rechtsmedizin2019·29:30–40 ... · dnaextractionandstrproﬁling the dna was isolated with the...

Original article

Rechtsmedizin 2019 · 29:30–40https://doi.org/10.1007/s00194-018-0291-1Published online: 18 December 2018© The Author(s) 2018

K. Anslinger · M. Graw · B. BayerInstitute of Legal Medicine, Ludwig-Maximilians-University Munich, Munich, Germany

Deconvolution of blood-bloodmixtures using DEPArrayTM

separated single cell STR profiling

Introduction

Mixture deconvolution

Mixture deconvolution is a main topicin forensic DNA investigations, becauseroutine cases involve a lot of mixturesand short tandem repeat (STR) profil-ing leads to mixed DNA profiles. Withthe exception of mixed profiles show-ing a major component or two-personmixtures, from whom one contributor isknown and the alleles from the secondcould be deduced, for most of these mix-tures the unambiguous determination ofthe alleles of a single contributor is notpossible. Therefore mixed profiles canmostly be used for direct comparison ofprofiles from known persons. A com-parison with a national database, whichis the aim of almost all investigations incases where no suspect could be deter-mined, is only possible for some selectednational databases [4]. In general, forsearching national databases, single per-son profiles are needed or at least prefer-able. Moreover, mixed profiles are morecomplicated to interpret and statisticallyevaluate than single source samples [12,17]. In summary, deducing single pro-files forall contributorsof agivenmixturewould be a great advantage.

For these reasons, several methodswere developed by forensic scientistsaiming to split mixed stains into theirdifferent cellular components. These in-clude the separationof different cell typesby laser capture microdissection (LCM)or fluorescence-activated cell sorting(FACS) [6, 18], as well as the time-shiftrelease of the DNA from the differentcellular components of a mixture, such

as the differential extraction method forsperm-epithelial cell mixtures [10]. Forthe splitting of mixed stains consistingexclusively of cells of the same cell type,some other methods may be helpful; forexample, the use of anti-human AB0and CD45 antibody-coated microbeadscombinedwith centrifugal separation forthe isolation of white blood cells origi-nating from donors with different bloodgroups [19]. As a further approach, thesex-specific labelling of cells combinedwith laser microdissection is used forthe isolation of cells from a male anda female contributor [1, 2]. Moreover,some probabilistic models included instatistical software programs were usedfor mixture deconvolution [5, 11].

DEPArrayTM technology

The DEPArrayTM technology (Menar-ini Silicon Biosystems, Bologna, Italy)has been successfully used for mixturedeconvolution by physical isolation ofcells using a digital approach. Withthe help of this technology, single cells,distinguished by immunofluorescent la-bels and verified by optical imaging, areisolated with the use of a computer-con-trolled semiconductor dielectrophoreticchip. Cells are inserted into a flowcell and captured in compartments (so-called cages) generated by the activationof hundreds of thousands of microelec-trodes distributed on the floor. Oncecaptured and identified, the cells canbe independently moved to a collectionvial with extreme precision. Thus, purepools of cells from different cell popu-lations can be separated from a mixtureand recovered in separate aliquots. On

the other hand, cells belonging to thesame cell population can be separatedand recovered individually for singlecell STR profiling. For forensic applica-tions, the DEPArrayTM forensic sampleprep kit (Menarini Silicon Biosystems),which enables the staining of epithe-lial cells, leucocytes and sperm cells,was specifically developed by Menar-ini Silicon Biosystems. The proof ofprinciple was demonstrated by Fontanaet al. [8]. Moreover, Williamson etal. used the DEPArrayTM technologyfor the enhanced mixture deconvolu-tion of sexual offence samples [21]. Ina former study this technology has beenused in the context of chimerism deter-mination after allogenic bone marrowand stem cell transplantation [3]. AllDEPArrayTM-related studies publishedso far were based on the separation ofpure pools of cells from different cellpopulations. This study attempted toconnect the separation of individualcells of the same type with single cellSTR profiling for the deconvolution ofmixtures containing white blood cellsfrom different contributors.

Single cell STR profiling

Single cell STR profiling is not com-monly used in forensic DNA investiga-tions, mostly due to the lack of reliablemethods for the isolation of single cellsfrom forensic samples. When single cellscan be obtained, some literature existsdemonstrating the possibility to obtainaprofileusingdifferentapproaches(Genget al. Single-cell forensic short tandemre-peat typing within microfluidic droplets[9]). A more commonly used procedure

30 Rechtsmedizin 1 · 2019

https://doi.org/10.1007/s00194-018-0291-1http://crossmark.crossref.org/dialog/?doi=10.1007/s00194-018-0291-1&domain=pdf

sampling point

Fig. 19 Bloodstainon the blade of aknife (sample II)

in order to obtain single source perpetra-tor DNA profiles from mixed samples, isthe collection of individual skin flakes orbio-particles [7, 15]: however single cells,as well as individual skin flakes, containa very small amount of DNA, which isnot sufficient to obtain STRprofiles usingthe routine extraction and amplificationprotocols. Therefore, special DNA ex-traction methods (e.g. single tube lysisand amplification) andmodified amplifi-cation protocols (low template, LT-DNAprotocols) were developed [8, 13]. Whileimproving the sensitivity through an en-hancement of PCR, LT-DNA protocolsoftenhavean influenceontheappearanceof the profiles; for example, increasedstutter peaks and imbalances of the twopeaks of a heterozygote loci as well asincreased drop-out rates were observed[20]. This in turnentails thedevelopmentof special guidelines for the interpreta-tion of these profiles.

Study design

In order to obtain single source perpetra-tor DNA profiles from blood-blood mix-tures of twoor three contributors, for thisstudy individualwhite blood cells (WBC)were separated using the DEPArrayTMtechnology. The DNA of each single cellwas set free in solution with the help ofa single tube lysis kit (DEPArrayTM Ly-sePrepKit, Menarini Silicon Biosystems)and STR profiling was carried out justadding the PCR mix to the lysed cell so-lution. To get a first impression regardingthe quality and reproducibility of singlecell profiles in optimal conditions, whiteblood cells from a fresh blood samplewere processed in advance (sample I).

Following this first control sample, threeadditional bloodstains (including mockand real casework samples) were investi-gated, which contained blood from twoor three contributors (samples II–IV).To check the correctness of the obtainedsingle source perpetrator DNA profilesand therefore to demonstrate the proofof principle, only mixtures from knowncontributors were investigated.

Material andmethods

Material

Sample Iwas a freshEDTAblood sample,from which 5μl were removed for inves-tigation. Sample II was a blood stainon the blade of a knife (. Fig. 1). Theknife was submitted in the context of theinvestigation in a current murder case.The age of the stain was approximately6 weeks. Routine STR profiling carriedout on several samples taken from theblade as well as the handle of the kniferevealed completely matching mixtures.Thesemixtures couldbe attributed to twopersons and completely explained by thealleles from the victim and suspect. Thecorresponding reference samples wereavailable. For DEPArrayTM separationanother sample was taken from the bladewith a nylon flocked, PBS buffer-soakedswab. Sample III was a blood-bloodmix-ture of two individuals (without a clearlyrecognizable major component) appliedon a piece of cellulose, which was part ofthe German DNA Profiling (GEDNAP)proficiency test of the year 2017 (GED-NAP55, stain2). For sample IVamixturewas made up of 5μl fresh EDTA bloodfrom three different individuals. From

this mixture 5μl were taken and addedtothestainingprocedure. All freshEDTAblood samples used here were submittedas reference samples for routine casesand therefore the corresponding profileswere known. In order to check the cor-rectness, the obtained single source per-petrator DNA profiles for the GEDNAPproficiency test stain (sample III) weresent to the organizer for confirmation.

Cell separation using DEPArrayTMtechnology

Cells were resuspended from the cel-lulose (sample III) or the swab (sam-ple II) according to the manufacturer’sinstructions. The resulting cell suspen-sions as well as the 5μl whole bloodsamples (sample I and sample IV) wereadded to the staining procedure. Whiteblood cells were stained using the Foren-sic Sample Prep Kit (Menarini SiliconBiosystems) according to the manufac-turer’s instructions. For the staining ofwhite blood cells, this kit contains a PE(Phycoerythrin)-conjugated anti-humanCD45 antibody. Furthermore, the nucleiwere stained with DAPI (4’,6-Diamidin-2-phenylindol). Thesampleswere loadedinto DEPArray cartridges and processedaccording to the manufacturer’s speci-fications for the DEPArray TM V2 sys-tem. White blood cells were classifiedwith the criteria “InCage” (position ofthe cell inside the dielectrophoretic cageso that it can be moved by the system),DAPI positive and PE positive using theCellBrowser software (Menarini SiliconBiosystems). The maximum number ofcells that can be recovered per experi-ment is determined by the type of chipused and by the type of recovery (singleor pooled). On the DEPArray TM V2 sys-tem, a maximum of 17 single cells canbe isolated, therefore, for samples II, IIIand IV, 17 single cells were recovered.For sample I a pool of 5WBCs (as a con-trol sample, which contains five times theamount of DNA and therefore sufficientDNA to yield a full STR profile) as wellas 11 single WBCs were isolated.

Rechtsmedizin 1 · 2019 31

DNA extraction and STR profiling

The DNA was isolated with theDEPArrayTM LysePrep Kit (MenariniSilicon Biosystems) according to manu-facturer’s instructions. Using the Mul-tiplex-PCR PowerPlex® ESXfast system(Promega, Madison, WI, USA) the sex-determining amelogenin system as well16 autosomal loci were amplified ona Veriti Thermal Cycler (Thermo FisherScientific, Waltham, MA, USA). ThePCR was carried out in a reaction vol-umeof 14μl and a 32 cycle PCRprogram,according to the in-house validated pro-tocol; apart from that, themanufacturer’sinstructions were followed. Determina-tion of fragment length was performedon a 3500xl Genetic Analyzer (ThermoFisher Scientific) according to manufac-turer’s instructions. Data analysis wascarriedoutusing theGeneMapper® ID-XSoftware v1.4 (Thermo Fisher Scientific)and a detection threshold of 50rfu.

Results and discussion

Sample I

For all 11 WBCs, profile completenessvaried from 18 to 31 of the expected32 alleles (. Table 1). This means thatthe worst profile still showed more than50% of the expected alleles, whereas onaverage approximately 82% of the allelescould be obtained per single cell. Whilea complete profile could not be obtainedfor any of the 11 single cells, it can bestated that each individual allele was de-tectable at least 6 times and 3 alleles wereshown in each of the 11 profiles. For thepool of five WBCs a balanced and fullprofile was obtained.

. Fig. 2 shows as an example of thepartial profile from one single cell (SC-6WBC) with one allelic drop-out in locusD16S539 (indicated with*). Beside thisdrop-out some other characteristics, typ-ically generated byLTprotocols, could beobserved. For example, some of the alle-les of the heterozygote loci show strongimbalances (indicated with arrows). Thepeakheightof the smallerallele is approx-imately 50% lower or more (50% allele31 in locus D21S11, 81% allele 13 in lo-cus D10S1248 and 64% allele 17 in locus

Abstract · Zusammenfassung

Rechtsmedizin 2019 · 29:30–40 https://doi.org/10.1007/s00194-018-0291-1© The Author(s) 2018

K. Anslinger · M. Graw · B. Bayer

Deconvolution of blood-bloodmixtures using DEPArrayTMseparated single cell STR profiling

AbstractBackground. So far STR (Short Tandem Re-peat) profiling of single cells has played hardlyany role in forensic molecular-biologicalcasework; however, the application of thistechnique could open up new possibilities fordeducing the unambiguous DNA profile ofeach single contributor involved in a mixtureof two or more components.Objective. For this purpose DEPArrayTM

technology was used to isolate single whiteblood cells from mixed stains containingblood from twoor three different contributors.The STR profiling was carried out on eachsingle cell.Material andmethods. In order to standar-dize the interpretation of the findings, which,like all low template (LT) DNA analyses, showscharacteristic artifacts, leukocytes from a freshblood sample were examined in advance and

the profiles were evaluated. Based on thesefindings, the results of the examination ofthree mixed stains (consisting of blood fromtwo or three persons) were evaluated.Results. In this way, complete or almostcomplete DNA profiles could be deduced forall persons involved in the correspondingmixture.Conclusion. The STR profiling of single cellsisolated by DEPArrayTM technology opensup completely new possibilities in the fieldof mixture deconvolution, especially formixtures composed of cells of the same celltype.

KeywordsDeducing single DNA profils · Mixed stains ·Interpretation guidelines · Artifacts · Lowtemplate (LT)

STR-Typisierungen von DEPArrayTM-separierten Einzelzellen zurAufsplittung von Blut-Blut-Mischungen in die jeweiligenEinzelbestandteile

ZusammenfassungHintergrund. STR(Short Tandem Repeat)-Analysen von Einzelzellen spielen bislangin der forensisch-molekularbiologischenFallarbeit kaum eine Rolle. Die Anwendungdieser Technik, im Speziellen zur Ableitungvon DNA-Profilen einzelner, an einerMischung beteiligten Personen, könntejedoch neue Möglichkeiten eröffnen.Ziel der Arbeit. Zu diesem Zweck sollenin dieser Arbeit mit Hilfe der DEPArrayTM-Technologie einzelne Leukozyten ausMischspuren, in denen sich Blut mehrererPersonen findet, separiert einer Einzelzell-STR-Analyse zugeführt werden.Material undMethode.Umdie Interpretationder Befunde zu standardisieren, die wiealle Low-template (LT)-DNA-Analysen cha-rakteristische Artefakte aufweisen, wurdenvorab Leukozyten einer frischen Blutprobeuntersucht und die Profile ausgewertet.Basierend auf diesen Erkenntnissen wurden

im zweiten Schritt die Ergebnisse derUntersuchung von 3 Mischspuren (bestehendaus Blut von 2 bzw. 3 Personen) bewertet.Ergebnisse. Für alle, an den jeweiligen Spurenbeteiligte Personen konnten vollständigebzw. fast vollständige DNA-Profile abgeleitetwerden.Schlussfolgerung. Durch die STR-Typisie-rung einzelner, mit Hilfe der DEPArrayTM-Technologie, separierter Zellen eröffnensich vollständig neue Möglichkeiten zurAbleitung von Einzelprofilen aus Mischungen,insbesondere wenn sich die entsprechendenMischspuren aus Zellen des gleichen Zelltypszusammensetzen.

SchlüsselwörterAbleiten von Einpersonenprofilen · Mischspu-ren · Interpretations-Richtlinien · Artefakte ·Low-template (LT)


https://doi.org/10.1007/s00194-018-0291-1

Table1

Profilingresults

ofthe11

whitebloodcells(W

BC)fromsampleIasw

ellasthe

alleleso

fthe

referenceprofile

DNA

system

D3S

1358

VWA

FIBR

ATH

O1

SE33

D8S

1179

D21

S11

D18

S51

Amel

D16

S539

D2S

1338

D19

S433

D22

S104

5D1S

1656

D10

S124

8D2S

441

D12

S391

Num

ber

ofde

-tected

true

alleles

Reference

Pool5

WBCs

15/17

15/17

247/8

1811/15

31/31.2

18/19

X/Y

9/12

17/19

13/14.2

14/15

14.3/15

13/14

10/12

20/24

32

SC-1WBC

15/17

1524

718

11/15

31/31.2

18/19

X/Y

9/12

1713/14.2

14/15

1513/14

10/12

20/24

28

SC-2WBC

15/17

15/17

248

1811/15

31/31.2

18/19

X9

17/19

1315

14.3/15

1410/12

20/24

26

SC-3WBC

15–

247/8

1815

31/31.2

18/19

Y–

–13

–14.3/15

13/14

–20/24

18

SC-4WBC

1715/17

247/8

1811/15

31/31.2

18/19

Y12

19–

14/15

14.3/15

13/14

1220/24

25

SC-5WBC

15/17

15/17

247/8

1811

31/31.2

18/19

Y9

1913/14.2

14/15

1513/14

10/12

20/24

27

SC-6WBC

15/17

15/17

247/8

1811/15

31/31.2

18/19

X/Y

917/19

13/14.2

14/15

14.3/15

13/14

10/12

20/24

31

SC-7WBC

15/17

15/17

247/8

1811/15

3118/19

X/Y

12–

13/14.2

1414.3/15

13/14

1220/24

26

SC-8WBC

15/17

14/15/17

–7/8

––

––

X/Y

–17/19

13/14.2

14/15

1513/14

10/12

20/24

21

SC-9WBC

15/17

15/17

248

1811/15

31/31.2

18/19

X/Y

9/12

17/19

13/14.2

14/15

14.3/15

13/14

10/12

20/24

31

SC-10

WBC

15/17

15/17

247/8

1811/15

31/31.2

18X/Y

9/12

17/19

13/14.2

14/15

14.3/15

13/14

1020/24

30

SC-11

WBC

15/17

1524

718

1131

18/19

X12

17/19

13/14.2

14/15

14.3/15

13/14

10/12

2025

Boldtype

indicatesalleleso

nlydetected

once


Original article

*

Fig. 28 Partial profile of single cell SC-6WBC (white blood cells) showing one allelic drop-out in locus D16S539 (indicatedwith *) aswell as strong heterozygote imbalances in theD21S11, D10S1248 andVWA loci (indicatedwith arrows) and someincreasedminus one repeat stutter >15% (circled)

VWA) in comparison with the second al-lele of the same locus. Themaximum im-balance obtained for the 2 alleles of a het-erozygote locus in all 11 profiles showeda ratio of 93.4:6.6% (data not shown).For the interpretation of all subsequentprofiles, smaller alleles of a heterozygotelocus, which did not reach a peak heightof at least 50% of the second allele of thesame locus, were not classified as truealleles.

Increased n-4/n-3 stutter was oftenobserved, as well. For example, in theprofile for SC-6 WBC (. Fig. 2, circled)minus one repeat stutters greater than15% could be found (stutter to allele 17in locus D3S1358 15.1%, stutter to al-lele 17 in locus VWA 29.3% and stut-ter to both alleles of D12S391 with 16.2and 20.7%, respectively). The highestpeak that occurred in n-4 stutter posi-tion showed a peak height ratio of 78.9%(SC-8WBC) and therefore was classifiedas allelic drop-in in advance. This peak

was not included in further considera-tions regarding the expression of stutterpeaks. The maximum stutter values ob-servedat all otherallelesweredeterminedper locus and are listed in . Table 2. Asa comparison, the maximum stutter val-ues reported by McLaren et al. in theirPowerPlex® ESXfast validation study areshown [14]. McLaren et al. calculatedthe stutter values (mean, standard devia-tion andmaximum stutter ratio) for eachlocus under consideration of the profilesfrom 656 individuals. All these profileswere generated using sufficient amountsof DNA and a 30-cycle PCR program. Incontrast, the number of profiles createdin this study as a part of the feasibilitystudy is still too small to provide reliableresults for a detailed statistical analysis.Forthisreason, thecalculationsonthede-termination of the maximum values perlocuswere initially limited, which shouldserve as a first idea of the interpretationof the profiles from the following sam-

ples. For routine use of this techniqueand standardized interpretation of theresulting profiles, a special set of appro-priate guidelines, based on a bigger dataset and amore detailed analysis, has to bedeveloped. McLarenet al. obtainedmax-imum stutter values up to 19.3%, while,in this single cell data set, extremely highstutter up to 34% was detected; for 7 ofthe 16 autosomal loci investigated indi-vidual stutter peaks values higher than20% were reached. Also, two additionalalleles in plus-one repeat stutter posi-tion were observed (allele 19 in D18S11,peak height ratio 16.4% and allele 12 inD16S539 with a ratio of 31.1%). Basedon this knowledge, for the interpretationof all profiles obtained for samples II, IIIand IV, peaks in plus-one and minus-one stutter position were only called astrue alleles when the peak height ratiowas above 35%. Further additional alle-les, which were not located in a minus or


Table 2 Maximumstuttervaluesper locusfor all alleles of the 11 profiles from sample Iin comparisonwith themaximumstuttervalues reported byMcLaren et al [14]

Locus Minus one repeat stutter (%)

Sample I McLaren et al.

D3S1358 19.2 19.3

TH01 4.5 5.9

D21S11 16.3 13.4

D18S51 33.8a(12.1) 13.8

D10S1248 15.0 14.5

D1S1656 28.2a(17.2) 19.3a(15.9)

D2S1338 16.7 17.1

D16S539 18.5 14.1

D22S1045 22.3 15.9

VWA 29.3 12.4

D8S1179 16.4 13.0

FGA 18.9 15.6

D2S1441 33.1a(14.8) 18.2a(11.3)

D12S391 27.1 18.7

D19S433 12.4 12.8

SE33 34.0a(18.4) 16.7a indicate that the given maximum stuttervalue was greatly increased to the rest of thedata set and the next highest value, whichwas additionally given in brackets

plus one repeat stutter position did notoccur.

Under consideration of possibly oc-curring increased stutter peaks and thefact thateachreal individualallelewasde-tectable at least 6 times in the 11 partialsingle cell profiles (while drop-in alle-les were only detected as unique events),a clear and full profile can be deduced bythe combined analysis of all the 11 singlecells.

Sample II

Theprofiles of 4 of the recovered 17whiteblood cells showed no alleles at all. Forthe remaining13cellspartial STRprofileswere obtained. These 13 partial profilescould be assigned to the 2 reference pro-files, 6 to the victim and 7 to the suspectprofile (. Table 3). The obtained partialprofiles showed 11 up to 27 of the ex-pected 29 (victim) and 14 up to 28 of theexpected 30 (suspect) alleles. Combin-ing the alleles assigned to each person ineach single cell profile, the full profiles oftwo individuals could be deduced. Every

single allele was detected in at least threeout of six (profiles assigned to the victim)or seven (profiles assigned to the sus-pect) cells. The highest number of allelicdrop-out could be observed in the STR-loci VWA, TH01, D8S1179 and D21S11and thus the observed allelic drop-outsare more likely to be a consequence ofstochastic phenomena [16] rather thanthe consequence of DNA degradation.Moreover, three additional alleles (clas-sified as allelic drop-ins) were observed(. Table 3), one of which (allele 29 inlocus D21S11, sample SC-5 WBC) waslocated in a minus one stutter positionof the actual suspects’ allele 30. Nev-ertheless, due to the peak height ratioof approximately 45% this peak was alsorated as a drop-in. All three drop-in alle-les occurred only once in three differentsamples and loci, where two other alleleshad already been proved several timesand therefore the true genotype of thisindividual couldbeclearly reconstructed.In summary, complete profiles could bederived for both persons involved in themixture.

Sample III

In contrast, from the 17 WBCs recov-ered from the GEDNAP stain (blood-bloodmixture fromtwoindividuals)onlytwo partial profiles could be obtained(. Figs. 3 and 4). The profiles showed dif-ferent alleles and based on the assump-tion, that no allelic drop out occurred inthe amelogenin, these are the profiles ofa male (. Fig. 3) and a female (. Fig. 4)person. For the remainingcells, noallelescould be detected at all. Under use of theabove established values for stutter inter-pretation, clear signals were labelledwiththe corresponding allele naming. Forthe male profile the two alleles in locusD10S1248 andD2S1338 showed extremeimbalances (both not in plus or minusone repeat stutter position). Therefore,the smaller alleles were classified as ques-tionable andnotas truealleles. Thesizeofallele 6, (locus TH01) and allele 18 (locusD12S391) in the female profile, as well asallele 14 (locus D19433) in the male pro-file show peak heights which were abouttwice as high as the peaks of the neigh-boring heterozygote loci. For these loci

homozygote genotypeswere reported. Inthis way two profiles were deduced; onemaleprofilewithat least twoquestionableand four possibly missing alleles (drop-outs) and a female profile with at leastfour possibly missing alleles (drop-outs).Thededucedpartial genotypeswere com-pletely confirmed by the organizer of theGEDNAP proficiency test. In summary,although the results are based on sin-gle analyses, the derived, albeit partialprofiles could be correctly deduced.

In comparison to sample I and II itis obvious that only 2 of the 17 cellscould be profiled. This fact did not meetthe expectations after the first two at-tempts. Looking for an explanation, theimages of the selected cells from sam-ple I and III were compared (. Fig. 5).Whereas the images of the WBCs fromsample I showed clear, well-demarcatedimages for the PE as well as bright fieldand DAPI staining, the signals of mostof the WBCs from sample III appearedmore diffuse. This may give a hint onthe constitution of the cells and thereforemaybe also the DNA quality. Accordingto the information provided by the or-ganizer of the GEDNAP proficiency test,at the time of investigation the stain wasapproximately 2 years old and thereforeolder than samples I and II. In addition, itwas communicated that the mixed sam-ples of the proficiency test were mostlytreated with surfactants. This also mayhave an influence of the integrity of thecell or nuclear membrane. Damage ofthe cell membrane in turn may have aninfluence on the staining and the perfectmovement of the cells in the extractioncartridge. Also, a loss of DNA wouldbe conceivable. Whether there really isa connectionhas tobe resolvedby furthertesting.

Sample IV

For all of the recovered 17 WBCs fromsample IV, partial profiles were obtained(. Table 4) and 3, 6 and 8 profiles can beassigned to an individual (named A, Band C, respectively). The obtained par-tial profiles showed 21 up to 29 of theexpected 31 (A), 24 up to 29 of the ex-pected 33 (B) and 18 up to 22 of the ex-pected 23 (C) alleles. A total of 16 alleles


Original articleTable3

Profilingresults

ofthewhitebloodcells(W

BC)fromsampleII

DNA

system

D3S

1358

VWA

FIBR

ATH

O1

SE33

D8S

1179

D21

S11

D18

S51

Amel

D16

S539

D2S

1338

D19

S433

D22

S104

5D1S

1656

D10

S124

8D2S

441

D12

S391

Num

ber

ofde

-tected

true

alleles

Reference

victim

14/17

17/18

256/7

14/16

1330.2/31.2

15/16

X11/12

19/23

12/15

1611/16

14/16

13/14

1729

SC-1WBC

1417/18

256/7

14/16

1330.2/31.2

15/16

X12

19/23

12/15

1611/16

14/16

13/14

1727

SC-3WBC

14/17

1725

6/7

14/16

1331.2

15/16

X11/12

2312/15

1611/16

14/16

13/14

1726

SC-12WBC

14/17

17/18

256

1413

30.2/31.2

15/16

X11/12

19/23

12/15

1611/16

14/16

13/14

1727

SC-15WBC

14/17

17/18

253/6/7

14/16

1331.2

15/16

X11/12

1912

1611/16

1613/14

1725

SC-16WBC

17/18

1725

6/7

14/16

1330.2/31.2

–X

11/12

19/23

12/15

1611/16

14/16

13/14

1725

SC-17WBC

14/17

17–

6/7

––

––

X–

––

16–

14/16

1317

11

Deduced

profile

14/17

17/18

256/7

14/16

1330.2/31.2

15/16

X11/12

19/23

12/15

1611/16

14/16

13/14

1729

Reference

suspect

15/16

17/19

22/24

6/9.3

18/30.2

11/12

28/30

15/18

X/Y

1224

13.2/14.2

1515

13/14

10/14

17/19.3

30

SC-4WBC

15/16

17/19

22/24

618/30.2

1128/30

15X/Y

1224

13.2/14.2

1515

13/14

10/14

17/19.3

27

SC-5WBC

––

––

18/30.2

1129/30

18X/Y

–24

13.2/14.2

1515

13–

–14

SC-6WBC

15/16

1722/24

9.3

1811

2815/18

X/Y

1224

13.2/14.2

–15

13/14

1017/19.3

23

SC-8WBC

15/16

17/19

22/24

9.3

18/30.2

11/12

28/30

15/18

X/Y

1224

13.2/14.2

1515

13/14

10/14

19.3

28

SC-9WBC

15/16

17/19

22/24

9.3

18/30.2

11/12

2815/18

X/Y

–24

13.2

1515

1314

17/19.3

24

SC-11WBC

15/16

17/19

226

18/30.2

–28/30

15/18

X/Y

1224

13.2/14.2

1515

1410/14

17/19.3

25

SC-14WBC

15/16

17–

6/9.3

18/30.2

11/12

28/30

15X/Y

–24

13.2/14.2

1515

13/14

10/14

17/19.3

25

Deduced

profile

15/16

17/19

22/24

6/9.3

18/30.2

11/12

28/30

15/18

X/Y

1224

13.2/14.2

1515

13/14

10/14

17/19.3

30

Boldtype

indicatesalleleso

nlydetected

once


X / Y 15 / 178 / 9.3

28 / 31 15 / ?

14 / ? 18.3 / ? 19 / ?14 / ?

11 / 16 17 / 18 13 / ? 20 / 22

11 / 11.3 23 / 25 14 / 1414 / 27.2

*

**

*

Fig. 38 Male profile ofwhite blood cells (WBC) SC-17 from sample III (GEDNAP55, stain 2) evaluatedunder use of the estab-lished values for the classification of true alleles (in stutter position>35%, in cases of heterozygote imbalances smaller peak>50%). The deduced genotype for all loci can be found in the corresponding text boxes.Possible drop-out positions (wherea distinction between the presence of an allelic drop-out or an actual homozygote genotypewas not possible) are indicatedwith *

weredetectedonlyonceand thereforenotclassified as true alleles (possible drop-in alleles). All other alleles were detectedat least three out of six (A), two out ofthree (B) or five out of 8 (C) times. Of the16allelesprovenonlyonce13occurred inloci where 1 (in the case of a homozygotegenotype) or 2 (in the case of a heterozy-gote genotype) other alleles had alreadybeenproven several times. Therefore, thetrue genotype of this individual could bededuced. The other three alleles werefound in the three profiles assigned toindividual B (allele 18 in locus VWA andalleles 14.2 and 15 in locus D19S433).A comparison with the reference pro-file of the corresponding person showedthat these were indeed the true alleles ofthe corresponding individuals. So oncemore itwas shown that for the certain andcomplete derivation of a genotype, a suf-ficiently large number of partial profiles

should be used in order to distinguishartifacts from true alleles. Thereby thenumber of the required profiles directlycorrelates with the quality.

For this sample, two complete andone almost complete partial profile werefinally deduced, which all completelymatched the alleles of the referencesamples and therefore demonstratedthe effectiveness and correctness of thismethod. Sample IV again was a freshand untreated sample. Compared tosample III, significantly better typingresults (partial profiles for all 17 recov-ered WBCs) could be obtained. Thisagain underlines the presumption thatthe success of the technology seems todepend very much on the quality of thecells and/or DNA.

Conclusion

In summary, it can be stated thatSTR profiling of single cells isolatedby DEPArrayTM technology can be usedsuccessfullyforthedeconvolutionofmix-tures containing white blood cells fromdifferent contributors. While the bene-fits of the technology in the separationof heterogeneous mixtures composed ofdifferent cell types have already beendemonstrated impressively several times[3, 8, 21], in the current study for the firsttime DNA profiles of each individualcontributor could be reconstructed frommixtures composed of cells of the sametype through the isolation and profilingof multiple single cells.

In contrast to software-based decon-volution models, this approach is well-suited for the investigation of mixtureswith balanced mixture ratios. Software-


Original article

X / X * *

* *

16 / 18 6 / 6 30 / ? 17/ ?

13 / 15 12 / ? 21 / 24 11 / ?

11 / 16 16 / 17 12 / 15 21 / 24

10 / 11 18 / 1812 / 14

20 / 22.2

Fig. 48 Female profile ofwhite blood cells (WBC) SC-17 from sample III (GEDNAP55, stain2) evaluatedusing the establishedvalues for the classificationof true alleles (in stutter position>35%, in cases of heterozygote imbalances smaller peak>50%).The deducedgenotype for all loci canbe found in the corresponding text boxes.Possible drop-out positions (where a distinc-tion between the presence of an allelic drop or an actual homozygote genotypewas not possible) are indicatedwith *

WBCs sample IPE

WBCs sample IIIDAPI Bright field PEDAPI Bright field

Fig. 59 Exemplary im-ages (eachwith the DAPI[4’,6-Diamidin-2-phenylin-dol], bright field aswell asPE [Phycoerythrin] images)of twoselectedwhitebloodcells (WBC) from samples Iand III are shown for com-parison

based models perform deconvolution byutilizing the peak height information forinferring the most likely profile geno-types for the unknown contributors [5].This can be used very well for deduc-ing the alleles of a major component;however, the more balanced a mixture

is the less meaningful are the calculatedprobabilities for the possible genotypesof the different contributors. In addi-tion, for samples with very low amountsof DNA, reliable estimations could notbe achieved because the peak height lev-els are too much influenced by stochastic

variations. Therefore, depending on themixture ratio and amount of templateDNA,STRprofilingofsinglecells isolatedby DEPArrayTM may be an advantage.

From the results of sample III it canbe concluded that the success of the tech-nology presented here, however, seems


Table4

Profilingresults

ofthewhitebloodcells(W

BC)fromsampleIV

DNA

system

D3S

1358

VWA

FIBR

ATH

O1

SE33

D8S

1179

D21

S11

D18

S51

Amel

D16

S539

D2S

1338

D19

S433

D22

S104

5D1S

1656

D10

S124

8D2S

441

D12

S391

Num

ber

ofde

-tected

true

alleles

ReferenceA

15/17

16/17

21/25

7/9

1910/15

30/30.2

12/18

X/Y

1217/23

13/16.2

14/16

11/17.3

14/16

1122/23

31

SC-1WBC

15/17

16/17

21/25

7/9

1910/15

30/30.2

12/18

X12

2313/16.2

1417.3

1411

2225

SC-2WBC

15/17

16/17

21/25

7/9

1910/15

30/30.2

12/18

X/Y

1217

13/16.2

1611/17.3

14/16

1122/23

29

SC-4WBC

1516/17

21/25

6/7

1910/15

30/30.2

12X/Y

1217

15/16.2

–11/14

–11

22/23

21

SC-5WBC

15/17

16/17

21/25

7/9

1910/15

29/30/

30.2

12/18

Y12

2313

14/16

11/17.3

14/16

1122/23

28

SC-14WBC

15/17

1621/25

7/9

1910

–12/18/

19X/Y

1217/23

13/16.2

13/14

11/17.3

14/16

1122/23

26

SC-15WBC

–16

21/25

919

1030/30.2

12/18

X12

2313/16.2

14/16

1114/16

1122/23

23

Deduced

profileA

15/17

16/17

21/25

7/9

1910/15

30/30.2

12/18

X/Y

1217/23

13/16.2

14/16

11/17.3

14/16

1122/23

31

ReferenceB

15/18

17/18

20/24

6/9.3

20/27.2

1329/31.2

12/14

X/Y

9/11

20/25

14.2/15

11/14

11/14

13/15

14/15

19/21

33

SC-9WBC

15/18

–20/24

9.3

20/27.2

1329

14Y

9/11

20/25

–14

11/14

13/15

14/15

19/21

24

SC-10WBC

15/18

1720/24

6/9.3

20/27.2

1329

12/14

X/Y

9/11

2015

11/14

1113

1419/21

26

SC-11WBC

15/18

17/18

20/24

620/27.2

13/15

29/31.2

11/12/

14X/Y

9/11

2514.2

11/14

11/14

13/15

14/15

19/21

29

Deduced

profileB

15/18

17/?

20/24

6/9.3

20/27.2

1329

12/14

X/Y

9/11

20/25

–11/14

11/14

13/15

14/15

19/21

29

ReferenceC

16/18

1822/24

615/26.2

1430

13/14

X12

17/21

1416

1114

1117/19.3

23

SC-3WBC

16/18

1822/24

615/26.2

1430

13/14

X12

1714

1611

1411

17/19.3

22

SC-6WBC

16/18

1821/24

615/26.2

–30

13/14

X12

1714

1611

1411

17/19.3

20

SC-7WBC

16/18

1822

615/26.2

1430

13X

1217/20/21

1416

1114

1117/19.3

21

SC-8WBC

1818

22/24

626.2

1430

14X

1217/21

1416

1114

1117/19.3

20

SC-12WBC

16/18

1822/24

615/18/

26.2

1430

13/14

X12

17/21

–16

1114

1117/19.3

22

SC-13WBC

16/18

1822/24

615/26.2

1430

13/14

X12

17/21

1416

1114

1116/17

22

SC-16WBC

16/18

1824

–15

1430

13/14

X12

1714

1611

1411

19.3

18

SC-17WBC

1617/18

246

15/26.2

1430

13/14

X12

2114

1611

1411

17/19.3

20

Deduced

profileC

16/18

1822/24

615/26.2

1430

13/14

X12

17/21

1416

1114

1117/19.3

23

Boldtype

indicatesallelesd

etectedonlyonce


Original article

to depend verymuchon the quality of thecells and/or DNA. Single cell STR pro-filing itself is already a challenge, evenwhenDNA is intact. Degradation or lossof DNA may very quickly lead to com-pletelynegative results. Furthermore, thestructure of the cell membrane must bepreserved to a level that the staining andthemovement of the cells in the cartridgeis still possible without restriction.

The single cells treated in this studyyielded only partial profiles. In order toobtain complete profiles, all partial pro-files (if available), which were attributedto the same single personwere combinedto form a consensus sequence. Artifactscan also be better recognized in this way,so that the true alleles can be deducedbeyond doubt. Therefore, with increas-ing complexity of themixture (increasingnumber of contributors), the number ofselected cells has to be increased. Withthe DEPArrayTM V2 system used in thisstudy, amaximumof17 single cells canberecovered from 1 cartridge. In contrast,the new generation, DEPArrayTMN xTinstrument, allows separation of up to48 single cells.

As expected, when low amounts ofDNA were amplified using an increasedPCR cycle program, beside drop-outssome other characteristic artefacts oc-curred more often. Initial guidelinevalues were determined in the courseof this feasibility study; however, forthe routine application of this technol-ogy in forensic casework, appropriateguidelines should be developed basedon a larger dataset. With the help ofsuch guidelines, STR profiling of singlecells isolated by DEPArrayTM technologyopens up completely new possibilities inthe field of mixture deconvolution.

Corresponding address

PD Dr. K. AnslingerInstitute of Legal Medicine, Ludwig-Maximilians-UniversityMunichNußbaumstr. 26, 80336 Munich, [email protected]

Acknowledgements. The authors thank Menar-ini Silicon Biosystems for the provision of theDEPArrayTM system as well as for the valuable sup-port and Eppendorf for providing necessary labora-tory equipment.

Compliance with ethicalguidelines

Conflict of interest. K. Anslinger,M. GrawandB. Bayerdeclare that theyhavenocompeting interests.

The investigationswere carriedout according to thespecifications of the Central Ethics Committee of theFederalMedical Council.

OpenAccessThis article is distributedunder the termsof the Creative CommonsAttribution 4.0 InternationalLicense (http://creativecommons.org/licenses/by/4.0/), which permits unrestricteduse, distribution,and reproduction in anymedium, provided yougiveappropriate credit to the original author(s) and thesource, providea link totheCreativeCommons license,and indicate if changesweremade.

References

1. Anslinger K, Mack B, Bayer B, Rolf B, EisenmengerW (2005) Digoxigenin labelling and laser capturemicrodissection of male cells. Int J Legal Med119:374–377

2. Anslinger K, Mack B, Bayer B, Eisenmenger W(2007) Sex-specificfluorescent labellingof cells forlasermicrodissectionandDNAprofiling. Int JLegalMed121:54–56

3. Anslinger K, Bayer B, von Máriássy D (2018)

Application of DEPArrayTM technology for theisolation ofwhite blood cells from cell mixtures inchimerismanalysis. Rechtsmedizin28:134–137

4. Benschop CCG, van de Merwe L, de Jong J,VanvoorenV,KempenaersM,KeesvanderBeekCP,Barni F, Reyes EL,Moulin L, PeneL,HanedH, SijenT(2017) Validation of SmartRank: a likelihood ratiosoftware for searching national DNA databaseswith complex DNAprofiles. Forensic Sci Int Genet29:145–153

5. Bleka Ø, Benschop CCG, Storvik G, Gill P (2016) Acomparative study of qualitative and quantitativemodels used to interpret complex STR DNAprofiles. ForensicSci IntGenet25:85–96

6. Elliott K, Hill D, Lambert C, Burroughes T, GillP (2003) Use of laser microdissection greatlyimproves the recovery of DNA from sperm onmicroscopeslides. ForensicSci Int137:28–36

7. FarashK,HansonEK, Ballantyne J (1997) Enhancedgenetic analysis of single human bioparticlesrecovered by simplified micromanipulation fromforensic “touchDNA”evidence. J Vis Exp9. https://doi.org/10.3791/52612

8. Fontana F, Rapone C, Bregola G, Aversa R, de MeoA, Signorini G, Sergio M, Ferrarini A, LanzellottoR, Medoro G, Giorgini G, Manaresi N, Berti A(2017) Isolation and genetic analysis of pure cellsfrom forensic biological mixtures: the precisionof a digital approach. Forensic Sci Int Genet29:225–241

9. Geng T, Novak R, Mathies RA (2014) Single-cell forensic short tandem repeat typing withinmicrofluidicdroplets. AnalChem86:703–712

10. Gill P, Jeffreys A, Werrett D (1985) Forensicapplication of DNA “fingerprints”. Nature318:577–579

11. Gill P, Sparkes R, Pinchin R, Clayton T, WhitakerJ, Buckleton J (1998) Interpreting simple STRmixtures using allele peak areas. Forensic Sci Int9:41–53

12. Gill P, Brenner CH, Buckleton JS, Carracedo A,Krawczak M, Mayr WR, Morling N, Prinz M,Schneider PM, Weir BS (2003) DNA commissionof the International Society of Forensic Genetics:recommendations on the interpretation ofmixtures; DNA commission of the internationalsociety of forensic genetics. Forensic Sci Int160:90–101

13. Marshall PL, King JL, Budowle B (2015) Utility ofamplification enhancers in low copynumberDNAanalysis. Int JLegalMed129:43–52

14. McLaren RS, Bourdeau-Heller J, Patel J, ThompsonJM, Pagram J, Loake T, Beesley D, PirttimaaM, Hill CR, Duewer DL, Kline MC, Butler JM,Storts DR (2014) Developmental validation of thepowerPlex(®) ESI 16/17 fast and powerPlex(®)ESX 16/17 fast systems. Forensic Sci Int Genet13:195–205

15. Schneider H, Sommerer T, Rand S, Wiegand P(2011) Hot flakes in cold cases. Int J Legal Med125:543–548

16. Timken MD, Klein SB, Buoncristiani MR (2014)Stochastic sampling effects in STR typing:implications for analysis and interpretation.ForensicSci IntGenet11:195–204

17. Ulbrich W, Anslinger K, Bäßler G, Eckert M, Fim-mers R, Hohoff C, Kraft M, Leuker C, MolsbergerG, Pich U, Razbin S, Schneider H, Templin M,Wächter A, Weirich V, Zierdt H, Schneider PM(2016) Gemeinsame Empfehlungen der Projek-tgruppe „Biostatistische DNA-Berechnungen“und der Spurenkommission zur biostatistischenBewertung von DNA-analytischen Befunden.Rechtsmedizin26:291–298

18. VerdonTJ,MitchellRJ,ChenW,XiaoK,VanOorschotRAH (2015) FACS separation of non-compromisedforensically relevant biologicalmixtures. ForensicSci IntGenet14:194–200

19. Yano S, Honda K, Kaminiwa J, Nishi T, IwabuchiY, Sugano Y, Kurosu A, Suzuki Y (2014) DNAextraction for short tandem repeat typing frommixed samples using anti-human leukocyte CD45and ABO blood group antibodies. Forensic Sci IntGenet10:17–22

20. Whitaker JP, Cotton EA, Gill P (2001) A comparisonof the characteristics of ptofiles produced withtheAMPFISTR SGMPlusmultiplex system for bothstandard and low copy number (LCN) STR DNAanalysis. ForensicSci Int123:215–223

21. Williamson VR, Laris TM, Romano R, MarcianoMA (2018) EnhancedDNAmixture deconvolutionof sexual offense samples using the DEPArray™system. ForensicSci IntGenet34:265–276


http://creativecommons.org/licenses/by/4.0/http://creativecommons.org/licenses/by/4.0/https://doi.org/10.3791/52612https://doi.org/10.3791/52612

Deconvolution of blood-blood mixtures using DEPArrayTM separated single cell STR profilingAbstractZusammenfassungIntroductionMixture deconvolutionDEPArrayTM technologySingle cell STR profilingStudy design

Material and methodsMaterialCell separation using DEPArrayTM technologyDNA extraction and STR profiling

Results and discussionSample ISample IISample IIISample IV

ConclusionReferences

rechtsmedizin2019·29:30–40 ... · dnaextractionandstrproﬁling the dna was isolated with the...

Documents