Tripolar chromosome segregation drives the spanning PLK4 ... · as preimplantation genetic screening (PGS) or preimplantation genetic diagnosis for aneuploidy (PGD-A) but now termed

A S S O C I A T I O N S T U D I E S A R T I C L E

Tripolar chromosome segregation drives the

association between maternal genotype at variants

spanning PLK4 and aneuploidy in human

preimplantation embryosRajiv C McCoy12dagger Louise J Newnham3dagger Christian S Ottolini4Eva R Hoffmann35 Katerina Chatzimeletiou6 Omar E Cornejo7Qiansheng Zhan8 Nikica Zaninovic8 Zev Rosenwaks8 Dmitri A Petrov9Zachary P Demko10 Styrmir Sigurjonsson10 and Alan H Handyside4

1Department of Ecology and Evolutionary Biology Princeton University Princeton NJ 08544 USA 2Departmentof Genome Sciences University of Washington Seattle WA 98195 USA 3Genome Damage and Stability CentreUniversity of Sussex Brighton BN1 9RQ UK 4School of Biosciences University of Kent Canterbury CT2 7NJ UK5Department of Cellular and Molecular Medicine DNRF Center for Chromosome Stability University ofCopenhagen 2200 Copenhagen N Denmark 6Section of Reproductive Medicine First Department of Obstetricsamp Gynaecology Aristotle University Medical School Papageorgiou General Hospital Thessaloniki 56403Greece 7School of Biological Sciences Washington State University Pullman WA 99164 USA 8Ronald OPerelman and Claudia Cohen Center for Reproductive Medicine Weill Cornell Medicine New York NY 10065USA 9Department of Biology Stanford University Stanford CA 94305 USA and 10Natera Inc San Carlos CA94070 USA

To whom correspondence should be addressed at Department of Ecology and Evolutionary Biology Princeton University 106A Guyot Hall Princeton NJ08544 USA Tel thorn1 6092589383 Fax thorn1 6092581712 Email rajivmccoyprincetonedu

AbstractAneuploidy is prevalent in human embryos and is the leading cause of pregnancy loss Many aneuploidies arise duringoogenesis increasing with maternal age Superimposed on these meiotic aneuploidies are frequent errors occurring duringearly mitotic divisions contributing to widespread chromosomal mosaicism Here we reanalyzed a published datasetcomprising preimplantation genetic testing for aneuploidy in 24 653 blastomere biopsies from day-3 cleavage-stage embryosas well as 17 051 trophectoderm biopsies from day-5 blastocysts We focused on complex abnormalities that affected multiplechromosomes simultaneously seeking insights into their formation In addition to well-described patterns such as triploidyand haploidy we identified 47 of blastomeres possessing characteristic hypodiploid karyotypes We inferred this signatureto have arisen from tripolar chromosome segregation in normally fertilized diploid zygotes or their descendant diploid cellsThis could occur via segregation on a tripolar mitotic spindle or by rapid sequential bipolar mitoses without an intervening

daggerThese authors contributed equally to this workReceived February 9 2018 Revised April 3 2018 Accepted April 18 2018

VC The Author(s) 2018 Published by Oxford University Press All rights reservedFor permissions please email journalspermissionsoupcom

Human Molecular Genetics 2018 Vol 0 No 0 1ndash13

doi 101093hmgddy147Advance Access Publication Date 24 April 2018Association Studies Article

Downloaded from httpsacademicoupcomhmgadvance-article-abstractdoi101093hmgddy1474983967by University of Leeds - Librarian useron 19 May 2018

S-phase Both models are consistent with time-lapse data from an intersecting set of 77 cleavage-stage embryos which wereenriched for the tripolar signature among embryos exhibiting abnormal cleavage The tripolar signature was strongly associ-ated with common maternal genetic variants spanning the centrosomal regulator PLK4 driving the association we previouslyreported with overall mitotic errors Our findings are consistent with the known capacity of PLK4 to induce tripolar mitosis orprecocious M-phase upon dysregulation Together our data support tripolar chromosome segregation as a key mechanismgenerating complex aneuploidy in cleavage-stage embryos and implicate maternal genotype at a quantitative trait locusspanning PLK4 as a factor influencing its occurrence

IntroductionAneuploidy is common in human preimplantation embryosand increases in frequency with maternal age (1) While manyaneuploidies originate in meiosismdashprimarily in females (oogen-esis) rather than males (spermatogenesis) (23)mdasha substantialproportion arise after fertilization (postzygotic) due to errantchromosome segregation during early cleavage divisions (4)These mitotic errors produce mosaic embryos with two or morecell lineages possessing distinct chromosomal complementsBecause aneuploidy is associated with negative pregnancy out-comes many patients undergoing in vitro fertilization (IVF)treatment for infertility have their embryos tested by copy num-ber analysis of all 24 chromosomes Testing is applied to singleor small numbers of biopsied cells with the aim of transferringonly embryos that are euploidmdashan approach previously knownas preimplantation genetic screening (PGS) or preimplantationgenetic diagnosis for aneuploidy (PGD-A) but now termed pre-implantation genetic testing for aneuploidy (PGT-A) (56)

Despite substantial improvements in detecting aneuploidyat the single cell level distinguishing aneuploidies of meioticand mitotic origin remains challenging because the chromo-somal signatures can be similar In some cases discriminationmay be achieved by incorporating parental genotype informa-tion to determine the parental origin of each embryonic chro-mosome Specifically observation of both maternal (or rarelypaternal) haplotypes transmitted in a homologous region of anembryonic chromosomemdasha signature we term lsquoboth parentalhomologsrsquo or lsquoBPHrsquomdashprovides strong evidence of the meiotic ori-gin of a trisomy (Fig 1A) Meanwhile because aneuploidy is rarein sperm (1ndash4) (7) and paternal BPH affects only 1 of preim-plantation embryos (8) aneuploidies involving gain or loss ofpaternal homologs are predominantly mitotic errors To dateseveral methods have been developed to infer the transmissionof individual parental homologs based on single-nucleotidepolymorphism (SNP) microarray data facilitating classificationof meiotic and mitotic aneuploidies These include ParentalSupport (910) and karyomapping (1112) which have recentlybeen validated for linkage-based preimplantation genetic diag-nosis (PGD) of single gene disorders (1314)

These technical improvements in PGT-A methodologieshave provided substantial insights into the mechanisms of an-euploidy formation An adapted form of karyomapping for ex-ample recently revealed a novel mechanism of meiotic errortermed lsquoreverse segregationrsquo whereby sister chromatids sepa-rate at meiosis I followed by (occasionally errant) segregation ofnon-sister chromatids at meiosis II (12) This added to prema-ture separation of sister chromatids (PSSC) and meiotic non-disjunction as the predominant known mechanisms of meioticerror (1215) Mitotic aneuploidies meanwhile have tradition-ally been attributed to mitotic non-disjunction whereby chro-matids fail to separate as well as anaphase lag wherebychromatids are lost after delayed migration toward the spindlepoles (16)

A distinct class of severe mitotic aneuploidy the origins ofwhich have remained poorly understood was previously re-ferred to as lsquochaotic mosaicismrsquo due to its seemingly randomchromosomal constitution (17) Though prevalent at the cleav-age stage chaotic mosaic embryos rarely achieve blastocyst for-mation (18) One hypothesized source of chaotic aneuploidy isthe formation of multipolar mitotic spindles which may causethe cell to cleave into three or more daughter cells Recognizedsince the foundational work of Boveri in the late ninteenth cen-tury (19) multipolar spindles are in turn thought to form due toabnormalities of the centrosome As the dynamic center of so-matic cells the centrosome is the key organelle coordinatingmicrotubule organization cell cycle progression and chromo-some segregation Supernumerary (gt2) centrosomes can lead tothe formation of a multipolar rather than normal bipolar spin-dle (20) If proceeding through anaphase chromosomes maythen segregate among multiple daughter cells

During gametogenesis sperm and egg cells undergo recipro-cal reductions in centrosome components (21) Upon fertiliza-tion these complementary elements are reunited and initiateformation of the sperm aster which then enlarges and maturesto organize the first mitotic spindle (21) In humans and mostother mammals (with the notable exception of rodents) thesperm is the dominant contributor of centrosome componentsincluding the proximal centriole (22ndash24) Although humanoocytes do not contain centrosomes or microtubule organizingcenters during their meiotic divisions (25) key centrosomal pro-teins such as c-tubulin and NuMA are recruited from theooplasm after the switch to mitotic division (26) Given thismode of inheritance excess centrosomes and tripolar mitosisare sometimes caused by abnormal fertilization This includesfertilization with two sperm or retention of the second polarbody producing diandric or digynic tripronuclear (3PN) zygotesrespectively (2728) Surprisingly however tripolar spindleshave also been observed by laser confocal microscopy both atthe cleavage and blastocyst stages in embryos derived from nor-mally fertilized dipronuclear (2PN) zygotes (18) albeit at a lowerrate than for 3PN zygotes Indeed tripolar mitosis [also calledlsquotrichotomic mitosisrsquo or lsquodirect unequal cleavage (DUC)rsquo] has re-cently been documented in a substantial proportion (17ndash26) ofcleaving embryos based on time-lapse imaging (29ndash32) evenafter intracytoplasmic sperm injection (ICSI) which virtuallyeliminates polyspermic fertilization Rather than abnormal cen-trosome transmission these observations suggest possible dys-regulation of centrosome duplication

Ottolini et al (33) recently detailed the phenomenon of tripo-lar mitosis in diploid cells using time-lapse analysis and PGT-Aof 51 normally fertilized embryos obtained from three IVFpatients In addition to multi-cell biopsies this study includedkaryomapping of 74 single cells disaggregated from 13 arrestedcleavage-stage embryos as well as 9 cells extruded from a totalof 26 developing embryos around the time of blastocyst forma-tion (33) PGT-A data revealed extensive evidence of chaotic

2 | Human Molecular Genetics 2018 Vol 00 No 00

Euploid

Aneuploid

Digynic Triploid Near-Triploid

Diandric Triploid Near-Triploid

Maternal Haploid Near-Haploid

Paternal Haploid Near-Haploid

Digynic TriploidTripolar

Diandric TriploidTripolar

Diploid Tripolar

Low Confidence Call

MaternalPaternal

XXXXXYXY

XXY (Mat)

XXY (Mat)XXX (Mat)XY

XXY (Pat)XXY (Pat)

XXY (Pat)XXX (Pat)

X0 (Mat) X0 (Mat)

X0 (Mat)

X0 (Pat)0Y

X0 (Pat)X0 (Pat)

XYXXXXY (Mat)X0 (Pat)

XXXX00XY (Pat)

X0 (Mat)XX

Chromosome Ploidy

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Sex

Single cell biopsy

(n = 24695)Day 3 Cleavage-Stage Embryo

Day 5 Blastocyst-Stage Embryo

Multi-cell trophectoderm

biopsy(n = 17053)

Genome-wide SNP genotypingalgorithm infers copy number

and transmitted parental haplotypes across all 23 chromosomes

A GameteProgenitor cell

Normalgametogenesis

Error in MI

Error in MII

Mitotic error

SPHError in MIINo recombination

Embryonic cell Daughter cell

Mitosis

trophectoderm(TE)

inner cell mass(ICM)

Figure 1 Patterns of chromosome abnormalities observed in the preimplantation human embryos (A) Schematic describing the signature of meiotic chromosome

gain adapted from Rabinowitz et al (10) and McCoy et al (8) The detection of non-identical homologous chromosomes inherited from a single parent (a signature that

we term lsquoboth parental homologsrsquo or lsquoBPHrsquo) indicates a meiosis I (MI) or meiosis II (MII) error Chromosome gains that involve identical chromosomes from a single par-

ent (termed lsquosingle parental homologrsquo or lsquoSPHrsquo) can occur either due to mitotic errors or MII errors in the absence of recombination (B) Schematic cross-section showing

the sample sizes of cleavage and blastocyst-stage embryo biopsies which underwent genome-wide PGS Both parents were also genotyped using the same SNP micro-

array platform to facilitate inference of chromosome copy number and determine the parental origin of each chromosome This approach also enables classification of

trisomies of meiotic origin (C) Examples of patterns of whole-chromosome copy number variation observed in the dataset Most samples were normal (euploid) or

contained a few single-chromosome imbalances (aneuploid) However complex abnormalities were also relatively prevalent (Table 1) Each column represents a chro-

mosome (1ndash22 left to right along with the sex chromosomes at the far right) Each row represents a distinct embryo biopsy Four representative biopsies are depicted

for each category of chromosome abnormality M maternal P paternal BPH both parental homologs

3Human Molecular Genetics 2018 Vol 00 No 00 |

chromosome loss in both extruded cells and arrested embryosconsistent with frequent tripolar mitosis and compromised po-tential for proliferation (33) While this study was critical forelucidating the karyotypic outcomes and developmental conse-quences of diploid chromosome segregation on tripolar spin-dles the small sample of 51 embryos from three selectedpatients provides limited information about the prevalence ofthis phenomenon or its underlying genetic basis Such informa-tion is crucial for better understanding the low baseline rates ofblastocyst formation in vitro (50) as well as patient-specificfactors driving variation in these rates

Here we leverage a large published PGT-A dataset (834) togain further insight into the molecular origins of chaotic aneu-ploidy These data comprise 46 439 embryo biopsies genotypedby SNP microarray with aneuploidies inferred using theParental Support algorithm (9) (Fig 1B) This approach providesseveral advantages for the study of aneuploidy The first advan-tage is the resolution of the data which include estimates ofcopy number across all 24 chromosomes as well as identifica-tion of meiotic chromosome gains Second the large samplesize facilitates detection and quantification of rare forms ofchromosome abnormality Finally by including both cleavage-and blastocyst-stage embryos the dataset provides valuableinformation about the early developmental implications of vari-ous forms of aneuploidy Based on our analysis we proposethat tripolar chromosome segregation in normally fertilized2PN zygotes or their descendant cells is a key mechanism con-tributing to chaotic aneuploidy in cleavage-stage embryosFurthermore we provide evidence that this mechanism drivesthe association we previously reported (34) between mitotic-origin aneuploidy and maternal genotype at common geneticvariants spanning PLK4 a critical regulator of centrosomeduplication

ResultsMitotic-origin aneuploidy is prevalent inpreimplantation human embryos

Excluding replicate and uninformative samples (10 chromo-somes with low-confidence or nullisomic calls) we quantifiedvarious forms of chromosomal abnormality in 41 704 embryo bi-opsies from a total of 6319 PGT-A cases Maternal age includingpatients and egg donors ranged from 18 to 48 (medianfrac14 37)while paternal age ranged from 21 to 77 (medianfrac14 39)Indications for PGT-A were diverse and included advanced ma-ternal age recurrent pregnancy loss sex selection previous IVFfailure male factor infertility unexplained infertility previousaneuploidy and translocation carrier status (8)

The overall rate of euploidy among day-3 blastomere biop-sies was 380 compared with 555 for day-5 trophectodermbiopsies (Table 1) We note that in light of mosaicism which iscommon during preimplantation development (1435) inferredploidy of the embryo biopsy may not reflect the ploidy of therest of the embryo Among samples with 1ndash2 chromosomeswith abnormal copy number a total of 2040 (83) day-3 blasto-mere samples possessed at least one putative meiotic (BPH)chromosome gain compared with 1619 (95) day-5 trophecto-derm samples (Table 1) Conversely a total of 1536 (62) suchday-3 blastomere samples and 930 (55) such day-5 trophecto-derm samples possessed only putative mitotic errors (loss orSPH gain of1 paternal chromosome with no co-occurring ma-ternal or paternal BPH gains Table 1) Tab

971ndash

In addition to these less severe aneuploidies we sought toinvestigate the origin of complex abnormalities that affectedthree or more chromosomes simultaneously These include so-called lsquochaoticrsquo aneuploidies characterized by seemingly ran-dom chromosome complements (17) We identified a total of10 291 samples exhibiting complex abnormalities a subset ofwhich could be classified into distinct categories (Fig 1CTable 2) The proportion of complex abnormalities was signifi-cantly greater in day-3 blastomere samples (7950 samples or322) than in day-5 trophectoderm samples [2341 samples or137 Fisherrsquos exact test ORfrac14 299 (95 CI 284ndash315) Plt 1 1010]

Evidence of frequent tripolar mitosis in cleavage-stageembryos

Among the different categories of complex aberrations digynictriploid and near-triploid patterns were relatively prevalent(1061 samples or 25 Table 2) Triploid zygotes are predisposedto form tripolar spindles (28) in which case the resulting em-bryos may be expected to be approximately or lsquoquasi-rsquo diploidwith chaotic chromosome complements This follows from amodel by which the triploid set of chromosomes (23 3frac14 69) isreplicated (69 2frac14 138) then randomly segregated to threedaughter cells (1383frac14 46 chromosomes per cell) Samplesexhibiting this pattern (see Materials and Methods) were rare inthe PGT-A dataset (106 samples or 025) potentially indicatingthat chromosome segregation in triploid cells is not generallyrandom

Meanwhile random tripolar division of normally fertilizeddipronuclear (2PN) zygotes would be predicted to generate hy-podiploid complements in the three daughter cells (923frac1431chromosomes per cell) Under this model one would expect toobserve a mixture of disomies paternal monosomies maternalmonosomies and nullisomies in the ratio of 4 2 2 1 (Fig 2A)Selecting aneuploid samples with at least one disomy paternalmonosomy maternal monosomy and nullisomy (and no co-occurring trisomy or uniparental disomy) we identified 1230(29) putative lsquodiploid tripolarrsquo samples with a mean numberof 280 6 63 (6SD) total chromosomes (modefrac14 30

chromosomes Fig 2B) Performing 10 000 simulations we con-firmed that the observed patterns of chromosome loss wereconsistent with a model of tripolar segregation of the diploidchromosome set (Fig 2C) We note that the distribution of totalchromosomes exhibits modest bimodality with a secondarypeak at 20 chromosomes (Fig 2B) This may reflect an additionalround of tripolar segregation whereby the diploid tripolar com-plement is replicated (3067 2frac14 613) then randomly dividedamong three daughter cells (6133frac14 204 chromosomes percell) This was similarly reflected in the observed counts of10 885 disomies 6398 paternal monosomies 6225 maternalmonosomies and 4451 nullisomies among the 1230 diploid tri-polar samples which exhibited a slight excess of chromosomeloss compared with the predicted 4 2 2 1 ratio (observedratiofrac14 4 24 23 16 Fig 2C) Nevertheless the strong concor-dance between the observed and simulated data indicates thatduring tripolar segregation chromosomes tend to assort ran-domly into each of three daughter cells rather than segregatingpreferentially to one or two cells Consistent with the notionthat multipolar divisions compromise development or cell pro-liferation more than 90 (1149 samples) of the 1230 putativediploid tripolar samples were from day-3 blastomere biopsieswith only a small minority (81 samples) represented among thetrophectoderm biopsies

Maternal genotype at common variants spanning PLK4is associated with incidence of the diploid tripolarsignature in blastomere samples

We recently reported that a common (30 global minor allelefrequency) haplotype spanning PLK4 tagged by the SNPrs2305957 is associated with mitotic error in human preimplan-tation embryos (34) In addition to the overall association withmitotic-origin aneuploidy we noted a particularly strong associ-ation with complex errors involving loss of both maternal andpaternal homologs (34) As the diploid tripolar signature docu-mented in our current analysis resembled this pattern wesought to test whether the original association between mitoticaneuploidy and maternal genotype at the quantitative trait

Table 2 Frequencies of various forms of complex abnormality (gt2 chromosomes with abnormal copy number) in day-3 (d3) blastomere biop-sies and day-5 (d5) trophectoderm biopsies

Type of complex abnormality No of d3 samples () No of d5 samples ()

Diploid tripolara 1149 (47) 81 (05)Digynic triploid tripolar 94 (04) 14 (01)Diandric triploid tripolar 27 (01) 0 (00)Digynic triploidnear-triploidb 740 (30) 321 (19)Diandric triploidnear-triploidb 25 (01) 32 (02)Maternal haploidnear-haploidb 288 (12) 84 (05)Paternal haploidnear-haploidb 126 (05) 25 (01)Multiple maternal chromosome gains (non-triploid) 480 (19) 248 (15)Multiple paternal chromosome gains (non-triploid) 61 (02) 40 (02)Both maternal and paternal gains (non-triploid) 390 (16) 298 (17)Multiple maternal chromosome losses (non-haploid) 360 (15) 93 (05)Multiple paternal chromosome losses (non-haploid) 186 (08) 24 (01)Both maternal and paternal losses (non-haploid) 448 (18) 44 (03)Otherc 3576 1037Total 7950 (322) 2341 (137)

aDefined by possession1 biparental disomy1 paternal monosomy1 maternal monosomy and1 nullisomybNear-triploidy and near-haploidy are defined by20 trisomic and monosomic chromosomes respectivelycUnclassified combinations of maternal and paternal chromosome gains and losses as well as uniparental disomy

locus encompassing PLK4 was driven by tripolar chromosomesegregation in diploid cells

Supporting our hypothesis the per-case frequency of diploidtripolar day-3 blastomere samples was positively associatedwith the minor (A) allele of rs2305957 [quasi-binomial GLMORfrac14 142 (95 CI 129ndash157) Pfrac14 19 1012] (Fig 3) The magni-tude of this association with tripolar mitosis thus exceeds thatof the original association with mitotic error [quasi-binomialGLM ORfrac14 124 (95 CI 118ndash131] Pfrac14 60 1015] despite com-prising only approximately one-fifth of all putative mitotic

aneuploidies in these samples (1149 vs 5438) Like the originallydescribed association (34) the effect was additive with meansof 35 47 and 75 diploid tripolar embryos per patient car-rying zero one and two copies of the risk allele respectively(Fig 3) The effect was also constant with maternal age [quasi-binomial GLM ORfrac14 099 (95 CI 098ndash101) Fig 3C] consistentwith previous results (34) Because the distributions of diploidtripolar samples per case exhibited significant inflation of zeros(AIC-corrected Vuong test Zfrac14661 Pfrac14 19 1011) we also fita hurdle model to the data to account separately for the zero

Disomy MaternalLoss

6 6 6 0

2 0 0 1

1 1 1 0

6 0 0 3x 3

Total 12 6 6 3

4 2 2 1Ratio

PaternalLoss

Nullisomy

Disomy Disomy

PaternalLoss

MaternalLoss

Disomy

One of each chromosome pairon different spindles

Both chromosomes onthe same spindle

6 possible configurations

ulatedO

bse rved

0 5 10 15 20

Maternal loss

Paternal loss

Nullisomy

Disomy

Maternal loss

Paternal loss

Nullisomy

Disomy

Count Per Sample

ulatedO

bserved

0 10 20 30 40 50

Number of Chromosomes

n = 1230

n = 10000

Figure 2 Tripolar mitosis in diploid embryos (A) Schematic showing all possible outcomes following diploid tripolar mitosis Only one pair of chromosomes is depicted

for clarity (maternal red paternal blue) Spindle attachment configurations are shown on the left with segregation outcomes in the three daughter cells shown in the

middle The table on the right provides the expected ratios of segregation outcomes for each chromosome following tripolar mitosis (B) Density histograms showing

observed chromosome counts in 1230 samples that are suspected to have undergone tripolar mitosis (see main text) compared with 10 000 samples simulated under a

model of random tripolar segregation Observed peaks at 30 and 20 chromosomes may reflect one and two rounds of tripolar mitosis respectively (C) Density histo-

grams displaying the number of chromosomes in each biopsy displaying each of the four possible outcomes Simulated data (10 000 simulations) are displayed in the

top panel while observed data (1230 putative diploid tripolar biopsies) are displayed in the bottom panel

and non-zero portions of the distribution The hurdle modelsupported the association both for the zero [binomial GLMORfrac14 141 (95 CI 135ndash147) Plt 1 1010] and non-zero counts[Poisson GLM ORfrac14 131 (95 CI 126ndash137) Plt 1 1010]

Upon excluding all putative diploid tripolar samples fromthe analysis we re-tested the remaining putative mitotic errorsfor association with maternal genotype at rs2305957 Weobserved a modest but significant association with residualmitotic aneuploidies [quasi-binomial GLM ORfrac14 112 (95 CI106ndash119) Pfrac14 16 104] potentially reflecting our initially strictclassification of diploid tripolar samples which excluded co-occurring trisomies Supporting this hypothesis significantassociations were observed upon excluding the originally de-fined set of diploid tripolar samples but relaxing the criteria toinclude hypodiploid complements (lt42 chromosomes) with co-occurring maternal trisomy [quasi-binomial GLM ORfrac14 116 (95CI 104ndash130) Pfrac14 95 103] or paternal trisomy [quasi-binomialGLM ORfrac14 129 (95 CI 109ndash153) Pfrac14 36 103] These resultssuggest that tripolar mitosis may also occur in embryos alreadyaffected by meiotic or mitotic errors or that additional aneuploi-dies may accumulate downstream of tripolar mitosis

Analysis of time-lapse data from a subset of cases

Seeking to validate our findings we took advantage of the factthat embryos from a subset of IVF cases included in the PGT-Adataset had previously undergone time-lapse imaging by the re-ferring laboratory with published data demonstrating frequentmultipolar cleavages (32) (Supplementary Material Videos S1S2 and S3) These included 77 day-3 cleavage-stage embryosfrom 10 IVF cases with single blastomeres analyzed by PGT-AWe note that time-lapse observation of direct unequal cleavage(DUC) does not prove the occurrence of tripolar mitosis as theformer refers to cleavage of the cytoplasm while the latterrefers to mitotic spindle conformation and chromosome

segregation Moreover rapid sequential cleavages with an un-usually short interval between mother and daughter cell divi-sion (5 h) are also classified as DUC but may arise by distinctmechanisms and confer distinct chromosomal outcomes Wenevertheless hypothesized that DUC and tripolar segregationare related Consistent with this hypothesis we observed thatthe diploid tripolar PGT-A signature was significantly enrichedamong embryos documented by time lapse to have undergoneone or more DUCs during the first three divisions [Fisherrsquos exacttest ORfrac14 664 (95 CI 134ndash374) Pfrac14 00087] The signature wasmost prevalent among embryos undergoing DUC during the firstdivision (DUC-1) or undergoing multiple DUCs (DUC-PlusFig 4A) indicating that these cleavage phenotypes may confer awider distribution of abnormal cells at the time of biopsy

Time-lapse data were also valuable for direct validation ofthe association with maternal PLK4 genotype For this purposedata from cases undergoing day-5 blastocyst biopsy could alsobe incorporated as the availability of time-lapse data fromthese cases does not depend on embryo survival to the blasto-cyst stage This extended the analysis to 58 IVF patients with atotal of 742 embryos having undergone time-lapse screening(32) Despite this small sample size rs2305957 was significantlyassociated with DUC [quasi-binomial GLM ORfrac14 153 (95 CI101ndash231) Pfrac14 0047] Directionality was consistent with expect-ations as the minor (A) allele which was previously associatedwith increased incidence of the diploid tripolar signature (Fig 3)was also associated with increased incidence of DUC (Fig 4B)

DiscussionAneuploidy is the leading known cause of implantation failuremiscarriage and congenital birth defects Recent studies havevastly improved our understanding of aneuploidy of maternalmeiotic origin and its association with maternal ageMeanwhile relatively little is known about errors of mitotic

Figure 3 Tripolar mitosis in diploid embryos drives association with common maternal genetic variants spanning PLK4 (A) For cases with1 diploid tripolar sample

boxplots of proportion of diploid tripolar samples stratified by maternal genotype at rs2305957 The colored boxes extend from the median to the first (Q1) and third

(Q3) quartiles Whiskers extend to 15 interquartile range Outlier data points beyond this range are plotted individually Only cases withgt2 samples are plotted for

clarity The risk allele (A allele of rs2305957) is associated with increased frequencies of biopsies exhibiting signatures of tripolar mitosis [quasi-binomial GLM

ORfrac14142 (95 CI 129 ndash 157) Pfrac1419 1012] (B) Proportions of cases per genotype with zero diploid tripolar samples Error bars indicate 6 standard errors of the mean

proportions (C) Mean proportions of putative diploid tripolar samples stratified by age (rounded to nearest two years) and genotype Ribbons indicate 6 standard errors

of the mean proportions

origin that occur after fertilization and contribute to widespreadchromosomal mosaicism Establishing the molecular mecha-nisms and risk factors contributing to mitotic aneuploidy is animportant step in improving our understanding of the basis ofvariation in human fertility

One prominent form of mitotic aneuploidy termed lsquochaoticmosaicismrsquo has been recognized since early applications of PGS(17) but its origins have proven elusive Chaotic embryos arecharacterized by severely aneuploid karyotypes that vary fromcell to cell in a pattern reminiscent of cancer cell linesIntriguingly the first description of this phenomenon byDelhanty et al noted that lsquothe occurrence of chaotically dividingembryos was strongly patient-related ie some patients hadldquochaoticrdquo embryos in repeated cycles whereas other patientswere completely free of this type of anomalyrsquo (17) This observa-tion hints at a patient-specific predisposition to the formationof chaotic embryos either due to genetic or environmental fac-tors Technical limitations of the fluorescence in situ hybridiza-tion (FISH) platform however hindered the ability of earlystudies to examine chaotic embryos in greater detail or probetheir molecular origins Here we revisited this topic equippedwith a high-resolution dataset comprising 24-chromosomePGT-A data fromgt41 000 embryos

One hypothesized source of chaotic aneuploidy is the forma-tion of multipolar mitotic spindles a phenomenon frequentlyobserved among 3PN zygotes (2728) We found that digynic trip-loidy was common relative to other forms of complex abnor-mality affecting 30 and 19 of day-3 and day-5 embryosrespectively This is consistent with previous studies showingthat digynic 3PN zygotes are relatively prevalent and occasion-ally undetected following ICSI (36) which was used in 80ndash90 ofcases in this dataset Digynic triploidy may arise either due tofailed meiotic division or failed extrusion of the second polarbody (31) While previous studies have demonstrated that trip-loid embryos are predisposed to multipolar cell division weidentified few samples (106 samples or 02) exhibiting patternssuggestive of random chromosome segregation of a replicateddigynic triploid genome This paucity of random triploid tripolarsamples agrees with previous reports that digynic 3PN ICSI-

derived embryos tend to either remain triploid or lsquoself-correctrsquoto diploidy (3738)

Even more prevalent at the cleavage stage (1162 samples or47) were samples exhibiting hypodiploid karyotypes involvingloss of both maternal and paternal homologs Based on simula-tion we showed that this pattern is consistent with a model ofrandom tripolar segregation of a normal diploid chromosomecomplement with a subset of cells potentially undergoing a sec-ond tripolar segregation event Time-lapse data from an inter-secting set of 77 embryos (32) supported our inference oftripolar chromosome segregation with the hypodiploid PGT-Asignature enriched among embryos previously recorded to haveundergone one or more DUC Despite this enrichment we notethat the time-lapse data were not perfectly predictive of the dip-loid tripolar signature This result is expected given the sam-pling noise associated with single-cell biopsy of mosaicembryos Earlier tripolar divisions would be expected confer awider distribution of aneuploid cells Somewhat counterintui-tively however even one embryo recorded as undergoing tripo-lar mitosis during the first cleavage (DUC-1) produced a euploidbiopsy Assuming the accuracy of time-lapse and PGT-A classifi-cation this indicates that as an alternative to random segrega-tion embryos undergoing tripolar mitosis may occasionallysegregate normal diploid complement to at least one daughtercell Interestingly such a phenomenon was recently docu-mented in a mosaic bovine embryo composed of separate an-drogenetic gynogenetic and normal diploid cell lines (39) Theauthors proposed that this pattern may have arisen via the for-mation of a tripolar gonomeric spindle with separate microtu-bules associated with maternal and paternal genomes (39)Alternatively anucleate fragments may be extruded duringsome normal bipolar mitoses potentially misclassified as tripo-lar cleavages

We found that diploid tripolar samples were rare at theday-5 blastocyst stage (80 samples or 04) We interpreted thisobservation to suggest that cleavage-stage embryos affected bytripolar segregation experience reduced viability andor that an-euploid cell lineages are purged or fail to propagate and thuscontribute fewer descendant cells to the blastocyst cell popula-tion (40) We note however that the interpretation of day-5

PGS Result

Other aneuploid

Euploid

Diploid tripolar

Time-Lapse Result

GG AG AA

Genotype

Figure 4 Time-lapse data from a subset of embryos support the PGT-A-based hypothesis of tripolar mitosis and association with PLK4 genotype (A) Samples recorded

to have undergone tripolar mitosis during the first (DUC-1 Supplementary Material Video S1) second (DUC-2 Supplementary Material Video S2) third (DUC-3

Supplementary Material Video S3) or multiple cleavage divisions are enriched for the diploid tripolar PGT-A signature compared with embryos that underwent normal

cleavage (non-DUC) [Fisherrsquos exact test ORfrac14664 (95 CI 134ndash374) Pfrac1400087] (B) Genotype data from 58 patients whose embryos underwent time-lapse imaging as

well as PGT-A analysis of either day-3 or day-5 embryo biopsies support a positive association between the minor allele of rs2305957 and incidence of tripolar mitosis

[quasi-binomial GLM ORfrac14 153 (95 CI 101ndash231) Pfrac140047]

trophectoderm samples could be confounded by the multi-cellnature of the biopsymdashan important limitation of these PGT-Adata In the case of a mosaic biopsy (which is one expected out-come of tripolar segregation) PGT-A signals from karyotypicallydistinct cells may be averaged and obscure expected patterns Ifsuch biopsies no longer meet the multiple chromosome loss cri-teria we used to distinguish diploid tripolar samples (seeMaterials and Methods section) their misclassification couldlead us to underestimate the frequency of the tripolar signatureamong day-5 blastocysts In rare cases reciprocal gains andlosses within multi-cell biopsies may average to resemble diso-mic signatures as previously proposed as a potential source offalse negative calls (41) Nevertheless our interpretation of em-bryonic or cellular inviability is consistent with multiple previ-ous studies demonstrating that embryos with complexaneuploidies including those deriving from normally fertilizedzygotes tend to arrest at cleavage stages around the time ofembryonic genome activation (842ndash44) Moreover previousstudies have observed a strong enrichment of abnormal mitoticspindles among arrested cleavage-stage embryos (18) as well asdemonstrating that zygotes undergoing tripolar mitosishave poor developmental potential beyond the cleavage stage(29ndash32) Finally the hypothesis that tripolar segregation impairsembryonic development was recently supported by single-cellkaryomapping and time-lapse imaging of 13 arrested cleavage-stage embryos which exhibited extensive chromosome lossconsistent with tripolar mitosis (33) While there is also evi-dence of tripolar spindle formation in confocal images of hu-man blastocysts it is possible that aneuploid daughter cells areeliminated by apoptosis or that the spindle configurations aretransient and do not tend to result in tripolar anaphase at thisstage Potentially relevant are previous observations that unlikecleavage-stage embryos adult somatic cells possessing extracentrosomes rarely undergo multipolar mitosis but instead ex-perience centrosome clustering leading to the formation of apseudo-bipolar spindle (45) This may reflect increased strin-gency of mitotic checkpoints following embryonic genome acti-vation possibly via upregulation of MAD2 (46) While clusteredcentrosomes predispose the cell to aberrant microtubule-kinetochore attachments and anaphase lag (47) the resultantaneuploidies are less severe than those induced by multipolarmitosis

We recently reported that common maternal genetic variantsdefining a 600 kb haplotype of chromosome 4 tagged by SNPrs2305957 is strongly associated with complex mitotic-origin an-euploidy in day-3 cleavage-stage embryos (34) Furthermoreindividuals carrying the risk allele had fewer blastocyst-stageembryos available for testing at day 5 suggesting that the com-plex aneuploid phenotype impairs embryonic survival The asso-ciation we observed was significant with maternal but notpaternal genotypes reflecting the fact that prior to embryonicgenome activation mitotic divisions are controlled by maternalgene products deposited in the oocyte (34) While the associatedhaplotype spans seven genes (INTU SLC25A31 HSPA4L PLK4MFSD8 LARP1B and PGRMC2) PLK4 stands out given itswell-characterized role as an essential regulator of centrosomeduplication PLK4 is a tightly regulated kinase that initiatesassembly of a daughter procentriole at the base of the existingmother centriole thereby mediating bipolar spindle forma-tion (4849) Altered expression of PLK4 is associated with centro-some amplification generating multipolar spindles in humancell linesmdasha hallmark of several cancers (48ndash52) Moreoverrecent work on bovine embryos indicates that PLK4 may consti-tute a maternally contributed factor with a crucial role in oocyte

maturation and organization of the initial mitotic spindle (53)Notably PLK4 knockdown with interfering double-stranded RNA(dsRNA) induced mitotic spindle defects abnormal cytokinesisand early developmental arrest (53)

Based on this knowledge we hypothesized that the risk al-lele of the PLK4-linked QTL identified in our previous study (34)may predispose cells of cleavage-stage embryos to tripolar mi-totic divisions Such divisions would produce daughter cellspossessing 4 2 2 1 ratios of disomies paternal and maternalmonosomies and nullisomies respectively Consistent withthis hypothesis the diploid tripolar signature was strongly as-sociated with maternal genotype at rs2305957mdasha finding thatwe further validated using time-lapse data from 742 embryosand corresponding genotype data from 58 patients Togetherour results support a causal role of PLK4 in the originalassociation

Given the necessarily correlational nature of our analysiswe cannot formally rule out alternative and non-mutually ex-clusive models by which the PLK4-spanning QTL could impactmitotic phenotypes Indeed PLK4 plays multiple roles withinand apart from the centrosome cycle and its variation couldthus influence cellular cleavage andor chromosome segrega-tion through alternative routes For example in Drosophila Plk4interacts with Cdc6 a replication licensing factor which bothlocalize to centrosomes Cdc6 inhibits centrosome duplicationduring S-phase DNA replication while Plk4 counteracts this ac-tivity (54) If the same interaction occurs in human embryosPLK4 activity could theoretically influence the rate of DUC bydecoupling the S- and M-phases of early embryonic divisionsOther studies of PLK4 in model organisms point to additionalroles that are independent of centrosome duplication (55ndash57)These include a study demonstrating that depletion of mater-nally contributed Plk4 prevents microtubule nucleation andleads to the formation of monopolar mitotic spindles in acen-triolar mouse embryos (57) Nevertheless we believe that thetripolar mitosis model provides the most complete and parsi-monious explanation for the data as these alternative modelsare not expected to produce chromosome complements closelymatching tripolar expectations (disomies paternal monoso-mies maternal monosomies and nullisomies in the ratio of 4 22 1 Fig 2)

Zhang et al (58) recently replicated a key finding of our initialassociation study demonstrating that rates of blastocyst forma-tion are reduced among embryos from patients carrying thehigh-risk genotypes of rs2305957 Furthermore patients diag-nosed with early recurrent miscarriage (before 12 weeks of ges-tation) were found to possess a higher frequency of the riskallele compared with matched fertile control subjects (58) Ouranalysis provides evidence of the molecular mechanism under-lying these results suggesting that complex aneuploidies aris-ing from tripolar mitosis compromise early embryonicdevelopment Along with methodological considerations theseverity of the complex aneuploid phenotype may explain thelack of association with recurrent pregnancy losses that occurlater in development and are presumably associated with lesssevere aneuploidies (859) Further research will be necessary todetermine the relevance of these findings to non-IVF patientpopulations and in vivo conception

Through detailed characterization of complex chromosomeabnormalities our analysis revealed that tripolar chromosomesegregation in embryos originating from normally fertilized 2PNzygotes is an under-recognized phenomenon contributing toaneuploidy in cleavage-stage embryos Using simulation weshowed that observed chromosomal patterns are consistent

with a model of random segregation of a replicated diploid com-plement to each of three daughter cells Finally we demon-strated that the signatures of tripolar chromosome segregationare significantly correlated with maternal genetic variants span-ning the centrosomal regulator PLK4 This implicates maternalvariation in PLK4 as a factor influencing mitotic spindle integritywhile shedding light on tripolar mitosis as a potentially impor-tant mechanism contributing to aneuploidy in preimplantationembryos Together our results help illuminate the molecularorigins of chaotic aneuploidy a long-recognized phenomenoncontributing to high rates of IVF failure

Materials and MethodsHuman subjects approvals

Based on the retrospective nature of the analysis and use of de-identified data this work was determined not to constitute hu-man subjects research by the University of Washington HumanSubjects Division (affiliation of RCM at the time of analysis) aswell as Ethical amp Independent Review Services who providedtheir determination to Natera Inc

Sample preparation genotyping and aneuploidydetection

DNA isolation whole genome amplification and SNP genotypingare described in detail in McCoy et al (8) Briefly genetic mate-rial was obtained from IVF patients or oocyte donors and malepartners by buccal swab or peripheral venipuncture Geneticmaterial was also obtained from single blastomere biopsies ofday-3 cleavage-stage embryos or 5ndash10 cell biopsies of trophec-toderm tissue from day-5 blastocysts Embryo DNA was ampli-fied via multiple displacement amplification [MDA see (8)]Amplified embryo DNA and bulk parental tissue were geno-typed on the HumanCytoSNP-12 BeadChip (Illumina San DiegoCA) using the standard Infinium II protocol Genotype callingwas performed using the GenomeStudio software package(Illumina San Diego CA)

Aneuploidy detection was performed using the ParentalSupport algorithm which leverages informative parentalmarkers (eg sites where one parent is homozygous and theother parent is heterozygous) to infer transmission of maternaland paternal homologs along with the copy number of all chro-mosomes across the embryo genome In doing so the methodovercomes high rates of allelic dropout that characterize dataobtained from whole genome amplified DNA of embryo biop-sies The Parental Support method is described in detail inJohnson et al (9) who also demonstrated its sensitivity and spe-cificity of 979 and 961 respectively

Classification criteria

Putative diploid tripolar samples were identified as those exhib-iting1 maternal monosomy1 paternal monosomy1nullisomy and1 biparental disomy with no co-occurring triso-mies or uniparental disomies Digynic and diandric triploid andnear-triploid samples were defined as those with20 maternaltrisomic or paternal trisomic chromosomes respectivelyMaternal haploid and paternal haploid or near-haploid sampleswere similarly defined as those with20 paternal monosomicor maternal monosomic chromosomes respectively Digynictriploid tripolar and diandric triploid tripolar samples were de-fined using simulation-based expectations of random

chromosome segregation of triploid cells Specifically digynictriploid tripolar samples were required to exhibit3 disomic3maternal trisomic 1ndash8 maternal uniparental disomic and 1ndash8paternal monosomic chromosomes Diandric triploid tripolarsamples were meanwhile required to exhibit3 disomic3 pa-ternal trisomic 1ndash8 paternal uniparental disomic and 1ndash8 mater-nal monosomic chromosomes

Time-lapse imaging and annotation

Time-lapse microscope image capture annotation and classifi-cation of tripolar mitosis (therein referred to as lsquodirect unequalcleavagersquo or lsquoDUCrsquo) is described in the Materials and Methodssection of Zhan et al (32) In short images were automaticallycaptured every 10 min with seven focal planes illuminated by635 nm LED light Several time points were annotated includ-ing appearance of pronuclei syngamy time of division morulacavitation early blastocyst expanded blastocyst and hatchingblastocyst Only cells with visible nuclei were considered blasto-meres DUC was defined as (1) cleavage of a single blastomereinto 3thorndaughter blastomeres or (2) unusually short interval be-tween mother and daughter cell division of5 h in accordancewith established criteria (6061) DUC-1 (first cleavageSupplementary Material Video S1) DUC-2 (second cleavageSupplementary Material Video S2) DUC-3 (third cleavageSupplementary Material Video S3) and DUC-Plus (multipleDUCs) were annotated based on the cleavage stage duringwhich DUC was observed

Statistical analyses

Statistical analyses were performed using the R statistical com-puting environment (62) with plots generated using thelsquoggplot2rsquo package (63) Density plots (Fig 2A and B) were gener-ated using the ggjoy package (httpscranr-projectorgpackagefrac14ggjoy date last accessed April 27 2018)

Simulations of tripolar mitosis were performed in R by usingthe lsquosamplersquo function (without replacement) to distribute repli-cated maternal and paternal chromatids to each of three daugh-ter cells To implement the simulation we set the probability ofa given daughter cell inheriting a given (maternal or paternal)chromatid to 23 thereby capturing the condition that sisterchromatids segregate to two of three daughter cells during tri-polar mitosis (Fig 2A) After repeating this procedure for all 46chromosomes counts of disomies maternal and paternalmonosomies and nullisomies were tabulated along with totalchromosome count The simulation was repeated 10 000 timesto generate frequency distributions of various classes of abnor-malities to which the empirical data were compared

To test the association between maternal genotype and inci-dence of embryos with the diploid tripolar PGT-A signature welimited the analysis to the originally tested set of unrelated IVFpatients or egg donors (no repeat cases less than second degreerelatedness) with genotype data meeting quality control thresh-olds (95 variant call rate 95 sample genotyping efficiency)We then used a generalized linear regression model to test forassociation between the per-case counts of samples that didand did not exhibit the diploid tripolar pattern and maternalgenotypes at rs2305957 encoded as dosage of the lsquoArsquo allele Toaccount for excess zeros we also fit a hurdle model to the zeroand non-zero portions of the distribution using the lsquopsclrsquo pack-age (64) For the corresponding analysis of time-lapse pheno-types we sought to maximize statistical power from a small

number of cases by combining rather than excluding embryodata from repeat IVF cases To this end we used KING (65) toinfer cases derived from the same patient based on maternalgenotype data We then summed tripolar and non-tripolar em-bryo phenotypes for each independent patient using the result-ing data in a quasi-binomial regression as above

Aneuploidy data were made available with the original pub-lication (34) posted as Supplementary Material Analysis scriptsare posted on GitHub httpsgithubcomrmccoy7541tripolar_mitosis

Supplementary MaterialSupplementary Material is available at HMG online

Acknowledgements

The authors thank Allison Ryan Milena Banjevic Matthew Hilland Matthew Rabinowitz for their work at Natera to developalgorithms to detect aneuploidies in PGS data We also thankMolly Gasperini and members of the Akey and Shendure labsfor helpful discussions on this project This work was supportedin part by NIHNHGRI Genome Training Grant (5T32HG000035-22) to the Department of Genome Sciences at the University ofWashington

Conflict of Interest statement DAP has received stock options inNatera Inc as consulting fees ZPD and SS are employees ofand hold stock or options to hold stock in Natera Inc RCM andDAP are co-inventors on a patent application filed by StanfordUniversity with the US Patent and Trademark Office onNovember 11 2015 (US 14938 842) RCM has received pastconference travel support from Natera Inc ERH receives fund-ing from Illumina Inc

References1 Fragouli E Alfarawati S Spath K Jaroudi S Sarasa J

Enciso M and Wells D (2013) The origin and impact of em-bryonic aneuploidy Hum Genet 132 1001ndash1013

2 Hassold T Hall H and Hunt P (2007) The origin of humananeuploidy where we have been where we are going Hum

Mol Genet 16 R203ndashR2083 Handyside AH Montag M Magli MC Repping S

Harper J Schmutzler A Vesela K Gianaroli L andGeraedts J (2012) Multiple meiotic errors caused by predivi-sion of chromatids in women of advanced maternal age un-dergoing in vitro fertilisation Eur J Hum Genet 20 742ndash747

4 Vanneste E Voet T Le Caignec C Ampe M Konings PMelotte C Debrock S Amyere M Vikkula M and SchuitF (2009) Chromosome instability is common in humancleavage-stage embryos Nat Med 15 577ndash583

5 Handyside AH (2013) 24-chromosome copy number analy-

sis a comparison of available technologies Fertil Steril 100595ndash602

6 Zegers-Hochschild F Adamson GD Dyer S RacowskyC de Mouzon J Sokol R Rienzi L Sunde A Schmidt LCooke ID et al (2017) The international glossary on infertil-ity and fertility care Fertil Steril 108 393ndash406

7 Donate A Estop AM Giraldo J and Templado C (2016)Paternal age and numerical chromosome abnormalities inhuman spermatozoa Cytogenet Genome Res 148 241ndash248

8 McCoy RC Demko ZP Ryan A Banjevic M Hill MSigurjonsson S Rabinowitz M and Petrov DA (2015)Evidence of selection against complex mitotic-origin aneu-ploidy during preimplantation development PLoS Genet 11e1005601

9 Johnson DS Gemelos G Baner J Ryan A Cinnioglu CBanjevic M Ross R Alper M Barrett B Frederick J et al(2010) Preclinical validation of a microarray method for fullmolecular karyotyping of blastomeres in a 24-h protocolHum Reprod 25 1066ndash1075

10 Rabinowitz M Ryan A Gemelos G Hill M Baner JCinnioglu C Banjevic M Potter D Petrov DA andDemko Z (2012) Origins and rates of aneuploidy in humanblastomeres Fertil Steril 97 395ndash401

11 Handyside AH Harton GL Mariani B Thornhill ARAffara N Shaw MA and Griffin DK (2010) Karyomappinga universal method for genome wide analysis of genetic dis-ease based on mapping crossovers between parental haplo-types J Med Genet 47 651ndash658

12 Ottolini CS Newnham LJ Capalbo A Natesan SAJoshi HA Cimadomo D Griffin DK Sage K SummersMC Thornhill AR et al (2015) Genome-wide maps of re-combination and chromosome segregation in humanoocytes and embryos show selection for maternal recombi-nation rates Nat Genet 47 727ndash735

13 Natesan SA Bladon AJ Coskun S Qubbaj W Prates RMunne S Coonen E Dreesen JC Stevens SJ PaulussenAD et al (2014) Genome-wide karyomapping accuratelyidentifies the inheritance of single-gene defects in humanpreimplantation embryos in vitro Genet Med 16 838ndash845

14 Kumar A Ryan A Kitzman JO Wemmer N SnyderMW Sigurjonsson S Lee C Banjevic M Zarutskie PWLewis AP et al (2015) Whole genome prediction for preim-plantation genetic diagnosis Genome Med 7 35

15 Hou Y Fan W Yan L Li R Lian Y Huang J Li J Xu LTang F Xie XS et al (2013) Genome analyses of single hu-man oocytes Cell 155 1492ndash1506

16 Mantikou E Wong KM Repping S and Mastenbroek S(2012) Molecular origin of mitotic aneuploidies in preimplan-tation embryos Biochim Biophys Acta 1822 1921ndash1930

17 Delhanty JD Harper JC Ao A Handyside AH andWinston RM (1997) Multicolour FISH detects frequent chro-mosomal mosaicism and chaotic division in normal preim-plantation embryos from fertile patients Hum Genet 99755ndash760

18 Chatzimeletiou K Morrison EE Prapas N Prapas Y andHandyside AH (2005) Spindle abnormalities in normallydeveloping and arrested human preimplantation embryosin vitro identified by confocal laser scanning microscopyHum Reprod 20 672ndash682

19 Boveri T (1900) Zellen-Studien Heft 4 Ueber die natur der cen-trosomen Verlag Von Gustav Fischer Stuttgart Germany

20 Balczon R Bao L Zimmer WE Brown K Zinkowski RPand Brinkley BR (1995) Dissociation of centrosome replica-tion events from cycles of DNA synthesis and mitotic divi-sion in hydroxyurea-arrested Chinese hamster ovary cellsJ Cell Biol 130 105ndash115

21 Schatten H and Sun QY (2011) New insights into the roleof centrosomes in mammalian fertilization and implicationsfor ART Reproduction 142 793ndash801

22 Palermo G Munne S and Cohen J (1994) The human zy-gote inherits its mitotic potential from the male gameteHum Reprod 9 1220ndash1225

23 Simerly C Wu GJ Zoran S Ord T Rawlins R Jones JNavara C Gerrity M Rinehart J Binor Z et al (1995) Thepaternal inheritance of the centrosome the cellrsquosmicrotubule-organizing center in humans and the implica-tions for infertility Nat Med 1 47ndash52

24 Sathananthan AH Ratnam SS Ng SC Tarin JJGianaroli L and Trounson A (1996) The sperm centriole itsinheritance replication and perpetuation in early humanembryos Hum Reprod 11 345ndash356

25 Holubcova Z Blayney M Elder K and Schuh M (2015)Error-prone chromosome-mediated spindle assembly favorschromosome segregation defects in human oocytes Science348 1143ndash1147

26 Sedo CA Schatten H Combelles CM and Rawe VY(2011) The nuclear mitotic apparatus (NuMA) protein locali-zation and dynamics in human oocytes fertilization andearly embryos Mol Hum Reprod 17 392ndash398

27 Plachot M Mandelbaum J Junca AM De Grouchy JSalat-Baroux J and Cohen J (1989) Cytogenetic analysisand developmental capacity of normal and abnormal em-bryos after IVF Hum Reprod 4 99ndash103

28 Kola I Trounson A Dawson G and Rogers P (1987)Tripronuclear human oocytes altered cleavage patterns andsubsequent karyotypic analysis of embryos Biol Reprod 37395ndash401

29 Wirka KA Chen AA Conaghan J Ivani K GvakhariaM Behr B Suraj V Tan L and Shen S (2014) Atypical em-bryo phenotypes identified by time-lapse microscopy highprevalence and association with embryo developmentFertil Steril 101 1637ndash1648

30 Hlinka D Kalatova B Uhrinova I Dolinska S RutarovaJ Rezacova J Lazarovska S and Dudas M (2012)Time-lapse cleavage rating predicts human embryo viabil-ity Physiol Res 61 513

31 Kalatova B Jesenska R Hlinka D and Dudas M (2015)Tripolar mitosis in human cells and embryos occurrencepathophysiology and medical implications Acta Histochem117 111ndash125

32 Zhan Q Ye Z Clarke R Rosenwaks Z and Zaninovic N(2016) Direct unequal cleavages embryo developmentalcompetence genetic constitution and clinical outcome PLoSOne 11 e0166398

33 Ottolini C Kitchen J Xanthopoulou L Gordon TSummers MC and Handyside AH (2017) Tripolar mitosisand partitioning of the genome arrests human preimplanta-tion development in vitro Sci Rep 7 9744

34 McCoy RC Demko Z Ryan A Banjevic M Hill MSigurjonsson S Rabinowitz M Fraser HB and Petrov DA(2015) Common variants spanning PLK4 are associated withmitotic-origin aneuploidy in human embryos Science 348235ndash238

35 McCoy RC (2017) Mosaicism in preimplantation humanembryos when chromosomal abnormalities are the normTrends Genet 33 448ndash463

36 Staessen C and Van Steirteghem AC (1997) The chromo-somal constitution of embryos developing from abnormallyfertilized oocytes after intracytoplasmic sperm injection andconventional in-vitro fertilization Hum Reprod 12 321ndash327

37 Grau N Escrich L Martın J Rubio C Pellicer A andEscriba MJ (2011) Self-correction in tripronucleated humanembryos Fertil Steril 96 951ndash956

38 Grau N Escrich L Galiana Y Meseguer M Garcıa-Herrero S Remohı J and Escriba MJ (2015)Morphokinetics as a predictor of self-correction to diploidy

in tripronucleated intracytoplasmic sperm injectionndashderived human embryos Fertil Steril 104 728ndash735

39 Destouni A Esteki MZ Catteeuw M Tsuiko ODimitriadou E Smits K Kurg A Salumets A Van SoomA Voet T et al (2016) Zygotes segregate entire parentalgenomes in distinct blastomere lineages causingcleavage-stage chimerism and mixoploidy Genome Res 26567ndash578

40 Bolton H Graham SJ Van der Aa N Kumar P TheunisK Gallardo EF Voet T and Zernicka-Goetz M (2016)Mouse model of chromosome mosaicism revealslineage-specific depletion of aneuploid cells and normal de-velopmental potential Nat Comm 7 11165

41 Treff NR and Franasiak JM (2017) Detection of segmentalaneuploidy and mosaicism in the human preimplantationembryo technical considerations and limitations FertilSteril 107 27ndash31

42 Sandalinas M Sadowy S Alikani M Calderon G CohenJ and Munne S (2001) Developmental ability of chromo-somally abnormal human embryos to develop to the blasto-cyst stage Hum Reprod 16 1954ndash1958

43 Ruangvutilert P Delhanty JD Serhal P Simopoulou MRodeck CH and Harper JC (2000) FISH analysis on day 5post-insemination of human arrested and blastocyst stageembryos Prenat Diagn 20 552ndash560

44 Rubio C Rodrigo L Mercader A Mateu E Buendıa PPehlivan T Viloria T De los Santos MJ Simon CRemohı J and Pellicer A (2007) Impact of chromosomal ab-normalities on preimplantation embryo developmentPrenat Diagn 27 748ndash756

45 Quintyne NJ Reing JE Hoffelder DR Gollin SM andSaunders WS (2005) Spindle multipolarity is prevented bycentrosomal clustering Science 307 127ndash129

46 Kramer A Maier B and Bartek J (2011) Centrosome clus-tering and chromosomal (in) stability a matter of life anddeath Mol Oncol 5 324ndash335

47 Ganem NJ Godinho SA and Pellman D (2009) A mecha-nism linking extra centrosomes to chromosomal instabilityNature 460 278ndash282

48 Habedanck R Stierhof YD Wilkinson CJ and Nigg EA(2005) The Polo kinase Plk4 functions in centriole duplica-tion Nat Cell Biol 7 1140ndash1146

49 Bettencourt-Dias M Rodrigues-Martins A Carpenter LRiparbelli M Lehmann L Gatt MK Carmo N Balloux FCallaini G and Glover DM (2005) SAKPLK4 is required forcentriole duplication and flagella development Curr Biol15 2199ndash2207

50 Ko MA Rosario CO Hudson JW Kulkarni S Pollett ADennis JW and Swallow CJ (2005) Plk4 haploinsufficiencycauses mitotic infidelity and carcinogenesis Nat Genet 37883ndash888

51 Duensing A Liu Y Perdreau SA Kleylein-Sohn J NiggEA and Duensing S (2007) Centriole overduplicationthrough the concurrent formation of multiple daughter cen-trioles at single maternal templates Oncogene 266280ndash6288

52 Rosario CO Ko MA Haffani YZ Gladdy RA PaderovaJ Pollett A Squire JA Dennis JW and Swallow CJ(2010) Plk4 is required for cytokinesis and maintenance ofchromosomal stability Proc Natl Acad Sci USA 1076888ndash6893

53 Liang S Zhao MH Guo J Choi JW Kim NH and CuiXS (2016) Polo-like kinase 4 regulates spindle and actin

assembly in meiosis and influence of early embryonic devel-opment in bovine oocytes Theriogenology 85 754ndash761

54 Xu X Huang S Zhang B Huang F Chi W Fu J WangG Li S Jiang Q and Zhang C (2017) DNA replication li-censing factor Cdc6 and Plk4 kinase antagonistically regu-late centrosome duplication via Sas-6 Nat Commun 815164

55 Galletta BJ Fagerstrom CJ Schoborg TA McLamarrahTA Ryniawec JM Buster DW Slep KC Rogers GC andRusan NM (2016) A centrosome interactome provides in-sight into organelle assembly and reveals a non-duplicationrole for Plk4 Nat Commun 7 12476

56 Aydogan MG Wainman A Saurya S Steinacker TLCaballe A Novak ZA Baumbach J Muschalik N andRaff JW (2018) A homeostatic clock sets daughter centriolesize in flies J Cell Biol 217 1233ndash1248

57 Coelho PA Bury L Sharif B Riparbelli MG Fu JCallaini G Glover DM and Zernicka-Goetz M (2013)Spindle formation in the mouse embryo requires Plk4 in theabsence of centrioles Dev Cell 27 586ndash597

58 Zhang Q Li G Zhang L Sun X Zhang D Lu J Ma JYan J and Chen ZJ (2017) Maternal common variantrs2305957 spanning PLK4 is associated with blastocyst

formation and early recurrent miscarriage Fertil Steril 1071034ndash1040

59 Sharif FA and Ashour M (2015) The single nucleotide poly-morphism rs2305957 GA is not associated with recurrentpregnancy loss Int J Res Med Sci 3 3123ndash3125

60 Meseguer M Herrero J Tejera A Hilligsoslashe KM RamsingNB and Remohı J (2011) The use of morphokinetics as apredictor of embryo implantation Hum Reprod 262658ndash2671

61 Rubio I Kuhlmann R Agerholm I Kirk J Herrero JEscriba MJ Bellver J and Meseguer M (2012) Limited im-plantation success of direct-cleaved human zygotes atime-lapse study Fertil Steril 98 1458ndash1463

62 R Development Core Team (2013) R A Language andEnvironment for Statistical Computing R Foundation forStatistical Computing Vienna Austria

63 Wickham H (2016) ggplot2 Elegant Graphics for Data AnalysisSpringer New York NY

64 Zeileis A Kleiber C and Jackman S (2008) Regressionmodels for count data in R J Stat Softw 27 1ndash25

65 Manichaikul A Mychaleckyj JC Rich SS Daly K Sale Mand Chen WM (2010) Robust relationship inference ingenome-wide association studies Bioinformatics 26 2867ndash2873

ddy147-TF1

ddy147-TF2

ddy147-TF3

ddy147-TF4

ddy147-TF5

ddy147-TF6

ddy147-TF7

ddy147-TF8

S-phase Both models are consistent with time-lapse data from an intersecting set of 77 cleavage-stage embryos which wereenriched for the tripolar signature among embryos exhibiting abnormal cleavage The tripolar signature was strongly associ-ated with common maternal genetic variants spanning the centrosomal regulator PLK4 driving the association we previouslyreported with overall mitotic errors Our findings are consistent with the known capacity of PLK4 to induce tripolar mitosis orprecocious M-phase upon dysregulation Together our data support tripolar chromosome segregation as a key mechanismgenerating complex aneuploidy in cleavage-stage embryos and implicate maternal genotype at a quantitative trait locusspanning PLK4 as a factor influencing its occurrence

IntroductionAneuploidy is common in human preimplantation embryosand increases in frequency with maternal age (1) While manyaneuploidies originate in meiosismdashprimarily in females (oogen-esis) rather than males (spermatogenesis) (23)mdasha substantialproportion arise after fertilization (postzygotic) due to errantchromosome segregation during early cleavage divisions (4)These mitotic errors produce mosaic embryos with two or morecell lineages possessing distinct chromosomal complementsBecause aneuploidy is associated with negative pregnancy out-comes many patients undergoing in vitro fertilization (IVF)treatment for infertility have their embryos tested by copy num-ber analysis of all 24 chromosomes Testing is applied to singleor small numbers of biopsied cells with the aim of transferringonly embryos that are euploidmdashan approach previously knownas preimplantation genetic screening (PGS) or preimplantationgenetic diagnosis for aneuploidy (PGD-A) but now termed pre-implantation genetic testing for aneuploidy (PGT-A) (56)

Despite substantial improvements in detecting aneuploidyat the single cell level distinguishing aneuploidies of meioticand mitotic origin remains challenging because the chromo-somal signatures can be similar In some cases discriminationmay be achieved by incorporating parental genotype informa-tion to determine the parental origin of each embryonic chro-mosome Specifically observation of both maternal (or rarelypaternal) haplotypes transmitted in a homologous region of anembryonic chromosomemdasha signature we term lsquoboth parentalhomologsrsquo or lsquoBPHrsquomdashprovides strong evidence of the meiotic ori-gin of a trisomy (Fig 1A) Meanwhile because aneuploidy is rarein sperm (1ndash4) (7) and paternal BPH affects only 1 of preim-plantation embryos (8) aneuploidies involving gain or loss ofpaternal homologs are predominantly mitotic errors To dateseveral methods have been developed to infer the transmissionof individual parental homologs based on single-nucleotidepolymorphism (SNP) microarray data facilitating classificationof meiotic and mitotic aneuploidies These include ParentalSupport (910) and karyomapping (1112) which have recentlybeen validated for linkage-based preimplantation genetic diag-nosis (PGD) of single gene disorders (1314)

These technical improvements in PGT-A methodologieshave provided substantial insights into the mechanisms of an-euploidy formation An adapted form of karyomapping for ex-ample recently revealed a novel mechanism of meiotic errortermed lsquoreverse segregationrsquo whereby sister chromatids sepa-rate at meiosis I followed by (occasionally errant) segregation ofnon-sister chromatids at meiosis II (12) This added to prema-ture separation of sister chromatids (PSSC) and meiotic non-disjunction as the predominant known mechanisms of meioticerror (1215) Mitotic aneuploidies meanwhile have tradition-ally been attributed to mitotic non-disjunction whereby chro-matids fail to separate as well as anaphase lag wherebychromatids are lost after delayed migration toward the spindlepoles (16)

A distinct class of severe mitotic aneuploidy the origins ofwhich have remained poorly understood was previously re-ferred to as lsquochaotic mosaicismrsquo due to its seemingly randomchromosomal constitution (17) Though prevalent at the cleav-age stage chaotic mosaic embryos rarely achieve blastocyst for-mation (18) One hypothesized source of chaotic aneuploidy isthe formation of multipolar mitotic spindles which may causethe cell to cleave into three or more daughter cells Recognizedsince the foundational work of Boveri in the late ninteenth cen-tury (19) multipolar spindles are in turn thought to form due toabnormalities of the centrosome As the dynamic center of so-matic cells the centrosome is the key organelle coordinatingmicrotubule organization cell cycle progression and chromo-some segregation Supernumerary (gt2) centrosomes can lead tothe formation of a multipolar rather than normal bipolar spin-dle (20) If proceeding through anaphase chromosomes maythen segregate among multiple daughter cells

During gametogenesis sperm and egg cells undergo recipro-cal reductions in centrosome components (21) Upon fertiliza-tion these complementary elements are reunited and initiateformation of the sperm aster which then enlarges and maturesto organize the first mitotic spindle (21) In humans and mostother mammals (with the notable exception of rodents) thesperm is the dominant contributor of centrosome componentsincluding the proximal centriole (22ndash24) Although humanoocytes do not contain centrosomes or microtubule organizingcenters during their meiotic divisions (25) key centrosomal pro-teins such as c-tubulin and NuMA are recruited from theooplasm after the switch to mitotic division (26) Given thismode of inheritance excess centrosomes and tripolar mitosisare sometimes caused by abnormal fertilization This includesfertilization with two sperm or retention of the second polarbody producing diandric or digynic tripronuclear (3PN) zygotesrespectively (2728) Surprisingly however tripolar spindleshave also been observed by laser confocal microscopy both atthe cleavage and blastocyst stages in embryos derived from nor-mally fertilized dipronuclear (2PN) zygotes (18) albeit at a lowerrate than for 3PN zygotes Indeed tripolar mitosis [also calledlsquotrichotomic mitosisrsquo or lsquodirect unequal cleavage (DUC)rsquo] has re-cently been documented in a substantial proportion (17ndash26) ofcleaving embryos based on time-lapse imaging (29ndash32) evenafter intracytoplasmic sperm injection (ICSI) which virtuallyeliminates polyspermic fertilization Rather than abnormal cen-trosome transmission these observations suggest possible dys-regulation of centrosome duplication

Ottolini et al (33) recently detailed the phenomenon of tripo-lar mitosis in diploid cells using time-lapse analysis and PGT-Aof 51 normally fertilized embryos obtained from three IVFpatients In addition to multi-cell biopsies this study includedkaryomapping of 74 single cells disaggregated from 13 arrestedcleavage-stage embryos as well as 9 cells extruded from a totalof 26 developing embryos around the time of blastocyst forma-tion (33) PGT-A data revealed extensive evidence of chaotic

Euploid

Aneuploid

Diploid Tripolar

Low Confidence Call

MaternalPaternal

XXXXXYXY

XXY (Mat)

XXY (Pat)XXY (Pat)

XXY (Pat)XXX (Pat)

X0 (Mat) X0 (Mat)

X0 (Mat)

X0 (Pat)0Y

X0 (Pat)X0 (Pat)

XXXX00XY (Pat)

X0 (Mat)XX

Chromosome Ploidy

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Sex

Single cell biopsy

biopsy(n = 17053)

Normalgametogenesis

Error in MI

Error in MII

Mitotic error

Mitosis

trophectoderm(TE)

971ndash

Disomy MaternalLoss

6 6 6 0

2 0 0 1

1 1 1 0

6 0 0 3x 3

Total 12 6 6 3

4 2 2 1Ratio

PaternalLoss

Nullisomy

Disomy Disomy

PaternalLoss

MaternalLoss

Disomy

ulatedO

bse rved

0 5 10 15 20

Maternal loss

Paternal loss

Nullisomy

Disomy

Maternal loss

Paternal loss

Nullisomy

Disomy

Count Per Sample

ulatedO

bserved

0 10 20 30 40 50

n = 1230

n = 10000

PGS Result

Other aneuploid

Euploid

Diploid tripolar

Time-Lapse Result

GG AG AA

Genotype

Acknowledgements

ddy147-TF1

ddy147-TF2

ddy147-TF3

ddy147-TF4

ddy147-TF5

ddy147-TF6

ddy147-TF7

ddy147-TF8

Euploid

Aneuploid

Diploid Tripolar

Low Confidence Call

MaternalPaternal

XXXXXYXY

XXY (Mat)

XXY (Pat)XXY (Pat)

XXY (Pat)XXX (Pat)

X0 (Mat) X0 (Mat)

X0 (Mat)

X0 (Pat)0Y

X0 (Pat)X0 (Pat)

XXXX00XY (Pat)

X0 (Mat)XX

Chromosome Ploidy

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Sex

Single cell biopsy

biopsy(n = 17053)

Normalgametogenesis

Error in MI

Error in MII

Mitotic error

Mitosis

trophectoderm(TE)

971ndash

Disomy MaternalLoss

6 6 6 0

2 0 0 1

1 1 1 0

6 0 0 3x 3

Total 12 6 6 3

4 2 2 1Ratio

PaternalLoss

Nullisomy

Disomy Disomy

PaternalLoss

MaternalLoss

Disomy

ulatedO

bse rved

0 5 10 15 20

Maternal loss

Paternal loss

Nullisomy

Disomy

Maternal loss

Paternal loss

Nullisomy

Disomy

Count Per Sample

ulatedO

bserved

0 10 20 30 40 50

n = 1230

n = 10000

PGS Result

Other aneuploid

Euploid

Diploid tripolar

Time-Lapse Result

GG AG AA

Genotype

Acknowledgements

ddy147-TF1

ddy147-TF2

ddy147-TF3

ddy147-TF4

ddy147-TF5

ddy147-TF6

ddy147-TF7

ddy147-TF8

971ndash

Disomy MaternalLoss

6 6 6 0

2 0 0 1

1 1 1 0

6 0 0 3x 3

Total 12 6 6 3

4 2 2 1Ratio

PaternalLoss

Nullisomy

Disomy Disomy

PaternalLoss

MaternalLoss

Disomy

ulatedO

bse rved

0 5 10 15 20

Maternal loss

Paternal loss

Nullisomy

Disomy

Maternal loss

Paternal loss

Nullisomy

Disomy

Count Per Sample

ulatedO

bserved

0 10 20 30 40 50

n = 1230

n = 10000

PGS Result

Other aneuploid

Euploid

Diploid tripolar

Time-Lapse Result

GG AG AA

Genotype

Acknowledgements

ddy147-TF1

ddy147-TF2

ddy147-TF3

ddy147-TF4

ddy147-TF5

ddy147-TF6

ddy147-TF7

ddy147-TF8

Disomy MaternalLoss

6 6 6 0

2 0 0 1

1 1 1 0

6 0 0 3x 3

Total 12 6 6 3

4 2 2 1Ratio

PaternalLoss

Nullisomy

Disomy Disomy

PaternalLoss

MaternalLoss

Disomy

ulatedO

bse rved

0 5 10 15 20

Maternal loss

Paternal loss

Nullisomy

Disomy

Maternal loss

Paternal loss

Nullisomy

Disomy

Count Per Sample

ulatedO

bserved

0 10 20 30 40 50

n = 1230

n = 10000

PGS Result

Other aneuploid

Euploid

Diploid tripolar

Time-Lapse Result

GG AG AA

Genotype

Acknowledgements

ddy147-TF1

ddy147-TF2

ddy147-TF3

ddy147-TF4

ddy147-TF5

ddy147-TF6

ddy147-TF7

ddy147-TF8

Disomy MaternalLoss

6 6 6 0

2 0 0 1

1 1 1 0

6 0 0 3x 3

Total 12 6 6 3

4 2 2 1Ratio

PaternalLoss

Nullisomy

Disomy Disomy

PaternalLoss

MaternalLoss

Disomy

ulatedO

bse rved

0 5 10 15 20

Maternal loss

Paternal loss

Nullisomy

Disomy

Maternal loss

Paternal loss

Nullisomy

Disomy

Count Per Sample

ulatedO

bserved

0 10 20 30 40 50

n = 1230

n = 10000

PGS Result

Other aneuploid

Euploid

Diploid tripolar

Time-Lapse Result

GG AG AA

Genotype

Acknowledgements

ddy147-TF1

ddy147-TF2

ddy147-TF3

ddy147-TF4

ddy147-TF5

ddy147-TF6

ddy147-TF7

ddy147-TF8

PGS Result

Other aneuploid

Euploid

Diploid tripolar

Time-Lapse Result

GG AG AA

Genotype

Acknowledgements

ddy147-TF1

ddy147-TF2

ddy147-TF3

ddy147-TF4

ddy147-TF5

ddy147-TF6

ddy147-TF7

ddy147-TF8

PGS Result

Other aneuploid

Euploid

Diploid tripolar

Time-Lapse Result

GG AG AA

Genotype

Acknowledgements

ddy147-TF1

ddy147-TF2

ddy147-TF3

ddy147-TF4

ddy147-TF5

ddy147-TF6

ddy147-TF7

ddy147-TF8

Acknowledgements

ddy147-TF1

ddy147-TF2

ddy147-TF3

ddy147-TF4

ddy147-TF5

ddy147-TF6

ddy147-TF7

ddy147-TF8

Acknowledgements

ddy147-TF1

ddy147-TF2

ddy147-TF3

ddy147-TF4

ddy147-TF5

ddy147-TF6

ddy147-TF7

ddy147-TF8

Acknowledgements

ddy147-TF1

ddy147-TF2

ddy147-TF3

ddy147-TF4

ddy147-TF5

ddy147-TF6

ddy147-TF7

ddy147-TF8

ddy147-TF1

ddy147-TF2

ddy147-TF3

ddy147-TF4

ddy147-TF5

ddy147-TF6

ddy147-TF7

ddy147-TF8

ddy147-TF1

ddy147-TF2

ddy147-TF3

ddy147-TF4

ddy147-TF5

ddy147-TF6

ddy147-TF7

ddy147-TF8

Documents

Tripolar chromosome segregation drives the spanning PLK4 ... · as preimplantation genetic screening (PGS) or preimplantation genetic diagnosis for aneuploidy (PGD-A) but now termed