Scenario : Statistics not high enough
In a sibling comparison, the two siblings don’t have to share DNA in common at all the markers like a father and child. This means that the outcome of a sibling comparison relies solely on statistics. In a sibling case, we are trying to determine if it is more likely that they share two common parents, one common parent or no common parents. To put it another way, are they full siblings, half siblings, or unrelated. We write two conclusions in our reports. The first conclusion assumes that the participants have the same mother, or one common parent, and gives the probability of whether they are full siblings or half siblings. The second conclusion assumes that the participants have different mothers and reports the probability of whether they are half siblings or unrelated.
Case 1
We've listed this case first because it illustrates perfectly why testing more markers will produce better quality results. We collected samples from our employees and their families, voluntarily of course, and performed sibling comparisons. One case stood out because two sisters, we knew were full siblings, were being reported as having a 95.39% probability they were half siblings with the 16 marker test. So, why did the 16 marker test produce a false result? The sibling DNA test relies on statistics and, with a limited number of data points (16 markers), the statistics were wrong. Even though a high percentage of sibling cases with 16 markers will report what is reality, because it’s statistics, the calculation in some cases will get it wrong. In order to lower the possibility of this happening, we need to test more data points. When we examined these sisters with the 41 marker test, we found the probability to be 97.15% that they would be full siblings.
| CASE 1 | More Likely To Be | Probability |
|---|
| 16 Markers |
Half-Sibling |
95.39% |
| 41 Markers |
Full-Sibling |
97.15% |
Case 2
The two siblings in this case were adults whose parents had passed away. They had different mothers and need to know if they were half siblings. They'd always had a suspicion that they might not share the same father and turned to DNA testing to put their minds at ease. The results of the 16 marker test found that the probability of being unrelated was only 78.83% and it seemed that their suspicion might be correct. However, the percentage was not high enough to conclusively say they weren't related. When we applied the 41 marker test to their case, we found that the probability they did not share the same father was 98.66%.
| CASE 2 | More Likely To Be | Probability |
|---|
| 16 Markers |
Unrelated |
78.83% |
| 41 Markers |
Unrelated |
98.66% |
Case 3
In this sibling case, we were asked to determine if the two possible siblings were half siblings or unrelated. We were able to get samples from both mothers. We found that the probability they did not share the same father was 71.67% after testing all four people at 16 markers A percentage this low cannot be trusted. After testing with the 41 marker test, results indicated that the probability of being unrelated increased to 99.96%.
| CASE 3 | More Likely To Be | Probability |
|---|
| 16 Markers |
Unrelated |
71.67% |
| 41 Markers |
Unrelated |
99.96% |
Case 4
In this case, the two men had the same mother and the standard 16 marker sibling test was performed. It was found that the probability of being half siblings was only 75.23%, which is not enough to be conclusive. When the 41 marker test was applied, they were found to be more likely to be full siblings at 99.77%. Noticing the dramatic change in the results lead to testing the two men’s Y-chromosomes. When asked to compare two male siblings there is an optional test that can be performed known as a Y-chromosome test. The Y-chromosome is male DNA passed down identically from father to son. Two brothers from the same father should have identical Y-chromosome profiles. The Y-chromosome profiles were found to be identical and a probability of paternal line relationship was calculated to be 99.97%. The conclusion that can be drawn from cases like this is that 16 markers won’t be enough in some cases and the field of relationship testing should reconsider what they’re willing to report as conclusive.
| CASE 4 | More Likely To Be | Probability |
|---|
| 16 Markers |
Half-Sibling |
75.23% |
| 41 Markers |
Full-Sibling |
99.77% |
|
