Practical determination of the low template DNA threshold

R. Puch-Solis, A. J. Kirkham, P. Gill, J. Read, S. Watson, D. Drew

Research output: Contribution to journalArticle

15 Citations (Scopus)

Abstract

The low template stochastic DNA threshold is used to infer the genotype of a single STR allelic peak. For example, within the context of the UK National DNA Database, the stochastic threshold is used to decide whether a DNA profile, consisting of a peak in position of allele a, is uploaded as aF or as an aa homozygote. The F designation acts as a ‘fail-safe’ wild card that is designed to capture the possibility of allele drop-out and to do this it will match any allele. If a profile is wrongly designated as an aa homozygote, the database search will be unnecessarily restricted and may fail to match a perpetrator reference sample on the database. If the stochastic threshold is too high, then this increases the number of adventitious matches, which in turn compromises the utility of the national DNA database. There are many different methods used to process DNA profiles. Often, the same stochastic threshold is used for each process (typically 150 rfu). But this means that more sensitive methods will have a threshold that is too low (and vice versa) and the risks of a wrongful designation are correspondingly greater. Recently, it was suggested that logistic regression could be used to relate the stochastic threshold to a defined probability of drop-out in order to properly evaluate the risks associated with a given stochastic threshold. In this article we introduce a new methodology to calculate the stochastic threshold that a practitioner could easily implement. The threshold depends on the sensitivity of the method employed, and is adjusted to be equivalent across all methods used to analyse DNA profiles. This ensures that risks associated with misdesignation are equivalent across all methods. In effect a uniformity of methods, underpinned by an analysis of risks associated with misdesignation can be achieved.
LanguageEnglish
Pages422-427
Number of pages6
JournalForensic Science International: Genetics
Volume5
Issue number5
DOIs
Publication statusPublished - Nov 2011

Fingerprint

DNA
Alleles
Nucleic Acid Databases
Homozygote
Databases
Logistic Models
Genotype

Keywords

  • homozygote threshold
  • low template DNA threshold
  • multiplex system
  • allelic drop out

Cite this

Puch-Solis, R., Kirkham, A. J., Gill, P., Read, J., Watson, S., & Drew, D. (2011). Practical determination of the low template DNA threshold. Forensic Science International: Genetics , 5(5), 422-427. https://doi.org/10.1016/j.fsigen.2010.09.001
Puch-Solis, R. ; Kirkham, A. J. ; Gill, P. ; Read, J. ; Watson, S. ; Drew, D. / Practical determination of the low template DNA threshold. In: Forensic Science International: Genetics . 2011 ; Vol. 5, No. 5. pp. 422-427.
@article{0d8da952552e48d0a47345b33d8347fb,
title = "Practical determination of the low template DNA threshold",
abstract = "The low template stochastic DNA threshold is used to infer the genotype of a single STR allelic peak. For example, within the context of the UK National DNA Database, the stochastic threshold is used to decide whether a DNA profile, consisting of a peak in position of allele a, is uploaded as aF or as an aa homozygote. The F designation acts as a ‘fail-safe’ wild card that is designed to capture the possibility of allele drop-out and to do this it will match any allele. If a profile is wrongly designated as an aa homozygote, the database search will be unnecessarily restricted and may fail to match a perpetrator reference sample on the database. If the stochastic threshold is too high, then this increases the number of adventitious matches, which in turn compromises the utility of the national DNA database. There are many different methods used to process DNA profiles. Often, the same stochastic threshold is used for each process (typically 150 rfu). But this means that more sensitive methods will have a threshold that is too low (and vice versa) and the risks of a wrongful designation are correspondingly greater. Recently, it was suggested that logistic regression could be used to relate the stochastic threshold to a defined probability of drop-out in order to properly evaluate the risks associated with a given stochastic threshold. In this article we introduce a new methodology to calculate the stochastic threshold that a practitioner could easily implement. The threshold depends on the sensitivity of the method employed, and is adjusted to be equivalent across all methods used to analyse DNA profiles. This ensures that risks associated with misdesignation are equivalent across all methods. In effect a uniformity of methods, underpinned by an analysis of risks associated with misdesignation can be achieved.",
keywords = "homozygote threshold, low template DNA threshold, multiplex system, allelic drop out",
author = "R. Puch-Solis and Kirkham, {A. J.} and P. Gill and J. Read and S. Watson and D. Drew",
year = "2011",
month = "11",
doi = "10.1016/j.fsigen.2010.09.001",
language = "English",
volume = "5",
pages = "422--427",
journal = "Forensic Science International: Genetics",
issn = "1872-4973",
number = "5",

}

Puch-Solis, R, Kirkham, AJ, Gill, P, Read, J, Watson, S & Drew, D 2011, 'Practical determination of the low template DNA threshold' Forensic Science International: Genetics , vol. 5, no. 5, pp. 422-427. https://doi.org/10.1016/j.fsigen.2010.09.001

Practical determination of the low template DNA threshold. / Puch-Solis, R.; Kirkham, A. J.; Gill, P.; Read, J.; Watson, S.; Drew, D.

In: Forensic Science International: Genetics , Vol. 5, No. 5, 11.2011, p. 422-427.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Practical determination of the low template DNA threshold

AU - Puch-Solis, R.

AU - Kirkham, A. J.

AU - Gill, P.

AU - Read, J.

AU - Watson, S.

AU - Drew, D.

PY - 2011/11

Y1 - 2011/11

N2 - The low template stochastic DNA threshold is used to infer the genotype of a single STR allelic peak. For example, within the context of the UK National DNA Database, the stochastic threshold is used to decide whether a DNA profile, consisting of a peak in position of allele a, is uploaded as aF or as an aa homozygote. The F designation acts as a ‘fail-safe’ wild card that is designed to capture the possibility of allele drop-out and to do this it will match any allele. If a profile is wrongly designated as an aa homozygote, the database search will be unnecessarily restricted and may fail to match a perpetrator reference sample on the database. If the stochastic threshold is too high, then this increases the number of adventitious matches, which in turn compromises the utility of the national DNA database. There are many different methods used to process DNA profiles. Often, the same stochastic threshold is used for each process (typically 150 rfu). But this means that more sensitive methods will have a threshold that is too low (and vice versa) and the risks of a wrongful designation are correspondingly greater. Recently, it was suggested that logistic regression could be used to relate the stochastic threshold to a defined probability of drop-out in order to properly evaluate the risks associated with a given stochastic threshold. In this article we introduce a new methodology to calculate the stochastic threshold that a practitioner could easily implement. The threshold depends on the sensitivity of the method employed, and is adjusted to be equivalent across all methods used to analyse DNA profiles. This ensures that risks associated with misdesignation are equivalent across all methods. In effect a uniformity of methods, underpinned by an analysis of risks associated with misdesignation can be achieved.

AB - The low template stochastic DNA threshold is used to infer the genotype of a single STR allelic peak. For example, within the context of the UK National DNA Database, the stochastic threshold is used to decide whether a DNA profile, consisting of a peak in position of allele a, is uploaded as aF or as an aa homozygote. The F designation acts as a ‘fail-safe’ wild card that is designed to capture the possibility of allele drop-out and to do this it will match any allele. If a profile is wrongly designated as an aa homozygote, the database search will be unnecessarily restricted and may fail to match a perpetrator reference sample on the database. If the stochastic threshold is too high, then this increases the number of adventitious matches, which in turn compromises the utility of the national DNA database. There are many different methods used to process DNA profiles. Often, the same stochastic threshold is used for each process (typically 150 rfu). But this means that more sensitive methods will have a threshold that is too low (and vice versa) and the risks of a wrongful designation are correspondingly greater. Recently, it was suggested that logistic regression could be used to relate the stochastic threshold to a defined probability of drop-out in order to properly evaluate the risks associated with a given stochastic threshold. In this article we introduce a new methodology to calculate the stochastic threshold that a practitioner could easily implement. The threshold depends on the sensitivity of the method employed, and is adjusted to be equivalent across all methods used to analyse DNA profiles. This ensures that risks associated with misdesignation are equivalent across all methods. In effect a uniformity of methods, underpinned by an analysis of risks associated with misdesignation can be achieved.

KW - homozygote threshold

KW - low template DNA threshold

KW - multiplex system

KW - allelic drop out

UR - http://www.scopus.com/inward/record.url?scp=80052271658&partnerID=8YFLogxK

U2 - 10.1016/j.fsigen.2010.09.001

DO - 10.1016/j.fsigen.2010.09.001

M3 - Article

VL - 5

SP - 422

EP - 427

JO - Forensic Science International: Genetics

T2 - Forensic Science International: Genetics

JF - Forensic Science International: Genetics

SN - 1872-4973

IS - 5

ER -