GedMatch Match Parent Side Sorter

This is a tool I have been working on as a way to categorize matches to opposing parent sides.

GedMatch Side Sorting Tool 

High-Level Overview

This tool attempts to identify two possible parental sides by repeatedly evaluating whether overlapping matches do or do not share DNA with each other within the same regions of the subject's genome.

Matches that share DNA with each other are treated as evidence for the same parental side. Matches that do not share DNA with each other are treated as evidence for opposite parental sides.

By repeatedly applying this logic across many matches, the tool builds two proposed parental sides and identifies matches that produce contradictory evidence.

How the Process Works

1. Select a Starting Match

The process begins with an Anchor Match. The anchor is assigned to Side A and serves as the initial reference point for evaluating other matches.

If Auto-Select Best Anchor is enabled, the tool evaluates multiple potential anchors and automatically chooses the one that produces the largest number of successfully sided matches.

2. Define the Comparison Region

For the current seed match, the tool identifies the portions of the subject's DNA where the seed shares DNA with the subject.

These shared regions become the comparison areas used for that seed.

3. Evaluate Other Matches

For every other match, the tool asks two questions:

Does the match overlap the same portion of the subject's DNA?

If not, no conclusion is drawn.

If yes, the match becomes eligible for comparison.

Does the match also share DNA with the seed in that same region?

If yes:

• The two matches are treated as evidence for the same parental side.

• The match is assigned to the same side as the seed.

If no:

• The two matches are treated as evidence for opposite parental sides.

• The match is assigned to the opposite side of the seed.

Certain matches may be excluded from opposite-side inference. These matches can still provide positive evidence for the same side but cannot be used to infer the opposite side based solely on the absence of a shared match.

Examples include:

• Matches identified as 23andMe-derived kits.

• Matches below the Minimum Kit Overlap threshold.

• Any other matches configured to be excluded from opposite-side inference.

4. Repeat Using Newly Sided Matches

Once a match has been assigned to a side, it can become a new seed.

The tool repeats the same evaluation process using that match as the reference point.

Each newly sided match can potentially classify additional matches.

This process continues until all eligible seeds have been evaluated.

5. Detect Contradictions

Sometimes a match receives conflicting evidence.

For example:

• One seed indicates Side A.

• Another seed indicates Side B.

When this occurs, the match is marked as a conflict and removed from further propagation.

Conflicted matches are reported separately for review.

6. Produce Final Results

The final output contains:

Side A

Matches most consistently associated with one proposed parental side.

Side B

Matches most consistently associated with the opposite proposed parental side.

Conflicts

Matches that received contradictory evidence.

Unassigned

Matches for which insufficient evidence existed to make a reliable determination.