I am working with unrelated people for local ancestry inference and I was wondering what are the metrics to assess phasing quality when having no familial data. I got affy6.0 array data for some 90 individuals phased with shapeit using polymorphic sites panel available in https://mathgen.stats.ox.ac.uk/impute/data_download_1000G_phase1_integrated_SHAPEIT2_9-12-13.html.

I've got suggestions to do an estimate of switch error, originally developed by Lin (2004, https://www.ncbi.nlm.nih.gov/pubmed/15256514) but if I got it right I would need to known parental original phases to do so. Any ideas?