Phasing: Separating Mom's DNA From Dad's

Open your raw DNA file and every position shows two letters. What it does not show is which letter came from your mother and which came from your father. Sorting that out is called phasing, and it turns a jumbled list of pairs into two coherent stories - one from each parent.

The problem: two letters, no labels

At each position on your autosomes you carry two alleles, one inherited from each parent. Your raw file records both, but as an unordered pair. A line might read:

rs4988235	2	136608646	AG

That AG tells you that you have one A and one G at this spot. It does not tell you whether the A came from your mother and the G from your father, or the other way around. Multiply that ambiguity across hundreds of thousands of positions and you have a file full of pairs with no sense of which side each letter belongs to.

What phasing does

Phasing assigns each allele to the maternal or paternal copy of your DNA, reconstructing your two haplotypes. A haplotype is the run of alleles that traveled together on a single inherited chromosome - essentially the intact string your mother gave you, and the separate string your father gave you.

Once phased, that ambiguous AG becomes ordered: the A sits on one parental copy, the G on the other. Do this across a chromosome and you recover two continuous sequences instead of a column of unordered pairs. If you want a refresher on what those columns mean in the first place, our raw file anatomy tour walks through each one.

Why phasing is worth the trouble

Sorting alleles onto the correct parental copy sharpens almost everything downstream:

• Better relative matching. Knowing which alleles travel together lets a matching algorithm recognize genuine shared segments and reject chance alignments, cutting down on false matches.
• Mapping segments to a side. Phasing helps assign a shared DNA segment to your maternal or paternal line, which is invaluable for genealogy when you are trying to work out which branch a match belongs to.
• Cleaner haplotype analysis. Many analyses assume you know which variants sit together on one chromosome, and phasing provides exactly that.

In short, unphased data tells you what you carry; phased data tells you how it is arranged, and arrangement is where a lot of the useful signal lives.

How phasing is actually done

There are two broad approaches, and they differ in reliability.

The first is statistical phasing. Algorithms use large reference panels of known haplotypes and the fact that nearby variants tend to be inherited together to infer the most probable arrangement of your alleles. It works without any family data, but it is a best guess, and it becomes less certain over long distances and at rarer variants.

The second, and more reliable, is phasing by comparison with tested relatives. If a parent has also been tested, the logic is direct: whatever allele you share with your mother at a position must be your maternal one, which leaves the other as paternal. Having one or both parents, or other close relatives, tested lets you phase large parts of the genome with confidence rather than probability. This is why serious genealogists prize a tested parent.

Where phasing fits in your exploration

Phasing is a quiet workhorse behind good relative matching and segment mapping, even when a tool never shows you the word. Understanding it explains why matching improves when relatives test and why some analyses can attribute a segment to one side of your family. For the bigger picture of how shared segments become relationship estimates, see how DNA relative matching works. And whatever you explore, you can keep your raw file on your own device with on device DNA analysis rather than uploading it.

This article is educational and is not medical advice.