2013/6/26 <ngcart@charter.net>


I didn't imply that the complete genome should be pasted into Gramps;
that would be silly and serve no genealogical purpose.

A haplogroup is a few bytes long, e.g. I1a1 or R1b1 (male) and H6 or L3
(female).  If a male is I1a1, his son, brothers, father, and all male
predecessors in that line will have the same designation.  If his
paternal uncle's grandson tests with a different haplogroup, that IS
significant.

There is consistency within the labs as far as designating what
genealogically relevant data results:  just because one lab tests only Y
DNA and another tests mitichondrial, and another tests all the autosomal
chromosomes doesn't mean there isn't consistent reporting.  Your
chromosome 11 is the same length as mine, and the centromere is in the
same position as mine.  It has the same starting position and the same
ending position.  So comparisons between individuals can be made.

Here is an example of how a cousins DNA might compare to yours, turned
into a Python list from my CSV download:

['name redacted', '10', '51.7', '70.6', '18.9', '17.8']
['name redacted', '2', '12', '26.5', '14.5', '17.7']
['name redacted', '8', '12.6', '22', '9.4', '17.6']
['name redacted', '1', '116.8', '149.7', '32.9', '16.2']
Fisrst number is chromosome number; 2nd is starting point where person
matches you; 3rd is end point of match and 4th is length of match.  all
of these are times 1 million. (# 5 is a different representation of #4).

Here is an example of how one person matches several others in his
database across chromosomes.  The darker the area , the more people he
matches at those specific chromosome ranges which suggest additional
research into their genealogical connection.  He did it in R and I am
currently trying to reproduce it in Python.

    http://imageshack.us/a/img838/8599/0bdy.png

I can think of all kinds of reasons for including you geneaalogically
relevant DNA results with your family tree information.  Thanks for
responses, and I hope they continue.

There has been discussion about this previously on the mailing list:
http://gramps.1791082.n4.nabble.com/Event-for-Y-DNA-data-td1812521.html
leading to wiki page:
http://www.gramps-project.org/wiki/index.php?title=Genetics
If you have experience, do update that wiki !

There was some more information discussed I can't find in the archives now.
In essense, one can now use person attributes to store DNA data.  Using events seems stupid, as DNA does not change.
We don't have predefined values though (like key social security number, ...), and ideally we have that so users use the same key, and we can write a report using the DNA data stored.

If I remember correct, for DNA data, you preferably have attributes which are a bit more complex than we have now. Our current attributes in Gramps are: (key, value)
So one could store (DYS19, 14) or (Y-DNA Hg, R1b1), but that would lead to a huge amount of extra predefined keys needed, so was never implemented, as that is cumbersome.

A better approach would be to use the value part to give more information in a predefined way that can be parsed by reports.

So, for example: (DNA-DYS,  19 - 14) and (DNA-Y, Hg - R1b1)

If somebody with DNA experience could work out what keys (DNA-DYS, ....) should be predefined in Gramps, and what format the value should optimally have, (19 - 14 or better 19; 14), then we can build support for that.  A specific optional DNA tab page in Gramps is then a possibility, which helps to enter DNA data, and stores the actual given information as an attribute.

Benny