How to Use G25 Coordinates in Vahaduo

Before you start

Vahaduo does not read raw DNA files. It works only with G25 coordinates, so you need two things: your own coordinate line, and a set of reference populations in the same scaled format to compare it with.

No coordinates yet? Follow Get G25 Coordinates From Your Raw DNA first, then come back here.

Step by step

1. 1

   Have your coordinates ready

   You need your own line of scaled G25 coordinates: a name followed by 25 comma separated numbers. If you only have a raw DNA file so far, convert it first.

2. 2

   Load reference populations into SOURCE

   The SOURCE tab holds the populations you want to compare yourself against, one per line, in the same scaled format. Paste them in, or use the DATA tab: press OPEN FILES to load a reference spreadsheet, type to filter it, and add the matching rows to the source.

3. 3

   Paste your line into TARGET

   The TARGET tab holds the sample being modelled, which is you. Paste your coordinate line exactly as you received it, with no header row. You can add several targets, one per line, for example yourself and your parents.

4. 4

   Read the DISTANCE list

   DISTANCE ranks every source by how far it sits from your target. The closest populations come first. This is the quickest way to see which groups you resemble most before you build any model.

5. 5

   Build a model in SINGLE

   SINGLE explains one target as a mixture of the sources, giving a percentage for each and a distance for the fit. A lower fit distance means the mixture reproduces your coordinates more closely. The cycles option trades speed for precision, and add dist col adds each source's distance to the table.

6. 6

   Compare several targets in MULTI

   MULTI runs the same model for every target in the TARGET tab at once and lays the results out side by side, which makes family comparisons and before and after checks easy.

Common mistakes

• Mixing scaled and unscaled coordinates. Scaled values sit roughly between minus 0.15 and plus 0.15. If one line has much larger numbers, it came from a different format and every distance will be wrong.
• Leaving a header row in SOURCE or TARGET. Both tabs expect data only, with no header line.
• Putting commas or line breaks inside a sample name. The name is the first comma separated field, so an extra comma shifts every value by one.
• Loading hundreds of sources into one SINGLE model. With too many options the model can fit almost anything. Start with a small set of plausible ancestral groups for the period you are asking about, then add or remove one source at a time.
• Reading a percentage as a fact about your ancestors. A model shows which mixture of the chosen sources best matches your coordinates. Different source sets can fit almost as well, so compare fit distances across models rather than trusting one result.

When you want more than a spreadsheet

Vahaduo is excellent for hands-on modelling, but you supply and curate every reference sheet yourself. MYDNATLAS runs the same kind of distance and mixture analysis against a maintained reference catalogue, adds period by period breakdowns and plots, and keeps your coordinates portable: you can export them back to Vahaduo at any time.

Questions