He is just wrong, the twins in this case got the same "results" as far as the genotyping is concerned. The differences were in the analysis provided.
Is it a case of them adding too many decimal places to the percentages? Making statistics out of noise?