Y-Chromosome Markers for Many Families

www.roperld.com

Go to Y-Chromosome Markers Families Comparisons
Go to Roper male-line Y-chromosome project web page.

Contents

• Special Case of DYS389-1 and DYS389-2 Locations
• Special Case of DYS464 Markers
• Variability of Markers
• Y-Markers Probabilities
• Data References

Many surname Y-chromosome projects are listed at http://freepages.genealogy.rootsweb.com/~allpoms/genetics1a.html and http://www.familytreedna.com/surname.asp.

Special Case of DYS389-1 and DYS389-2 Markers

The reason DYS389-1 is subtracted from DYS389-2 is because the true alleles are DYS389a=DYS389-1 and DYS389b=(DYS389-2) - (DYS389-1). (See the December 2002 issue of Facts & Genes newsletter of Family Tree DNA) One must subtract 389-1 from 389-2 before calculating relative mutations or calculating phylogenetic networks. If one does not do this one can get either one too many or one too few relative mutations. E.g.:

• 389a = N + 1 & 389b = M + 0, where N and M are positive; correct relative mutation = |N + 1 - N| + |M + 0 - M| = 1.
  (389-1) & (389-2) incorrect relative mutation = |N + 1 - N| + |M + 0 + N + 1 - M - N| = 2.
  Thus, the relative mutation appears to be 2 instead of the real 1.
• 389a = N - 1 & 389b = M + 1, where N and M are positive; correct relative mutation = |N - 1 - N| + |M + 1 - M| = 2.
  (389-1) & (389-2) incorrect relative mutation = |N - 1 - N| + |M + 1 + N - 1 - M - N| = 1.
  Thus, the relative mutation appears to be 1 instead of the real 2.

The first example could cause one to think there are two relative mutations when there is only one, and vice versa for the last example.

The general case is:
389a = N + n & 389b = M + m, where N and M are positive and n & m can be either postive or negative.
Therefore, correct relative mutation = |N + n - N| + |M + m - M| = |n| + |m|.
(389-1) & (389-2) incorrect relative mutation = |N + n - N| + |M + m + N + n - M - N| = |n| + |m + n|
Thus, the difference between the incorrect and correct relative mutations is |n| + |m + n| - |n| - |m| = |m + n| - |m|.

Note that marker DYS454 has the same value (11) for all families, both paleolithic and neolithic. Thus, I believe that this marker is not part of the junk DNA; it must be related to some important feature of humans that selects against it passing mutations down through generations. Perhaps a similar statement applies for markers DYS426 and DYS459a which show no more than one mutation in the 8,000 years or so between the ancient European ancestor and the paleolithic families neolithic families.

Special Case of DYS464 Markers

The following is a quote from the December 2002 issue of Facts & Genes newsletter of Family Tree DNA:

"Two reporting changes have been implemented for Markers 464a-d. Markers 464a-d are copies found at different locations on the Y chromosome. In about 1% of the test subjects, more than 4 copies will be present, representing Markers 464e, 464f, 464g, etc. If those additional Markers are found, they are now shown on the individual's Y-DNA DYS Values page, and on the Group Administrator's Generate Y-DNA Scores page. If these additional Markers are not found, the columns for them in the reports will not appear. At the current time, 7 copies are the maximum number of Markers found for this DYS.

The reporting of genetic distance for Markers 464a-d has also changed. Previously, when comparing the results for these Markers, each result that did not match was considered a mismatch. With new guidelines from our scientific board, the genetic distance calculation for Markers 464a-d has been revised. Results are always reported from low to high, when reading from left to right. When a mismatch occurs, it must be taken into consideration whether the number of apparent mismatches are a result of the order of presentation of the Markers. The order of the results for these Markers may make it appear as if there are more mismatches than are actually present. Here are two examples:

In the first example, the mutation did not cause the results to be reordered, so it is very clear that there is one mismatch. In the second example, if a 2-point mutation occurred (17-2 = 15) the loss of 2 repeats caused the results to be reordered. On the surface, it looks like there are two mismatches, but this illusion is caused by the results being reordered, and there is only 1 mismatch, the 17 becoming the 15. Remember, DYS 464a-d is a highly polymorphic Marker, in fact it appears to be the fastest moving marker in our entire test. (Polymorphic means rapidly changing!)

For Group Administrators, when comparing two individuals results, if it appears that there is more than 1 mismatch for Markers 464a-d, be sure to check the genetic distance report (Members Page of your GAP) to verify the number of mismatches following guidelines by the folks at the Lab for Molecular Science and Evolution at the University of Arizona."

For another example, see http://freepages.genealogy.rootsweb.com/~gkbopp/KINNEY/Research/gendistance.htm. I am grateful to Georgia and Tom Bopp for informing me that I had been calculating relative mutations incorrectly for the DYS464 markers.

The following is an Excel spreadsheet formula to calculate the relative mutations for the DYS464 markers:

=4-MIN(COUNTIF($A$1:$D$1,$A$1),COUNTIF($A2:$D2,$A$1)) -IF(($A$1=$B$1),0,MIN(COUNTIF($A$1:$D$1,$B$1),COUNTIF($A2:$D2,$B$1))) -IF(($A$1=$C$1),0,IF(($B$1=$C$1),0,MIN(COUNTIF($A$1:$D$1,$C$1),COUNTIF($A2:$D2,$C$1)))) -IF(($A$1=$D$1),0,IF(($B$1=$D$1),0,IF(($C$1=$D$1),0,MIN(COUNTIF($A$1:$D$1,$D$1),COUNTIF($A2:$D2,$D$1)))))

Thanks go to Professor Thomas T. Bopp of the Department of Chemistry at University of Hawaii (tbopp@hawaii.edu) for devising this formula. Tom has also given me the following explanation of why DYS464 must be treated in this special way: "Basically, I gather that DYS464 is not only a multiple-site beast (4 sites), it also shows an enormous spread in its frequency in the population. The spread is apparently so wide that the common model of a single-step random walk is in doubt. The extreme alternative model is the "Infinite Alleles" model, which essentially says that a single mutation event can, with equal probability, result in any one value from a group of possibilities. In other words, either it's the same (=0), or it's not (=1). As I understand it, the mechanism is actually not known, and it is a hard question, currently under intense study in the research community. The statements that are made about this are guesses based on the observed frequency distributions. The single step mutation at least has the virtue of being a tractable model, and not unreasonable."

If you put the spreadsheet cell names in as $A$1, $B$1, $C$1, $D$1, $A2, $B2, $C2 and $D2, then you can copy from one line to the next without making any changes.

To calculate the relative mutations for all 25 markers for two persons (starting in column 1 and row A), use:

{=SUM(ABS($A$2:$U2-$A$1:$U$1))
+4-MIN(COUNTIF($V$1:$Y$1,$V$1),COUNTIF($V2:$Y2,$V$1)) -IF(($V$1=$W$1),0,MIN(COUNTIF($V$1:$Y$1,$W$1),COUNTIF($V2:$Y2,$W$1))) -IF(($V$1=$X$1),0,IF(($W$1=$X$1),0,MIN(COUNTIF($V$1:$Y$1,$X$1),COUNTIF($V2:$Y2,$X$1)))) -IF(($V$1=$Y$1),0,IF(($W$1=$Y$1),0,IF(($X$1=$Y$1),0,MIN(COUNTIF($V$1:$Y$1,$Y$1),COUNTIF($V2:$Y2,$Y$1)))))}

The first term is an array that calculates the relative mutations for the first 21 markers. (Be sure to use values for DYS389b instead of DYS389-2.) The curly brackets {} around the entire expression indicate that it is an array. You don't type in those brackets; they are automatically put in by the spreadsheet when you end entry or changes by holding down the SHIFT and CTRL keys while you press the ENTER key; you must do this each time you enter or revise an array entry. The remaining lines calculate the genetic distance for the DYS464a,b,c,d markers using the special procedure described by FTDNA. This calculation begins with the value of 4, then subtracts one for every matching value between the two individuals. The COUNTIF function counts the number of occurances of each value for one individual, and the MIN function yields the number of matches. The later portions of the test use IF statements to prevent duplicate counting of matches. The contribution from DYS464 will thus be a number between 0 and 4.

If anyone detects any errors in the formula above, please inform me.

Variability of Markers

The 25 markers have quite different variablility among the families in this study. Here is a graph of the maximum minus the minimum values of the markers for each of the 25 markers:

Markers DYS385a, 385b, 458, 447, 449, 464a and 464b have variability (=maximum-minumum values) greater than 5 and markers 426, 459a and 454 have variability less than or equal to 2. The average variability is 4.56 .

If we assume that 0.2%=1/500 mutation probability per marker per generation applies to the average variability of 4.56, we get a factor of 0.0110 (mutation probability per generation for 25 markers per variability) unit for a marker (25/500/4.56). Then we can assign the following probabilities per marker per generation for the markers according to the following chart:

Then we have the following mutation probabilities per generation for the 25 markers:

To use these numbers to calculate the generations separation of two families, invert the number and multiply it by the number of relative mutations for each marker. (Ignore markers that have no relative mutations.) Add these inverses to get twice the number of generations that separate the two families. (Twice because their common ancestor is the same number of generations from each of the two families.) To convert to years multiply by, say, 25 years per generation. Here is an example for how to use these probabilities to calculate the time to a common ancestor for my paternal and maternal families:

So my two families are separated by about 2000 years. (If one used a fixed mutation rate of 0.2%=1/500 for each of the 25 markers the result would be 2500 years [10*500/2].)

Of course, the important question in this calculation is whether the average mutation rate of 0.2% is correct. Some work has been done to measure mutation rates for several of the Y markers:

• Mutation rates for DYS19, DYS389-1, DYS389-2, DYS390, DYS391, DYS392, DYS393 and DYS385.
• Mutation rates for DYS19, DYS389-1, DYS389-2, DYS390, DYS391 and DYS385.

The highest rate for these eight markers is for DYS390: 0.86% per generation. However, in the second paper, half of the four DYS390 mutations for 4999 father-son combinations were positive and half were negative. Only one other locus had a single negative mutation. Note in the variability bar chart above that DYS390 has average variability. DYS385a/b have twice the variability as does DYS390. I suspect that this is because of the tendency for DYS390 to mutate both negatively and positively. The variability described here comes about because of relative mutations between families that are generally quited distantly related, not father and son combinations. So the effective mutation rate for this case is smaller than the mutation rate for father-son situations.

Data References

Go to Y-Chromosome Markers Families Comparisons
Go to Roper male-line Y-chromosome project web page.

Return to top of page.
www.roperld.com