The Use of GeneMapper ID-X Software v1.

1 for the Analysis of Mixed DNA Samples

Heather L. Harrah1*, MS; Valerie K. Bostwick1, MSFS; Sally E. Edwards1, BS; Anthony W. Eiler1, BS; Terry W. Fenger1, PhD; Rhonda K. Roby2, PhD, MPH
1Marshall University Forensic Science Center, Huntington, WV 25701; 2NIJ Technical Director, NEST Project, Fort Worth, TX 76107

ABSTRACT RESULTS SOFTWARE CALCULATIONS

GeneMapper ID-X Software v1.1 (Applied Biosystems, Foster City, California) mixture analysis tool assists the forensic analyst with the Mxobs (mixture proportion) values are calculated
Genotype
analysis and interpretation of mixed DNA samples. The mixture deconvolution module of the GeneMapper ID-X Software (GMID-X) below. The alleles and peak heights are from data
Figure 1A Locus Pattern Mxobs Calculations
evaluates two-person mixtures. GMID-X has the ability to produce best-fit major and minor profiles. The software also calculates the Figure 2A not presented for the 4 allele example. The three
(Major/Minor)
mixing ratio of two donors within a mixture. Using the defined Peak Height Ratio rule and Residual rule, a list of all possible candidate allele example relates to Figure 1B.
genotypes are ranked and displayed per locus. The genotype combinations are automatically sorted into either the Selected (most likely set 4 Alleles AB, CD
for inclusion) or Unselected Genotype Combinations tables. These tables can be edited to filter incorrect genotype combinations. In D8S1179 Figure 1B
(C + D ) Mxobs Calculation for a 4 allele locus:
Genotype Pattern: AB, CD (4 alleles)
addition, statistics such as Random Match Probability, Combined Probability of Inclusion/Exclusion, and Likelihood Ratio can be quickly ( A + B + C + D) Major: 28,29 Minor: 30,31.2
generated, ranging from one (1) contributor up to three (3) or more contributors.
A C B D Peaks:
3 Alleles AA, BC 28 (1432 RFU) 29 (1259 RFU)
Cases involving sexual assault, homicide, and touch DNA often have mixtures of two or more DNA profiles, e.g. vaginal swabs, fingernail
scrapings, blood stains, and swabs from door handles. The vaginal swabs and fingernail scrapings usually have the advantage of being No shared alleles (B + C) 30 (357 RFU) 31.2 (433 RFU)

considered intimate samples. Confidence in the resulting second contributor profile from an intimate sample is higher than if both
contributors are unknown. When both contributors are unknown, deconvoluting the mixture becomes much more complex. In this study,
( A+ B + C) Alleles: A=28 B=29 C=30 D=31.2

data will be analyzed with and without a reference applied, as GMID-X allows the user to define one contributor to the mixture, e.g., a
A B C
Equation: (C + D)
= Mxobs
victim’s profile. Additionally, low levels of input DNA from a mixture of two individuals will be explored. These low levels of DNA will 3 Alleles AB, CC ( A + B + C + D)
No shared alleles
C
exhibit stochastic effects in the amplifications resulting in allele dropout. GMID-X software considers the possibility of allele dropout by
reporting an F allele. The F allele is reported for potential alleles up to 1 RFU less than the mixture analysis threshold. ( A+ B + C) (357 + 433)
= Mxobs
C A B (1432 + 1259 + 357 + 433)
The results of these studies demonstrate that the mixture deconvolution module in the ID-X software can assist an analyst
GeneMapper 3 Alleles AB, AC
with mixture casework data. The peak height ratio calculations can be time-consuming when performed by hand and introduce the risk for
8
2546
13
2459
14
319 One shared Allele C 0.227 = Mxobs
human error or inconsistent calculations. An advantage to this new technology is that the peak height ratios for all combinations at each
locus are calculated automatically and consistently. In addition, mixing proportions for the two contributors are calculated per locus as
(B + C)
C A B
well as over the entire sample. As a result, the forensic analyst is freed to spend less time in performing mundane calculations and more Mxobs Calculation for a 3 allele locus:
time considering the various combinations produced by the given data. 2 Alleles AA, BB
No shared alleles
B Genotype Pattern: AB, AC (3 alleles, one shared)
Major: 13,8 Minor: 13,14
Figure 2B
( A + B)
D8S1179
Peaks:
INTRODUCTION
B A
8 (2546 RFU) 13 (2459 RFU) 14 (319 RFU)
2 Alleles AA, AB
One shared allele
N/A Alleles: A=13 B=8 C=14
In a DNA laboratory, forensic analysts are tasked with the job of interpreting mixture results. With advancements in software
interpretation tools for forensic DNA casework, rule firings and algorithms can be used to ensure consistency in interpretation from analyst 2 Alleles AB, BB C
to analyst. When working with samples at low levels with two contributors, mixture analysis becomes much more difficult and more One shared allele
N/A Equation:
[B + C]
= Mxobs
variability may be seen in analyst interpretations. By incorporating the GeneMapper® ID-X v1.1 (Applied Biosystems, Foster City,
California) software into their workflow, analysts can evaluate mixture data more consistently within a laboratory. 2 Alleles AB, AB
74 RFU
F-allele Two shared alleles N/A 319
= Mxobs
Figure 1: Results of a 10:1 (F:M) mixture of samples A and X at 1.50 ng total Figure 2: Results of the same 10:1 (F:M) mixture of samples A and X at an input
level of 0.25 ng total DNA. Due to stochastic effects, Figure 2A considers allelic
( 2546 + 319 )
Increasingly, expert systems and software tools have proved to be critical components during the analysis of both single source and DNA. The genotype combination possibilities for D8S1179 are represented in
mixture data. In mixture analysis, GeneMapper® ID-X (GMID-X) provides analysts an opportunity to examine all possible genotype mixture interpretation table (Fig. 1A). The Selected Genotype Combinations table dropout as indicated by the placeholders F1 and F2 in the minor genotypes. The F Table 3: Chart depicting scenarios, genotype patterns, and equations used to
allele is seen in these selected genotype combinations where it was not observed with 0.111 = Mxobs
combinations for major/minor contributors in two-person mixtures. The raw sample file (.fsa file) is directly imported into GMID-X where (top portion) shows phenotypes that have passed all rules. The Unselected Genotype determine the correct Mxobs calculation performed by GMID-X.
calculations are performed. The time savings benefits of the program include peak height ratio calculations and genotype ranking based Combinations table(bottom portion) shows phenotypes that have not passed the 8 13 higher input DNA (Fig. 1A). The Above Detection Below Interpretation (ADBI) rule
Inclusion Quality (IQ) rule. In the electropherogram (Fig. 1B) the mixture rule 349 471 has fired (Fig. 2A) indicating that a peak has fallen above the detection limit (50
on the residual (or other value) for each marker. The mixture analysis tool also allows for a reference to be applied, in which the software
(MIX) at the D8S1179 locus is flagged yellow, indicating a possible mixture. Both RFU) but below the interpretation limit (75 RFU). The correct major and minor

R =( Aexp − Aobs ) + ( Bexp − Bobs ) + ( Cexp − Cobs )

identifies the most likely genotypes for each contributor. Most importantly, GeneMapper® ID-X in no way replaces the analyst, it only 2 2 2 Figure 6: Residual calculation equation used to
major and minor contributors are represented in these data. For further detailed profiles are highlighted in the red box and is one of the combinations in the selected
assists in their interpretation of the data. rank genotype combinations. This equation uses a
explanations of these tables, refer to definitions of calculations below. genotypes table. In the electropherogram (Fig. 2B), the mixture rule (MIX) at the
least squares deconvolution method where A, B and
D8S1179 locus has fired and is flagged yellow indicating a possible mixture.
The NIJ Expert System Testbed (NEST) Implementation Team selected four unique profiles and created two sets of male/female mixtures. C represent Peak Proportions.
The information gathered from mixture analysis tools may prove to be a great advancement in forensic science technology.

Applying a Reference CONCLUSIONS

Mixture Proportion Calculations Residual and Peak Height Ratio Rules Inclusion Quality Rule
• Directly imports raw sample files (.fsa file) where calculations are performed.
MATERIALS & METHODS The observed mixture proportion (Mxobs) is a Genotype combinations are ranked by the residual value with To determine Inclusion Quality (IQ), the peak GeneMappper® ID-X allows the analyst to apply a reference • Contains mixture deconvolution tool for two-person mixtures.
measure of the relative proportion of a minor respect to the residual threshold. This threshold is software height ratios (PHR) for major and minor profile to a two-person mixture. The software will filter the • Can apply a reference sample to assist the forensic analyst with the analysis and interpretation of mixtures.
A total of 280 samples were prepared from two sets of two-person (male and female) mixtures. These profiles were chosen to demonstrate contributor for a genotype combination at a single defined at 0.04 and cannot be modified by the user. The contributors are considered with the Residual selected and unselected genotype tables to identify the most • Calculates and determines all possible genotype combinations and rankings according to the lowest residual value.
the effects of overlapping alleles, minor alleles which fall in the stutter position, and other important mixture combinations illustrated by locus in a two-person mixture. The Mxavg is lowest residual represents the least difference between the values. The IQ value is an indicator of the likely genotype combinations for each contributor. This • Designates an F-allele to consider allelic dropout.
Table 1. calculated using the Mxobs, and is the average expected and observed peak proportions. Samples below accumulated results of the applied rules. A sample comparison is made locus by locus using all available data. • Calculates peak Height Ratios, Mixing Proportions, Residual Value s, Peak Proportions, and Inclusion Quality quickly, consistently, and
mixing proportion for the minor contributor across 0.04 will pass the Residual rule and be assigned a passing will be added to the selected table if the IQ has a Genotype combinations can be moved to or from either the accurately.
all loci. This value is calculated from a combination green flag. Samples that fail this rule will be assigned a red green flag meaning both the Residual and PHR selected or unselected table at the analyst’s discretion. Also,
Table 1: Combinations of alleles in mixture samples • Offers flexibility to the analyst and has many options to customize.
of the lowest Mxobs values from each locus. Only flag. The peak height ratio (PHR) is then calculated and the passed. A yellow flag will be displayed for the IQ genotypes indicated as a match in the software can be marked • Includes the FBI population database and also has the option to create and import custom population databases for use in mixture
One Peak Two Peaks Three Peaks Four Peaks combinations that contain three or four alleles PHR status will be displayed as a green flag if the PHR at a if either the Residual or the PHR fail. If both the as inconclusive by the analyst. analysis.
(without dropout) are used to estimate an overall locus is greater than 50% or as a red flag if the PHR is less PHR and Residual fail then the IQ will be
1 Homozygote and 1 Homozygote and • Calculates Random Match Probability, Combined Probability of Inclusion/Exclusion and Likelihood Ratios.
2 Homozygotes
1 Heterozygote 1 Heterozygote
2 Heterozygotes average for the entire profile. than 50%. The Residual Status and PHR Status are then displayed as a red flag.
(shared allele) (no shared alleles) • Easy transition for analysts who currently use GeneMapper® ID (Applied Biosystems)
(one shared allele) (no shared alleles) combined to determine an overall Inclusion Quality (IQ).
2 Heterozygotes 2 Heterozygotes A sample will be added to the selected table if it
(two shared alleles) (one shared allele) Peak Proportion Calculations has a green IQ. If the IQ flags either yellow or DISCUSSION
2 Homozygotes Mxobs Mxavg PPexp PPobs Residual red, then the sample will be added to the Major Minor Known Genotype
(no shared alleles)
unselected table. The analyst has the option to
Selected GeneMapper® ID-X is a user-friendly software package that makes allele calls for single source profiles, has audit trails to track user edits,
Peak proportions (PP) are calculated after the Genotype
select or unselect genotype combinations in either Combinations and assists analysts in mixture interpretation. The NEST Project Implementation Team thoroughly studied the calculations used by GMID-
mixture proportions. A peak proportion is the Figure 3: GMID-X uses the observed Mx (Mxobs) for each
combination at each locus to calculate the average Mx (Mxavg) across table. X in the mixture analysis tool to better understand the software, appreciate its time-saving capabilities, and evaluate its consistency.
measure of the relative proportion of a particular
The mixture sets were mixed at the following ratios: 30:1, 10:1, 3:1, 1:1, 1:3, 1:10, and 1:30. The samples were then amplified at varying peak at a locus. The Mxavg is used to determine the entire profile. The Mxavg is then used to calculate the expected
Peak Proportion (PPexp) (Table 2). PPexp is compared to the observed This software considers all possible genotype combinations. The correct combination is not necessarily in the Selected Genotype
DNA input levels: 1.50 ng, 1.00 ng, 0.50 ng, 0.25 ng, with five different multiplexes: AmpFLSTR Identifiler®, Profiler Plus®, COfiler®, the expected peak proportion (PPexp) at a locus.
Peak Proportion (PPobs) to determine the Residual value (Fig. 6). The Combinations table nor is it always the combination that the software ranks as number one. Since the software performs time-consuming
SGM Plus® (Applied Biosystems, Foster City, California), and PowerPlex® 16 (Promega Corporation, Madison, Wisconsin). The samples The software also considers the observed peak
Residual gives the analyst an estimate of the closeness of fit between F-Allele Rule
Unselected complex calculations, the forensic analyst is freed to spend more time considering the genotype combinations produced by the given data.
were electrophoresed on an ABI Prism® 3100 Genetic Analyzer (Applied Biosystems) and analyzed with GeneMapper® ID-X version 1.1 proportions (PPobs) and compares these values to Genotype
expected and observed peak proportions. Combinations This tool is useful in helping an analyst deconvolute mixtures; however, the software results should be reviewed by an analyst to make the
(Applied Biosystems). the PPexp to calculate the Residual (R) value. This
The F allele considers the possibility of allelic final decision and take ownership of the allele calls.
value is a measure of the closeness of fit between
the PPexp and PPobs at a locus and is the result of dropout. It is given an RFU value determined by
= R / PHR using the mixture analysis threshold minus one.
the accumulated rules.
The GMID-X software calculates PHR and Mxobs REFERENCES
ACKNOWLEDGEMENTS
Residual PHR values using the F allele in order to determine
IQ = R / PHR 1) Clayton, TM, Whitaker JP, Sparkes RL, Gill P. Analysis and interpretation of mixed forensic stains using DNA STR profiling. Forensic Sci Int
The authors of this poster would like to acknowledge the staff of the National Institute of Justice’s Expert System Testbed Project Implementation PPexp PPexp PPexp Status Status which table (selected or unselected) that genotype
A B C or 1998; 91(1):55-70.
Project Team for their assistance with this project, specifically Chuck Heurich, Program Manager to the NEST Project. >0.04=R <50%=PHR combination belongs in. F alleles are represented Figure 5: Applying a Reference. In the Selected Genotype
R / PHR 2) Curran JM, Triggs CM, Buckleton JS, Weir BS. Interpreting DNA mixtures in structured populations. J Forensic Sci 1999; 44(5):987-995.
(1.0 − Mx ) + Mx
avg avg
(1.0 − Mx ) Mxavg <0.04=R >50%=PHR by F1 and F2, allowing for either homozygote or Combinations table, the bright green in the Major column matches
3) Evett IW, Gill PD, Lambert JA. Taking account of peak areas when interpreting mixed DNA profiles. J Forensic Sci 1998; 43(1):62-69.
Marshall University Forensic Science Center is a partner with the Forensic Technologies Center of Excellence (FTCoE) as the host site of the NEST 2.0 2.0
avg heterozygote dropout. the bright green Known Genotype column. The bright yellow in the
4) Wang T, Xue N., Wickenheiser R. Least-square deconvolution: A framework for interpreting short tandem repeat mixtures. J Forensic Sci.
Project. This project is supported by Award No. 2008-MU-MU-K003 awarded by the National Institute of Justice, Office of Justice Programs, U.S. = 0.5 2.0 2.0 = R / PHR Minor column indicates possible phenotypes for the unknown
2006;51(6):1284-1297
Department of Justice to the FTCoE administered by the National Forensic Science Technology Center. contributor. When a known genotype is applied, the GMID-X
5) Christen, A., Roby, R., et. al. The National Institute of Justice’s Expert System Testbed Project: Phase II, Analysis of Casework Samples. 2006.
software will move samples that have a green IQ flag but do not
Table 2: Example PPexp equations for three peaks at Proceedings of the 17th International Symposium on Human Identification.
The opinions, findings, and conclusions or recommendations expressed in this publication are those of the authors and do not necessarily reflect Figure 4: Rule Firings for IQ flags. The Residual status and the PHR match the known genotype to the Unselected Genotype
a locus with one shared allele (AB, AC). 6) Brooks, E., et. al. An Initial Evaluation by the NEST Project of GeneMapper ID-X’s Expert System Features. 2008. Mid-Atlantic Association of
those of the Department of Justice. status are combined to yield an IQ for a specific allele combination. Combinations table.
Forensic Scientists’ Annual Meeting.