NRC I

1
NRC I

• Refers to the first report of the National
  Research Council on forensic DNA typing, which
  came out in 1992
• In 1992 only a few states had criminal offender
  databases, all with RFLP genetic profiles

2
Typical Investigation
3
RMP

• When we find a match in the typical investigation
  the value of the evidence can be measured in part
  by answering, What is the probability that a
  single person chosen at random from a reference
  population will have a genetic profile that
  matches the evidence profile
• This question is appropriate because only a
  single person is chosen for genetic testing.
• In addition the non-genetic evidence does not
  help predict the genetic profile, e.g. the are
  independent.

4
NRC I Database Search
5
Is the RMP appropriate?

• If more than 1 suspect is identified by the
  database search a modification of the RMP, (see
  NRC II) is required.
• Must believe that there is no linkage
  disequilibrium between loci 1-8 and loci 9-15.

6
NRC II arguments against NRC I

• We lose information by not using loci 1-8 in the
  final statistical analysis.

7
Available loci
8
Power of NRC I method

• Matching studies of 2 million pairs yield 14,
  7-locus matches but only 1, 8-locus match.
• If only 7 loci were used to confirm we might
  typically expect RMPs of about 1 in 1-2 billion
  (1017(7/13))
• If one considers lab error rates (1 in 10,000)
  there is no loss in power between a 7 and 13
  locus match.
• LR 10,000 (RMP 1 in 100 quadrillion)
• LR 9,999.99 (RMP 1 in 10 billion)

9
Example

• Shortly after the NRC I report the Minnesota BCA
  used the NRC I method for a cold hit case (RFLP,
  five markers).