How can there be two answers?

  Science is being asked to answer the emotionally charged question
whether Tsilla Levine is really the daughter of Margalit Umassi.
However, it has, so far at least, fallen short.  To date, there
are two contradictory results from two different laboratories.
How can this be?  Is science not supposed to provide us with
clear, distinct answers?

  Last week, heredity jargon such as mitochondrial DNA and
genomic analysis was being casually bandied about in the media.
But what do they mean? How do molecular biologists determine that
two people are related to one another?

  The prime player on this forensic science stage is the very
stuff of heredity - DNA.  The hereditary material found in the
nucleus of every human cell is called genomic DNA, which contains
a complete set of chromosomes along with their associated genes.
Genomic DNA is a large molecule resembling a long chain of myriad
links.  Each link is made of nucleotides comprising three
elements, two of which are identical along the entire length of
the chain.  The third element is a chemical material called the
organic base, which appears in any one of four different
combinations.  Continuous segments of these links - which are
created by combinations of the four organic bases - are the
genetic codes that determine heredity. DNA consists of two linked
molecular strands - Watson and Crick's famous double helix model.
During conception, a child receives one strand of DNA from his
mother and one from his father.  However, DNA is also found in
mitochondria - the tiny powerhouses within each cell, which
contain enzymes for cell metabolism. Mitochondrial DNA is
composed of the same building blocks as genomic DNA, but are
different in that they pass directly from mother to offspring.
Fathers do not contribute mitochondrial DNA to their children.

  Scientific examinations intended to confirm or rule out family
relationship are based upon the degree of compatibility between
the DNA of the parents and the children. Forensic medicine
recognizes such comparisons as binding legal documents. In the
case of Margalit Umassi and Tsilla Levine, two different tests
were used to determine a familial relationship. Dr. Hasan Khatib
of the Hebrew University Institute of Life Sciences conducted one
set of DNA tests in which he compared genomic DNA found in blood
samples from the two women. Khatib correlated the nucleotide
strands found in the two samples, specifically comparing 15 sites
situated along the strands. Khatib was looking for the same
continuous sequence of four organic bases arranged along the
strand. Feeding the results of his comparison into the computer,
Dr. Khatib found a 99.9 percent probability that Margalit was in
fact Tsilla's mother.

  Meanwhile, Dr. Maya Freund at the Institute of Forensic
Medicine, took a slightly different tack, and examined
mitochondrial DNA, as well. Freund asserts that this is the
standard procedure used by forensic biologists when they are
asked to determine the relation between a mother and her
offspring. Since mitochondrial DNA is passed directly from mother
to child, without intervention or splicing with the father's DNA,
Freund claims that this is the preferred method for mother-child
forensic comparison. The researchers do not have any samples of
genetic material from the father, a fact which Freund says
complicates genomic DNA comparison, since half of it comes from
the father.

  Dr. Freund analyzed both mitochondrial and genomic DNA, testing
pre-set locations along the strands that are accepted as comparison
benchmarks by forensic scientists worldwide.   In both cases, the
samples were found to be statistically dissimilar.

  Geneticist Adam Friedman of the Hadassah-Hebrew University
Hospital feels that the possibility of two people's DNA matching
at 15 different sites along the strand, as was found in Khatib's
test, provides convincing statistical proof of a family
relationship.  Friedman also pointed out that testing of
mitochondrial DNA presents its own problems because, unlike
genomic DNA, it can mutate or fractionalize with age. This in
itself might explain why specimens from two related individuals
might not match

  (c) copyright 1998 Ha'aretz.  All Rights Reserved