Just a thought:
Drawing data from a volunteer's mother and her siblings and/or father and his siblings will taint the test, i.e., make it unfairly difficult to discern the true identity of the mother and/or father. Siblings and their children share the same gene pool: Mom and Pop's sisters and brothers may very well share similar astro-points. Many families share the same signs and points like the Sun, Moon, ASC, Moon's nodes, etc., and planetary "complexes", i.e., clusters of aspects like T-squares, kites, trines and so on, handing down these 'signatures' through generations. These and other astro-connections are used to identify mother/child and father/child bonds.
For example: I have Taurus on the MC (the midheaven, the angle of the chart most often associated with the mother; the IC/MC is considered a 'parental axis'). In any case, the point is that my MC (mother-angle) is in Taurus and my mother had the Moon in Taurus. But, my maternal aunt was also a Sun sign Taurus. So, in my opinion, it would be easy for AT to mistake a maternal sibling for the volunteer's mother.
I would suggest using four birth data/charts drawn from four different months of the year. One data/chart is actually the volunteer's mother. Three others you could make up -- but these made up charts should not be drawn from the same astro-birth month (by astro-month I mean when the suns enters and leaves a particular sign) as the volunteer's mother. Same with the father. As well, the false charts should have birth times different than the parents' birth times. At least three hours difference. That's what I'd want.