For purposes of genetic genealogy it would be wonderful to have DNA >from
my great-grandparents up on FTDNA or Ancestry. As they died a long time
ago, this is seemingly a fantasy. However, it seems to me that
technology could be developed to obtain a good approximation of my
great-grandparents' autosomal DNA based on the autosomal DNA of their
surviving grandchildren. Specifically it seems to me if one were able
to get test results >from a number of their grandchildren, one could then
find which sections of DNA are common between grandchildren >from
different siblings. As the different grandchildren combinations likely
would have different areas of overlap, by combining the areas of overlap
across the sets of grandchildren, one could obtain a large number of
sections of overlap for autosomal DNA which would very likely be >from
their grandparents. Admittedly one would not know which autosomal DNA
section came >from which grandparent, but this could even be refined by
looking at DNA results of descendants of each grandparents' siblings.
In addition genetic relationships between in-laws and others in the tree
could be an issue, but with additional research this problem could be
minimized as well.

Are there any tools that do something like this yet? If not, is anybody
working on developing technology like this?

Seth Bittker

Researching:
BITTKER / BLAKHARZH, TOKARSKY / LEIBOVITCH - Aleksotas
REZNITSY, LINKOVSKY - Babtai
GROSSMAN, MIROSLAVSKI - Prienai
ROTH - Onod
HIPSCH, VEIS - Hernadszentandras

I think what you want is Kathy Johnston's Visual Phasing.

http://forums.familytreedna.com/showthread.php?t=36812

Blaine Bettinger discussed this at the GRIP course in July.

Israel Pickholtz