View Full Version : How far between us?

Basta

11-20-2018, 05:31 PM

I came upon Y-DNA result, we're the same 14/16, with 2 and 3 difference on the rest. Here are STRs:

DYS456

x

DYS389I

x

DYS390

x

DYS389II

x

DYS458

x

DYS19

-2

DYS385a/b

x

DYS393

x

DYS391

x

DYS439

x

DYS635

x

DYS392

x

YGATAH4

-3

DYS437

x

DYS438

x

DYS448

x

RobertCasey

11-21-2018, 05:10 AM

I came upon Y-DNA result, we're the same 14/16, with 2 and 3 difference on the rest. Here are STRs:

DYS456

x

DYS389I

x

DYS390

x

DYS389II

x

DYS458

x

DYS19

-2

DYS385a/b

x

DYS393

x

DYS391

x

DYS439

x

DYS635

x

DYS392

x

YGATAH4

-3

DYS437

x

DYS438

x

DYS448

x

On the average - 12 marker testers have 90 % average false positive rate. 25 markers is around 50 % false positives and 37 drops to 20 % false positives (the first useful level). So any comparison at 16 markers will have around a 80 % false positive rate. You really need more markers to determine relatedness. Also, having the marker values for 14 of 16 and then having two multi-step markers is another red flag. Also, 385a and 385b are two independent mutations, so you really have 17 markers. Your list also does not include most of the faster mutating markers, so genetic distances should be much less than normal.

Basta

11-21-2018, 09:15 AM

So these are all slow moving markers? I've found an Excel list of a thousand Y-DNA tests in various parts of Croatia, data from some research. This one was the best, 14 of 16, with few 13s. On FTDNA I have 31 matches, genetic distance 1, on 25 markers. If I'm right, all those are closer than this one?

Michał

11-21-2018, 09:45 AM

On the average - 12 marker testers have 90 % average false positive rate. 25 markers is around 50 % false positives and 37 drops to 20 % false positives (the first useful level). So any comparison at 16 markers will have around a 80 % false positive rate.

You need to specify what your definition of "false positive" is? With no such definition, your above comment won't be very helpful to Basta.

We all are "related" to each other, so the question is what your specific cut-off for "true relationship" is.

Also, the above numbers may differ very significantly depending on haplogroup or specific subclade.

Also, having the marker values for 14 of 16 and then having two multi-step markers is another red flag.

Yes, this looks suspicious. For Y-GATA-H4, this might be related to different formats used, so in such case a conversion is needed.

Volat

11-23-2018, 09:36 PM

You need to specify what your definition of "false positive" is? With no such definition, your above comment won't be very helpful to Basta.

In stastical hypothesis testing false positive is type I error i.e inference of existence of something that is not there.

Michał

11-23-2018, 10:10 PM

In stastical hypothesis testing false positive is type I error i.e inference of existence of something that is not there.

If so, then we need to know what the null hypothesis was in this case. For example, if the null hypothesis was that the two analyzed haplotypes belong to people who are not related to each other (so type I error would mean that we wrongly reject the null hypothesis, ie. we wrongly consider these two people to be related to each other), we still need to know how Robert defines relationship. In other words, we need to know how to distinguish people who are "related" from those who are "not related" to each other.

Powered by vBulletin® Version 4.2.5 Copyright © 2020 vBulletin Solutions Inc. All rights reserved.