Autosomal DNA segment matching

Here is a quick and dirty picture of how I'm doing autosomal segment mapping. Forgive the tools. I'm using 23 and me because the bulk of my easily gathered data is there. I'm also using Gedmatch for an easy feature. I'm trying to get across a concept on this one.

Three people here. T, C and P.

C is the person I'm helping. They are the half sister of T. We start by figuring out our known people.

My end goal is to identify my C's unknown paternal family. Her closest maternal relative is T. I want to use T to sort matches.

This is basically an "In Common With" or "One or Both Kits" match. The question this answers is "Who matches both these kits". I'm checking T's half sister against him to determine their maternal side. This is the first question to ask about any match. Where do they fit and who do they match with.

P is going to be our unknown Person here. P is most likely a very good maternal match for C. Because C is my T's half sister, I am using their relationship as the known maternal relationship for C.

The next step is to see how P matches each person.

For that I'm switching back over to 23 and me. This is for a visual representation. You can use any tool for this. A spreadsheet would do something similar, but I want a visual representation. You could also check these kits individually at gedmatch to see where the segments were and then compare them against each other.

I'm going for simplicity here to illustrate a concept.

The key to this visual is that I can check P against both T and C. Yes I know they match each other but I want to know "how". With this tool (or a one to one comparison) I can see the match from P's perspective as well as C's. How does P match each person in this triangle?

I can visualize P vs C and P vs T.  T is in green and C is in blue. You can see that P matches both my T and C in many positions on many segments. There are lots of overlaps. Since T and C's relationship is known, those are most likely maternal positions.

P also matches C by herself sometimes though. You can see that where the little blue segments stand alone on chromosome 17 and 22.

Now for the counter check. I need to see if my C matches T in the same spots that she matches with P.

You can see here that C matches T on many segments. T is green and P is blue. For a lot of those segments P's match overlaps and matches the segment shared with T. Now I can look at those segments and assign them to C's and T's shared maternal family. The same segment matches all three people.

What about those little blue segments that don't overlap and match with T? Those are anyone's guess.

• They could be great leads to my C's paternal family. P may be related to C in more than one way. If I can identify P's match on those segments as paternal then P's family tree might be a good resource to finding C's unknown paternal family.
• They could be maternal matches that T just didn't inherit and outline a longer maternal segment for C. In that case it would be good to solidly identify them as maternal (using other maternal relatives).
• If they're small enough, they might be noise.

What if T had overlapped on those little blue segments but not matched?

It's most likely then, if a known maternal relative like T overlapped but did not match, that I have identified segments that show P's relationship to C's paternal family. P's family tree could then be a resource to find C's paternal unknown. 

As it stands, those segments warrant further investigation to see if I can get them assigned to a side of C's family. P is a "mostly" maternal match for C, with a few unassigned segments to be sorted out.