DYS ordering in Y projects

[Nagelfar]

I think there should be some way to arrange the order & relation of DYS values in the project member table. Right now, it is simply in the common FTDNA order (starting at 393 and going on to; 390, 19, 391, 385a etc...). Wouldn't it be better to have the slowest mutating markers first and go toward faster mutating markers as you go right to show a higher probability of relation?

[wolong]

DYS Order

The way that results table is presently displayed, it is pretty much useless to most members. The subgrouping feature with mutation colorization is a big improvement. I think numerical ordering of the columns is OK since administrators can use their spreadsheet applications to sort the columns in any desired order. But, I would like to see FTDNA display "Y-Results" with automatic subgrouping of the members based on genetic distance from haplotype mode. As administrators, we can do that now ourselves, but it involves time and effort and mistakes can be made. The more time a group administrator devotes to website maintenance, the less time there is available for recruiting. Bill Long - Group Administrator - Long Surname DNA Project

[Itzhak Epstein]

The way that results table is presently displayed, it is pretty much useless to most members. The subgrouping feature with mutation colorization is a big improvement. I think numerical ordering of the columns is OK since administrators can use their spreadsheet applications to sort the columns in any desired order. But, I would like to see FTDNA display "Y-Results" with automatic subgrouping of the members based on genetic distance from haplotype mode. As administrators, we can do that now ourselves, but it involves time and effort and mistakes can be made. The more time a group administrator devotes to website maintenance, the less time there is available for recruiting. Bill Long - Group Administrator - Long Surname DNA Project

Yes.

In an ideal environment, project administrators would be provided with useful and effective management and analysis tools. Some would be simple for the less computer-savvy and others would be more sophisticated. In such an environment, FTDNA would allow the administrators to choose more default settings for their projects' web sites. This, of course, requires paying greater attention and allocating more resources to computerization.

[jablair]

I would like to see FTDNA display "Y-Results" with automatic subgrouping of the members based on genetic distance from haplotype mode. As administrators, we can do that now ourselves, but it involves time and effort and mistakes can be made.
Bill Long - Group Administrator - Long Surname DNA Project

I'm sure this is something that could be programmed, but it leaves a lot of questions. What genetic distance would you want FTDNA to use to determine grouping? Arbitrary genetic distances to dertermine grouping can be misleading. Sometimes you have to look past the generic distance and consider the paper trail and/or rare values on specific markers.

And there is also the problem with people who match at 12 markers but fail to match at 25 or 37 markers. I've had this happen in my own Blair DNA Project.

It's not that hard for Project Adminstrators to group their own participants. I have over 120 participants. Each time I get new test results I simply go to the "member page" of the GAP and open the "genetic distance report" for that member. I can quickly see how close the new test results match those of other participants and decide what group, if any, the participant belongs in. In borderline cases I can review the paper trail or look for rare value matches and then make my decision.

[jablair]

I think there should be some way to arrange the order & relation of DYS values in the project member table. Right now, it is simply in the common FTDNA order (starting at 393 and going on to; 390, 19, 391, 385a etc...). Wouldn't it be better to have the slowest mutating markers first and go toward faster mutating markers as you go right to show a higher probability of relation?

The current order that FTDNA uses is based on the various panels they offer. Panel 1: markers 1-12, Panel 2: markers 13-25, Panel 3: markers 26-37, and panel 4: markers 38-67.

Different participants order different numbers of markers to be tested, but most order them by panel. If they order test panel 3 they first order panel 1 and 2. Up until recently you could only order panel 3 if you had already ordered panels 1 and 2.

The difference in mutation rates for most markers is hardly worth considering when it comes to figuring out relatedness. The fact that a mutation occurred on a "slow" marker rather that s "fast" marker is not going to help you find your MRCA.

[Nagelfar]

The difference in mutation rates for most markers is hardly worth considering when it comes to figuring out relatedness. The fact that a mutation occurred on a "slow" marker rather that s "fast" marker is not going to help you find your MRCA.

Projects like the I1a project however are based on finding & classifying haplotypes (since I1a is such a large haplogroup with so few in subclades). I was meaning this from that perspective.

[rdgriffith]

I'm sure this is something that could be programmed, but it leaves a lot of questions. What genetic distance would you want FTDNA to use to determine grouping? Arbitrary genetic distances to dertermine grouping can be misleading. Sometimes you have to look past the generic distance and consider the paper trail and/or rare values on specific markers.

And there is also the problem with people who match at 12 markers but fail to match at 25 or 37 markers. I've had this happen in my own Blair DNA Project.

It's not that hard for Project Adminstrators to group their own participants. I have over 120 participants. Each time I get new test results I simply go to the "member page" of the GAP and open the "genetic distance report" for that member. I can quickly see how close the new test results match those of other participants and decide what group, if any, the participant belongs in. In borderline cases I can review the paper trail or look for rare value matches and then make my decision.

I agree that arbitrary groupings can be misleading. On our Griffi(th)(n)(s)(ng) project site, I group only by major haplogroup (G, I, R, etc.) and sort by genetic distance (GD) from the modal values of each haplogroup group. In order to account for the differing number of markers, I compute genetic distance from modal only for the number of markers in the target result...then divide the result by the number of markers tested. This provides a sort value that gives greater weight to the higher resolution tests. The only thing I have found to be consistently valid is that participants closer to the model values are more likely to find connections within the group than those more distant from the modal.

If participants have proven connections, their respective group numbers take precedence over the GD sort. I took this approach because we have so few proven connections among our participants. By adding an additional feature that computes genetic distance for a selected kit number, a participant can compare his results against the rest of the group. Genealogically-significant matches are displayed with contact information for those participants who sign release forms.