I had RFLP tests done to confirm siblingship with my 5 newly found brothers and sisters. The lab results indicated more than I ever thought possible... namely that we were, with well over 98% probability, FULL siblings!

While I loved this news, it just didn't look right from the results I saw, so I decided to ask one of my brothers to compare y-str dna with mine...and we didn't match like we should have...only 4 out of 12 markers matched...the other brother matched the first one and not me, too.

Now I'm back to half siblingship, but puzzled at how the labs could have such totally differing results ... is RFLP worthless or is there a valid question about the reliability of Y-STR PCR techniques?

************ What is RFLP testing *************

At each location along the DNA chain (those we are interested in right now, anyway), there are two *ALLELES*. This is important, because each of these two alleles came randomly from one or the other parent. For instance if your father had AA at this location and your mother had BB at the same location, then you will have AB at that location on your DNA chain. If your father had AB and your mother had AB, then you could have AA, AB or BB. But you won't know which of your two parents each marker comes from. What is informative is if you mother is BB, your father is AA and you are CA,then you know your father is not the man your mother is married to parents

Actually, the test is much more discriminating than this simple analogy. The alleles I have called A and B are usually represented by numeric values that reflect the number of "base pairs" found at the location of interest.

Base pairs are the repeating segemnts comprised of the amino acids that make up our DNA. So if you see a value of 3.55 for location D3Sxxx, it means that there are 3,555 bp (base pairs) at the location defined by D3Sxxx.

Now for some good news. The size of the allele at location D3Sxxx will have certain frequency in the population as a whole. It may be rare, common or anywhere in-between. If you consider the possiblility of matching some else at 8, 15 or more of these "markers" you will find it highly unlikely that anyone other than two siblings with common (to one another) parents would match up.

A pair of full siblings will show a high percentage of DUAL MATCHES in their profiles, where half siblings will show, on average, one half of that...but in either case, far more than RANDOM matches that might occur in the population at large.

The final analysis of these is done by mathematical computation and models. The CODIS markers used by the FBI for identification is based upon these techniques. In their world it is virtually impossible to mis-identify someone...but it HAS been known to happen...after all there are over 5 BILLION of us alive and in this gene pool but for all practical purposes we can rely on the technology and techniques but we have to understand, errors can occur.


If you have any further comments, questions, opinions or whatever ... please do contact me by email at wcscott@wcscott.net .

Thanks for reading this and I hope it is a pleasantly challenging puzzle for you ...
Chris Scott


Back to Home page