Clustering of validation outliers based on 3D proximity?

[olibclarke]

Hi Paul,

In the multi criterion validation outlier markup, the entries are ordered based on their sequence position.

For large, multichain structures, I wonder if it would be worth considering using Kmeans or agglomerative clustering to identify 3D clusters of outliers - in order to help pick out "hotspots of badness" which may have residues that are close to one another in 3D, but not necessarily in sequence space?

Cheers
Oli