Application of DNA·VIEW to a paternity case | |
Toshiko Sawaguchi, Charles Brenner, Akiko Sawaguchi |
13 December 1997
(Poster from the Japanese DNA Polymorphism meeting, Nagasaki, 4 December 1997) |
Mr. A was not available for testing so his types must be inferred indirectly from those of his half-brother and half-sister (G and H), of his child E from another cousin (Mrs. D), and from the types of his other cousins F and B as shown.
The possible father-child relationship between Mr. A and B is
investigated with the
Symbolic Kinship Program. The Kinship program is also used to
analyse the possible trio relationship among Mr. A, Mrs. C, and B.
Maternity analysis | |
Maternity 99.9%
16 blood groups "fatherless" computation using just 2 people |
Paternity analysis (trio) | |
Paternity 0.6%
9 simple systems Kinship program, use types of several relatives |
To this end the Symbolic Kinship Program is used to make an analysis (details and exhibits) which makes use of all the typing data from the various relatives.
In summary, the combined paternity index across all computed loci is about 1/200. That is, the liklihood ratio favoring non-paternity is about 200.
Using the verbal predicates of Hummel,
the paternity of Mr. A is "extremely unlikely" if Mrs. C is the
mother.
Paternity analysis (motherless) | |
Paternity 4%
8 simple systems Kinship program, use types of several relatives |
The answer is that since the verbal predicates have an unjustified assumption of "equal priors" built into them, the reasoning is not air-tight. This point is discussed on the page "Man, woman, and child."
Therefore, just to be careful, a further analysis was done to
test the paternity of Mr. A, assuming that Mrs. C is not the mother.
In summary, the likelihood ratio came to 1/23, suggesting that
paternity is "very unlikely."
Conclusion |
Links: Forensic mathematics home page.