Probability and Statistics of DNA Fingerprinting - PowerPoint PPT Presentation

1 / 11
About This Presentation
Title:

Probability and Statistics of DNA Fingerprinting

Description:

Probability and Statistics of DNA Fingerprinting (posterior odds) = (likelihood ratio) (prior odds) The strength of a piece of evidence includes: Its accuracy. – PowerPoint PPT presentation

Number of Views:37
Avg rating:3.0/5.0
Slides: 12
Provided by: Walk103
Category:

less

Transcript and Presenter's Notes

Title: Probability and Statistics of DNA Fingerprinting


1
Probability and Statistics of DNA Fingerprinting
2
  • (posterior odds) (likelihood ratio) (prior
    odds)
  • The strength of a piece of evidence includes
  • Its accuracy.
  • Its meaning.
  • DNA evidence merely tries to connect or
    disconnect a piece of evidence to a suspect. It
    makes NO assertions of guilt or innocence!!!

3
  • (Weir, slide 26)
  • Key questions
  • The event is that a piece of DNA evidence
    matches the DNA of the suspect
  • How likely is this evidence to have that type, if
    it comes from the suspect? (prosecution)
  • How likely is this evidence to have that type, if
    it comes from someone else? (defense)

4
  • (Weir, slide 27)

5
  • Assuming independence in the allele frequencies
  • Example
  • 18 alleles, each with a frequency of Pa 0.1.
  • L1 1 / (Pa 18)
  • L1 1018

6
  • (Weir, slide 35)
  • Allele frequency dependence due to evolution
  • Dealing with pairs of alleles.
  • Theta is the probability that two alleles, each
    from a different randomly selected person, are
    identical due to evolutionary means.

7
  • Assuming dependence in the allele frequencies due
    to evolution
  • Example
  • 18 alleles
  • Pa 0.1
  • Theta 0.04
  • P(aa) 0.0338462
  • The new Pa SqrtP(aa) 0.18397
  • L2 (L1) (5.582886 x 10-4)

8
  • Adjustments of allele frequencies due to sampling
    effects
  • Cant make a DNA profile of everyone!
  • Must estimate how inaccurate your DNA database
    might be.
  • Zc is in terms of standard deviations
  • N is the number of alleles at the specific locus
    in the database
  • (confidence interval for proportions for an
    infinite binomial population)

9
  • Adjustments in allele frequencies due to sampling
    limitations
  • Example
  • Pa 0.1
  • N 10000
  • Zc 3 (99.73 - Really greater, approx 99.86)
  • New Pa 0.103
  • L3 (L1)(0.5874)

10
  • Effects of human error
  • Example
  • N 10000 (number of cases)
  • PoliceError 0.02
  • LabError 0.001
  • DNATestError 1/(some L) 1/(50 x 106)
  • (1-PoliceError)(1-LabError)(1-DNATestError)
  • (1-PoliceError)(1-LabError) approx.
  • Human Error is overwhelming the deciding factor
    with the accuracy of DNA fingerprinting!

11
  • Conclusions
  • Human error is the overwhelming factor in the
    accuracy of DNA fingerprinting. (However, its
    involved in all forms of evidence.)
  • Allele frequencies are NOT independent of each
    other. However when evolutionary considerations
    and sampling considerations are taken into
    account, the accuracy of DNA fingerprinting is
    still beyond human experience.
  • A statement of the accuracy of some piece of DNA
    evidence, by itself, is NOT a statement of guilt
    of innocence!
Write a Comment
User Comments (0)
About PowerShow.com