Leeds AutoClustering Tool
AutoCluster Tool (Manual, ICW-Based)
Build the same kind of cluster grid you'd get from MyHeritage AutoCluster or Genetic Affair — but entirely manually and locally in your browser, based on which matches share each other's "in common with" list (ICW).
How this works
1. Add your matches with their name and shared cM. 2. Set a "too close" threshold — matches above it (parents, siblings, etc.) are excluded, since they share DNA with essentially every line and just add noise. 3. In the ICW grid, tick the boxes where two matches appear on each other's shared-match list at your testing company. This is exactly the kind of data real autocluster tools are built on — not segment overlap. 4. Click "Build cluster grid". The tool groups matches that share a large overlap of ICW connections into a cluster (the same underlying principle as the Leeds Method, but optimized and shown as a grid instead of a list).
On ICW vs. triangulation: ICW means two matches show up on each other's match lists — a strong indicator, but it does not guarantee you actually share the same DNA segment (that requires real segment overlap / triangulation in a chromosome browser). This is nonetheless exactly what AutoCluster and Genetic Affair themselves are built on, since most testing companies don't expose segment data through their APIs. If you've separately confirmed a match via triangulation, you can mark it as "Triangulated" below — it will then get a small marker in the grid so you can see which connections are extra solid.
1. Your matches
Add each match's name, total shared cM, and tick the box if you've separately confirmed segment overlap via triangulation.
| Name | Shared cM | Triangulated |
|---|
2. Settings
The similarity threshold controls how much ICW overlap two matches (or clusters) need to be merged. Higher → fewer, tighter clusters and more "unclustered" matches. Lower → larger, more intertwined clusters.
3. Shared-match grid (ICW)
Tick the box where the row match and column match appear on each other's shared-match list. Only matches at or below the threshold are shown, sorted largest to smallest.
Results — cluster matrix
The cells along the diagonal form "blocks" — filled cells within a block are the actual ICW connections that created the cluster. Small gray dots outside the blocks show ICW connections crossing two clusters (a possible pile-up match, or relatedness on both sides). Rename each cluster once you recognize surnames or known relatives in it.