other newsletters

October 30, 1998



A new feature in DNA·VIEW ties Kinship with Paternity Case. The result is very fast analysis of kinship situations because you don't have to make a separate manual calculation for each locus.

Immigration is a sub-menu option of Paternity case, and it is an alternative to Calculate report. It becomes available after a case has been defined, and data has been created for the people in the case. Then, Select case, and choose Immigration. A window opens wherein you can type in the kinship formula (i.e. the specification of pedigrees). For a standard paternity trio analysis, you would type

C : M + F/?

The standard "role" letters -- M, F, C, D, etc -- are used to name people.

You don't type in the genotypes; the program will add the genotype specifications iteratively, locus-by-locus, as it makes the calculation across all loci for which there is data.

The Immigration feature is so named because one typical use is analysis of claims of relationship that potential immigrants make to bolster their applicant status, and because I am not very inventive at creating names. However, it is of course also useful for every kinship situation, including inheritance, missing persons, etc.

This feature is experimental, in development starting with version 21.03 of DNA·VIEW.

SwissAir crash victims

Last month I was privileged to be invited by Dr. Fourney of the RCMP to help sort out DNA identifications of the SwissAir flight 111 crash victims. The questions to be answered were of the form: Given a set of related-looking DNA types from the crash (possibly a family that perished), and given a collection of living relatives as plausible referents, how certain can be we as to the identities of the various victims? Since there was typically little in the way of physical clues from the crash samples, many different combinations could potentially be considered; a large number of kinship scenarios needed to be calculated.

The project proved to be an excellent test of the new program. Since each configuration that I wished to test could be evaluated in only a minute or so, it was feasible to test a large number of possibilities -- although not possible to compare individually every possibility. However, I eventually came to understand a way to arrange all the conceivable relationships among the people in a hierarchical fashion, using a type of arrangement called a "lattice." The lattice represents a set of assumptions about which permutations of individuals are more plausible than others. At most a few comparisons are necessary to confirm that the lattice's structure is correct, and to compare the best explanation for the DNA evidence with a handful of possible second-best explanations. The smallest likelihood ratio among this handful -- usually it was over 1 million -- represents the superiority of the best explanation over any other explanation, a conclusion that is established confidently without requiring the tedium of examining most of the possible explanations.


Validation & Accreditation

A number of users have expressed interest in these matters, particularly Peter van Eede of the Amsterdam Red Cross who proposes to take a lead role. He has prepared a questionnaire, which I am distributing to all DNA·VIEW and PATER users. Please return it -- preferably directly to Amsterdam, although I will forward it to him if you send it to me early enough.

Followup – September 1999

The questionnaire results have now been posted on the web.

Also posted are validation/accreditation studies from the Amsterdam Red Cross lab.


The PATER report now (version 11.02) formats STR data that is imported from DNA·VIEW as one line per locus -- unlike RFLP data, but similar to the way it would look if treated as a PATER genetic system. The new format is illustrated below. Note that PCR types appear in a section of their own. You can change the wording of the title of the section by using the Serology tool (f1 f2) Categories of tests, 9]. (Category 10 is used for RFLP results. PCR genetic systems defined in PATER can be assigned to category 8.)

                                     Tested    Paternity   Nonpaternity    Ratio
                 Mother    Child     Man       Assumed (X) Assumed (Y)     (X/Y)
PCR alleles  
    FGA/4q                 5,6       6,5        0.5         0.00256       195      
    TH01/11p15.5 8,10      9,10      9          0.5         0.0871          5.74
DNA Probes (Kbase)
    TBQ7/D10S28   1.94      1.12      1.12      0.25        0.0378          6.62   
        Hae III   1.12      1.94      1.55
    D17S26/EFD52  4.28      5.21      5.21      0.25        0.0192         13      
        Hae III   7.03      7.03      2.55
PATER report. Imported PCR alleles (category 9) formatted one line/locus

More lines/report page

Assuming you use a Laserjet-compatible printer, the PATER command Printer Setup, option printer setup, includes a parameter of vertical lines/inch. By increasing the value of this parameter (beyond the recommended "proportional" value), you can squeeze more genetic systems onto a one page report. (version 9.71)