allow us to provide the detailed country breakdown from your ancestry calculator in ADNTRO, but it is a very robust method with a solid scientific basis.

To achieve the description of the ancestry composition, we use computational pipelines (workflows for "big data" - large data sets - that use the computational power of Google Cloud) to compare your DNA with that of many other individuals around the world, in this case using the 1000 Genome database as a starting point, which comprises 2,504 genetic profiles of anonymous individuals from 26 different world populations (see the additional information section below for more details).

Based on your genetic similarity, to others in these databases, the computer pipelines make predictions about how much of your ancestry comes from different places around the world..

Missing from this endeavour, however, is how to put your results into the context of other populations. How similar is your DNA profile to that of people from your part of the world and to that of people from other parts of the world? Many people around the world have ancestries that are unique to the region where they are from, but many others also have mixed ancestries that may reflect human migration patterns in the distant past or in recent times (such as having a parent from a different country or even a different continent).

ADNTRO has collaborated with the Dr. Joshua Bantaa professor of Biology at the University of Texas at Tyler who studies and publishes about the genetics of natural populations, using big data to make a visual tool available to you that puts your genetic results in a worldwide context. This tool is based on your own DNA, combined with the ‘1000 Genomes’ database of worldwide genetic diversity, comprising 2,504 anonymous individuals’ DNA profiles from 26 different worldwide populations (see “additional information” section below for further details)..

Now for the gritty details. We use a non-negative matrix factorization algorithm (sNMF) implemented by the 'snmf' from the LEA package package in the open source statistical program R. This algorithm estimates the contributions to the genome from hypothetical ancestral groups of every individual in the analysis, based on the genetic similarities/differences of the DNA sequences to one another. For further details on the methods, see “additional information” section below.

To find the optimal number of ancestral groups (K), different numbers of hypothetical ancestors were simulated, and the different simulations were compared to one another to determine which one fits the data the best and, therefore, which number of K groups should be used for further purposes. We tried different numbers of K groups ranging from 5 – 12 using the snmf function’s entropy criterion that evaluates the quality of fit of the statistical model to the data using a cross-validation technique ((see "Additional information" section below (1)).

A model with a certain number of K groups is favoured when the entropy criterion is minimized (see "Additional information" section below (2)).

Based upon these criteria, we find that the optimal number of K groups is 5, meaning five different main ancestral groups (Figure 1).

For more details about the methods, you can consult the additional information at the end of the article).