As far as I remember the criteria for the panels was 1) even coverage of the entire genome and 2) sufficiently polymorphic. No specific traits or characteristics were taken into account. They idea was to have a panel that could measure selectively neutral variation. QTL or markers close to QTL would skew any measure of distance or similarity between populations.

To complement the answer of Herwin, I would guess the markers were also selected to be polymorphic on a large diversity of breeds, be not linked physically, and also show limited number of null alleles (which may by experience, still exist)...

Thank you Herwin, thank you Gregoire

Salem, Hossem

Je peut ajouté par apport a Mr Herwin et Mr Leroy que les microsatellites soit du type dinucléotide par exemple (AG)n et simple (non complexe = (AG)nTTCAT(TC)n)

OK, thank you doctor Semir