• Test Your DNA!

Fewer SNPs Test Reveals Familial Relationships Up to Second Cousins

drawn image of a double helix of DNA

DNA Double Helix: demonstrating 2 part ‘ladder rungs’ composed of paired DNA bases.

Fewer SNPs Tested Means Faster and Cheaper Results

A new DNA analysis methodology — based on a quick swab test using fewer SNPs– allows high-certainty detection of 2nd cousins or closer. In order to determine if two people are related (they call second cousins distant relatives although we genealogists would consider them close relatives!) the research team used an index of relatedness: “index of chromosome sharing” (ICS). ICS compares the total cM two people share to a chart detailing the likely relationship between them. The chart is adjusted for the number of SNPs tested. In their prototype testing, they used 174,254 SNPs (on an Illumina chip) – less than half those used for matching relatives at the big three DNA testing companies. They then identify the probability that two people are second cousins or closer by referencing their charts.

Diagram of the methods used in the paper to determine 'distant' relatives.

Pairwise Kinship Methodology Overview

The researchers have developed two charts. The first, is for linear (L) relationships such as a mother- son pair or a grandparent-grandchild pair — i.e. one person is the ancestor of the other. The second chart is for collateral (C) pairs — i.e. where both people descend from a common ancestor such as a sibling, aunt/uncle to niece/nephew, or a cousin. They, of course, use thresholds for determining the IBS segments because thresholds are critical for differentiating between mechanical matching and actual matching especially at the low SNP density they are testing at. They have different thresholds depending on whether the comparison is linear (3 cM threshold) or collateral (4cM threshold).

What’s Different?

The methodology the researchers are using isn’t looking for Identical by Descent (IBD) segments. They only care if the two people share Identical by State (IBS) segments. This is different than what genetic genealogists do because it’s important to us that the segments two individuals share were inherited from a common ancestor.

Relevance to Genetic Genealogists?

The researchers plan to use this method to ID remains in the cases of emergencies such as flooding or building collapses. But it seems to me that this test, which may be cheaper because of the fewer SNPs tested, would be valuable for confirming closer relatives for those with  suspected parents, half siblings, or even recent brick walls in their family trees. Hopefully, since there are other uses than just genealogy, the test might become available for verifying relationships between those that are second cousins or closer.

It’s also interesting that they found that linear relationships allowed smaller segments to be used despite testing fewer SNPs. This fits in well for what I have said for a long time now, that vertical (linear) inheritance of smaller segments should not be confused with what you see in horizontal transmission of segments and therefore (I believe) FTDNA has based their cousin thresholds erroneously on what they see with known descent from grandparent to parent to child. This is why FTDNA believes that it’s valid to add up those small segments (1 cM and 3 cM) having examined them vertically (Linearly) but they falsely inflate your total cM when looking at cousins.

Why Didn’t You Include the Chart?

Well, interestingly, because this research was done on Japanese pairs and the authors themselves state that each population should have it’s own table generated. Which points out another conundrum that some of us in the genetic testing community have been discussing for a while – can you really apply one set of rules across all populations? I have always thought that the answer is no.

Kelly Wheaton frequently mentions how African Americans have smaller DNA segments in common with each other and I’ve discussed the small French Canadian segment sharing as well. Timber, for all it’s helpful nature, still punishes those descended from bottle-necked populations by over culling ‘population-wide’ segments. The problem for these testing companies is should they set the threshold standards for everyone uniformly? How do forward thinking DNA testing companies hone their product with these ethnicity-mandated threshold differences? I believe in the years to come, genealogy companies, unable to trust customer self-reporting, will need to look at their own ethnicity assignments per chromosome when assigning thresholds to segments. Surely this will be costly in computing bandwidth but you know, computing is getting faster and easier and cheaper all the time.

After reading the paper, please share with us your comments, ideas and suggestions below!

Sources: Pairwise Kinship Analysis by the Index of Chromosome Sharing Using High-Density Single Nucleotide Polymorphisms by Chie Morimoto et. al. 2016

Bookmark the permalink.

Comments are closed