PDA

View Full Version : Y STR GD or TMRCA Matrix - Sorting Methods Question



haleaton
10-01-2015, 04:29 PM
When we have a Genetic Distance (GD) or TMRCA Matrix what are the methods of sorting the rows/columns to show clustering?

The Matrix is symmetric around the diagonal which can be any dummy value. It can be shifted up and down by permuting entries without changing order.

Changing order highlights clustering, often done manually by eye.

Is there a generic mathematical method for doing this?

MJost
10-01-2015, 05:15 PM
I like to create a haplotype GD list against a major haplogroup modal such as L21 and then sort the HT GD's smallest to largest and then enter these ordered HTs into McGee's adding the modal HT you used as the very first HT in the list. I also use MikeW's spreadsheet to do all the heavy work of running the GDs and sorting low to high and then pulling the HTs info into McGee's.

MJost

haleaton
10-01-2015, 07:28 PM
I like to create a haplotype GD list against a major haplogroup modal such as L21 and then sort the HT GD's smallest to largest and then enter these ordered HTs into McGee's adding the modal HT you used as the very first HT in the list. I also use MikeW's spreadsheet to do all the heavy work of running the GDs and sorting low to high and then pulling the HTs info into McGee's.

MJost

Thanks! I looked at this method but wanted to mix 37, 67, 111 data sets so used TMRCA or Normalized (by FTDNA Max Reported) GD. It creates a one sided distribution away from a modal comparison. When people do it by hand and aesthetics it is usually two sided with the lowest values in the middle--when there is only one cluster.

I thought there must be a general mathematical technique for organizing this type of data.

MJost
10-02-2015, 12:46 AM
Maybe some one can provide a solution.

MJost