Announcement

Collapse
No announcement yet.

Weekly Information Technology/Engineering Update (10 Dec 2013)

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Rebekah Canada
    replied
    Originally posted by fullerbb View Post
    I am administrator for 3 kits studying one surname. We all descend from a common Grandfather & Grandmother. Kit 1- matrix does not show ANY relationships to 4 known cousins that match on FF, & one is a 1st cousin! Kits 2 & 3 -matrix show relationships to 2 of 3 known cousins & FF matches. So this tool does not work for me.
    Hi,

    Could you send me, [email protected], the kit numbers and relationships for each person? Jon and I will look into it.

    Leave a comment:


  • fullerbb
    replied
    Matrix doesn't work for me

    I am administrator for 3 kits studying one surname. We all descend from a common Grandfather & Grandmother. Kit 1- matrix does not show ANY relationships to 4 known cousins that match on FF, & one is a 1st cousin! Kits 2 & 3 -matrix show relationships to 2 of 3 known cousins & FF matches. So this tool does not work for me.

    Leave a comment:


  • royfarnol
    replied
    Arrange for the matches list not to default to the first name after selection of lower listed names but to remain on last selected name to obviate the need for rescrolling. I have 387 matches to scroll through and some folk have many more.

    Leave a comment:


  • MtnMama
    replied
    Matrix question

    I am using the Matrix to evaluate apparent overlapping segments. I downloaded all matches from the Chromosome Browser (GREAT new feature!!!) in a .csv file, saved as a spreadsheet file, and sorted by Chromosome and then longest cM.

    Then I removed everything under 7 cMs and color-coded by apparent matching segments. I have then used the Matrix to evaluate the matches.

    I have a long match (for me) of 31 cMs (who hasn't responded) on FTDNA. I will call him VB. I have two possible cousins who overlap the first half of his string, and with each other. I have another possible cousin who overlaps on the second half of his string.

    NONE of these people matrix with each other??? No little blue check marks at all. They all appear to overlap with VB on the chromosome browser but not on the Matrix.

    I have another match with no contact information with whom I have four separate fairly long matching strings, on three different chromosomes, ranging from 18, 13, 8 to 6.3. None of hers show on the Matrix with anyone else although some of the strings have people who overlap and matrix with each other.

    Help please. What am I not understanding?

    Leave a comment:


  • wighty44
    replied
    Yes, I will - thank you.

    Also, I just checked my son's FF matches. I, and his mother, show-up in his listing. Neither of his parents are identified as FF matches in the new matrix.

    Perhaps I'm doing something wrong it the selection process?

    PM being sent with kit No.s...

    Leave a comment:


  • Rebekah Canada
    replied
    Kit numbers

    Originally posted by wighty44 View Post
    I must be missing something as I'm not seeing the utility of this matrix.

    Firstly, the "Matches" column does not list all of my FF matches - my son (who is listed as a match in the standard FF match list) is not shown.

    Secondly, I do not understand why people who are listed as my FF matches are not identified as matches to me when they are added to the "selected Matches" column.

    What am I missing?
    Hi,

    Could you send me a private message with you and your son's kit numbers? I will have someone check into this right away.

    Leave a comment:


  • wighty44
    replied
    I must be missing something as I'm not seeing the utility of this matrix.

    Firstly, the "Matches" column does not list all of my FF matches - my son (who is listed as a match in the standard FF match list) is not shown.

    Secondly, I do not understand why people who are listed as my FF matches are not identified as matches to me when they are added to the "selected Matches" column.

    What am I missing?

    Leave a comment:


  • ech124
    replied
    One small request would be a clear button to remove "all". If I build matrix and then want to work on another grouping, I need to exit and re-enter to clear the slate.

    Leave a comment:


  • Táltos
    replied
    I finally just got around to playing with this new feature. I like it. It is easy to use. I too would like to see more of where we all exactly share or don't like at 23andme. Sometimes I can see where my Mom shares with say two matches over there. Then you can see if those two matches are sharing even more on another chromosome. Thanks FTDNA keep the new autosomal tools coming!

    Leave a comment:


  • Walter Freeman
    replied
    Originally posted by Zephyr View Post
    but no way to view to tell if they are matching on the same segment as the match to yourself like at 23andMe? ... I've found many a match who match one another but at different chr/seg locations so not true group matches indicating a common ancestor as the source of a shared DNA segment...
    I agree with Zephyr. This is a step in the right direction, but short of the chromosomal level of matching detail needed. Nice, but this is essentially the same matrix (sans detail) found on the GAP pages. The only difference is that one can now see this with respect to an individual account's POV.

    I use the GAP matrix from time to time, in projects that have encouraged atDNA participants. This form of matrix is useful, but not definitive for a drill down study.

    For instance, I have done a recent detailed study of a multiple triangulated match on Chr 4 which has now grown to 14 people, all of whom are related to the same ancestral couple at about the fourth cousin level. One of those in the study, however, matches me on another segment on another chromosome which 13 of the others do not.

    If I were using the matrix as currently configured at the beginning of this study, I would have extra people intermingled into the Chr 4 group, who were not legitimately connected to that group through the same common ancestors.

    Good try FTDNA but you are not there yet in terms of the detail of individual blocks on individual chromosomes we need to show that three (or more) people all match one another, i.e., A=B=C for a specific block.

    Leave a comment:


  • Oshri
    replied
    The matrix tool looks like a very usefull tool but i think it can be further improved. Here are my suggestions:
    1. Total shared cM metween matches could be great
    2. Allow threshold for the matrix
    3. Min cut of the Adjcency martix would be ausome
    4. Min cut of the Adjcency martix would be ausome
    5. Compare specific segments

    Leave a comment:


  • S9 H9
    replied
    Originally posted by T E Peterman View Post
    Regarding chromosome browser:

    Currently, I can select a match & see which segments are shared. This is apparently not considered a privacy issue.

    I can also select two people from my list of matches & see which segments are shared with the reference (me). Again, not apparently a privacy issue.

    I don't see why it would suddenly become a privacy issue if I got a report of identical segments shared by all three: the reference & each of the two matches. If this presents some sort of privacy problem, I think that this problem exists in the first two cases as well.

    Timothy Peterman
    The segment browser shows you where you are half identical with another person. From this you can conclude what the sharer's allele is at any location within that segment only if you are homozygous at that location.

    However, if you could exchange enough data with enough shares, and some of them gave you their raw file, you could write an algorithm that could estimate to a high confidence what all the data is for those who didn't give you their raw files.

    I think the concern over privacy is overblown, but I accept that FTDNA has to deal with it.

    Leave a comment:


  • T E Peterman
    replied
    Regarding chromosome browser:

    Currently, I can select a match & see which segments are shared. This is apparently not considered a privacy issue.

    I can also select two people from my list of matches & see which segments are shared with the reference (me). Again, not apparently a privacy issue.

    I don't see why it would suddenly become a privacy issue if I got a report of identical segments shared by all three: the reference & each of the two matches. If this presents some sort of privacy problem, I think that this problem exists in the first two cases as well.

    Timothy Peterman

    Leave a comment:


  • JohnG
    replied
    This is really a new feature but connected to the matrix. It would be very useful if two known relatives, who do not match by the algorithm in use, could submit something that said "Include us in each others matches". This would let us look at matches in common more easily.

    A current example is a 5th cousin who lives in another country. She does not match me or 3 other of her 5th cousins by the algorithm. But two of us match some of her matches, and there is pretty good evidence we are looking back 2-3 generations past our common ancestor, matching 7th or 8th cousins in that other country. Those matches have great paper genealogies and we have a fair chance of learning the connection. But every time we get new matches we have to go through a 5 person dance to exchange the new matches and look for duplicates.

    Leave a comment:


  • JohnG
    replied
    Originally posted by Rebekah Canada View Post
    Hello,

    That needs to be approached carefully so that individuals are respected but as many people as possible opt-in.
    Having the default be opt-in and providing an opt-out achieves that.

    Leave a comment:

Working...
X