No announcement yet.

Autosomal STR

  • Filter
  • Time
  • Show
Clear All
new posts

  • Autosomal STR

    I know FF is autosomal SNP, but I hope that asking a question about autosomal STR tests is not off-topic, even if not strictly FF.

    So the thing is that I have two persons for both of which are taken autosomal STR tests, i.e., alelle values for a (subset of=) chromosomes with STR genetic loci. I think similar tests are used in paternity/maternity cases, but in this particular case it is to establish if two persons are direct siblings (share both parents), possibly half-siblings (share one parent). DNA from parents is not available. An mtDNA test is pending.

    I assume that the inheritance rules for STRs are the same as for SNPs, i.e., on average 50% inherited from one parent, and 50% from the other. Anyone has any further insights? Of course, while FF tests 1000s of SNPs here it is only a few 10s of STRs.

    Two direct siblings will, on average, share 50% of STRs, two half-siblings only 25%. The chance that two direct siblings share only 25% of STRs is a little less than 1%.
    What I don't know is what is the chance that two random persons - who are not close siblings at all - share a certain percentage of STR? The reason is that I don't know the distribution of STR alelle values for a given loci.

    The investigation office has come back with a statement that the likelhood of direct siblings based on STR results is about 1:100, i.e., chance not being direct siblings is 100x larger than being direct siblings. Of course mtDNA results will help a lot. If mtDNA is different the answer is simple. If it is the same, then the question of sibling/half-sibling remains open. And if mtDNA is different but the mother is different, then the half-sister hypothesis remains open too. So I want to calculate the likehood under all these assumptions. And what I need is the likelhood that two random people share a certain part of autosomal STR.

    Anyone has some ideas/suggestions on further reads? Hope it is clear?!

    === edit ===
    If this test was for two males and it was a YDNA STR test (i.e., loci on one chromosome, inherited from common father), then the calculations are easy because one knows STR distributions and mutation rates. But my case is different, although bears some resemblance at the same time.
    Last edited by lc0; 12 April 2021, 05:54 PM.

  • #2
    Hi there, I've got no idea, but you can check the other threads. There is a detailed discussion regarding this though can't remember what specific thread is it located.


    • #3
      For the case of autosomal STR's, there are two sources of variation from one generation to the next. Both sources are very unlikely events.

      First, the STR (a short sequence that contains a repeating pattern) can itself mutate, so that the number of "repeats" changes. We can assume that the STR's have been selected for study based on data showing that the number of repeats does NOT change very often. Thus, the probability that the STR itself has mutated is going to be a very rare event per generation -- and in the case of establishing parentage, the number of "generations" is trivial.

      Second, the STR could be altered by recombination. There will be two copies of the STR, one on the chromosome from the mother, the other on the homologous chromosome from the father. STR's are so short that the probability of a recombination event within the STR is again exceedingly small.

      The bottom line is that the autosomal STR's selected for this kind of analysis are extremely unlikely to change over a period of a generation or two, and are therefore well suited for this purpose, especially when the analysis includes say, a dozen or so STR's scattered throughout the genome.

      If something like the Family Finder test is available for the individuals, the STR results should be entirely consistent with the Family Finder results, because they are intended to measure essentially the same thing, the degree of autosomal matching between individuals.


      • #4
        It is better to include visual presentation to better understand. It is quite technical