This Key Event Relationship is licensed under the Creative Commons BY-SA license. This license allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use. If you remix, adapt, or build upon the material, you must license the modified material under identical terms.

Relationship: 3480

Title

A descriptive phrase which clearly defines the two KEs being considered and the sequential relationship between them (i.e., which is upstream, and which is downstream). More help

Increase, DNA strand breaks leads to Cell cycle disruption

Upstream event

The causing Key Event (KE) in a Key Event Relationship (KER). More help

Increase, DNA strand breaks

Downstream event

The responding Key Event (KE) in a Key Event Relationship (KER). More help

Cell cycle disruption

Key Event Relationship Overview

The utility of AOPs for regulatory application is defined, to a large extent, by the confidence and precision with which they facilitate extrapolation of data measured at low levels of biological organisation to predicted outcomes at higher levels of organisation and the extent to which they can link biological effect measurements to their specific causes.Within the AOP framework, the predictive relationships that facilitate extrapolation are represented by the KERs. Consequently, the overall WoE for an AOP is a reflection in part, of the level of confidence in the underlying series of KERs it encompasses. Therefore, describing the KERs in an AOP involves assembling and organising the types of information and evidence that defines the scientific basis for inferring the probable change in, or state of, a downstream KE from the known or measured state of an upstream KE. More help

AOPs Referencing Relationship

AOP Name	Adjacency	Weight of Evidence	Quantitative Understanding	Point of Contact	Author Status	OECD Status
Excessive reactive oxygen species leading to growth inhibition via oxidative DNA damage	adjacent			You Song (send email)	Under development: Not open for comment. Do not cite
Increase in reactive oxygen species leading to growth inhibition via oxidative DNA damage and cell cycle disruption	adjacent	High	Moderate	You Song (send email)	Under development: Not open for comment. Do not cite	Under Development

Taxonomic Applicability

Latin or common names of a species or broader taxonomic grouping (e.g., class, order, family) that help to define the biological applicability domain of the KER.In general, this will be dictated by the more restrictive of the two KEs being linked together by the KER. More help

Term	Scientific Term	Evidence	Link
human	Homo sapiens	High	NCBI
mouse	Mus musculus	High	NCBI
rat	Rattus norvegicus	High	NCBI
mammals	mammals	High	NCBI
fish	fish	High	NCBI
crustaceans	Daphnia magna	Moderate	NCBI
green algae	Ulva compressa	Moderate	NCBI

Sex Applicability

An indication of the the relevant sex for this KER. More help

Sex	Evidence
Unspecific	Moderate

Life Stage Applicability

An indication of the the relevant life stage(s) for this KER. More help

Term	Evidence
All life stages	Moderate

Key Event Relationship Description

Provides a concise overview of the information given below as well as addressing details that aren’t inherent in the description of the KEs themselves. More help

This KER describes the causal relationship whereby increased DNA strand breaks lead to cell cycle disruption. DNA strand breaks include single-strand breaks (SSBs) and double-strand breaks (DSBs); DSBs are generally considered the most genotoxic lesions. Strand breaks are detected by sensor complexes that activate the apical kinases ATM (mainly in response to DSBs) and ATR (in response to single-stranded DNA at resected breaks and stalled replication forks). These kinases activate the effector kinases CHK2 and CHK1 and p53–p21 signaling, which inhibit the CDC25 phosphatases and cyclin-dependent kinases and thereby delay or arrest progression through G1/S, S, or G2/M (Hartwell and Weinert, 1989; O'Connell et al., 2000; Cuddihy and O'Connell, 2003; Jackson and Bartek, 2009; Ciccia and Elledge, 2010).

The relationship is conditional. Whether strand breaks cause a detectable disruption of the cell cycle depends on the number and type of breaks, whether they are repaired before the checkpoint response is sustained, the cell-cycle phase at the time of damage, and checkpoint competence. A single DSB at a defined genomic site can trigger a checkpoint response that delays cell cycle progression in human cells (van den Berg et al., 2018), whereas transient SSBs generated during normal DNA repair are usually resolved without arrest. If damage is repaired, arrest is reversible; persistent or irreparable damage can lead to prolonged arrest or permanent exit from the cell cycle (Jackson and Bartek, 2009). The relationship is not stressor-specific and can be triggered by ionizing radiation, oxidative stress, redox-active chemicals, metals, nanoparticles, or endogenous processes that generate strand breaks directly or indirectly.

Evidence Collection Strategy

Include a description of the approach for identification and assembly of the evidence base for the KER. For evidence identification, include, for example, a description of the sources and dates of information consulted including expert knowledge, databases searched and associated search terms/strings. Include also a description of study screening criteria and methodology, study quality assessment considerations, the data extraction strategy and links to any repositories/databases of relevant references.Tabular summaries and links to relevant supporting documentation are encouraged, wherever possible. More help

Evidence for this KER was assembled using the same AI-human hybrid workflow applied to the ROS-growth AOP network. Search terms were developed for both KEs and included DNA strand breaks, DNA double-strand breaks, DNA single-strand breaks, comet assay, γH2AX, ATM, ATR, CHK1, CHK2, p53, cell cycle arrest, G1/S checkpoint, G2/M checkpoint, and cell cycle disruption, combined with taxa of interest (algae, zebrafish, fish, copepods, and Daphnia). AOP-helpFinder and targeted literature searches were used to identify studies reporting both DNA strand break measurements and cell-cycle endpoints. Records were prioritized when they co-measured strand breaks (comet assay, γH2AX, PFGE) and cell-cycle distribution (flow cytometry, G-phase markers) in the same biological system, and when they reported dose or concentration, exposure duration, and evidence relevant to dose-response, temporal, or incidence concordance.

LLM-assisted screening (ChatGPT, OpenAI, San Francisco, CA, USA) was used for metadata extraction and preliminary prioritization only. All LLM outputs were checked manually against the original literature before inclusion, and final evidence selection and interpretation were performed by expert human review. Mechanistic reviews were used to support biological plausibility, and primary studies were used for empirical support where possible. Supporting evidence from endorsed AOP contexts (AOP 296 and AOP 212) was also considered.

Evidence Supporting this KER

Addresses the scientific evidence supporting KERs in an AOP setting the stage for overall assessment of the AOP. More help

Biological plausibility is high. DNA strand breaks are recognized by the DNA damage response. Double-strand breaks activate ATM (via the MRN complex), and single-strand breaks and stalled replication forks activate ATR. These kinases signal through CHK2 and CHK1 and through p53–p21 to inhibit the CDC25 phosphatases and cyclin-dependent kinases, which delays or arrests the cell cycle at G1/S, in S phase, or at G2/M (Hartwell and Weinert, 1989; O'Connell et al., 2000; Jackson and Bartek, 2009). This is canonical knowledge and is conserved across eukaryotes. In human cells, even a single Cas9-induced double-strand break can delay cell cycle progression (van den Berg et al., 2018).

Empirical support is moderate. Studies in green algae, embryonic zebrafish cells, copepods, and oyster embryos report DNA strand breaks or DNA damage responses together with cell cycle disruption or a proxy of it after exposure to genotoxic stressors (Hlavová et al., 2011; Quevedo et al., 2021; Han et al., 2014; Mai et al., 2012). The evidence is not rated high because it comes from a small number of stressors and systems, cell cycle disruption was inferred from proxy endpoints in the invertebrate studies, and few studies show dose-response, temporal, and incidence concordance for both events in the same experiment. In algae, the response also depends on the cell-cycle phase at the time of damage (Hlavová et al., 2011).

Uncertainties. Not all strand breaks cause cell cycle disruption. The outcome depends on lesion type and number, repair capacity, cell-cycle phase, and checkpoint competence, and checkpoint arrest is reversible when damage is repaired.

Biological Plausibility

Addresses the biological rationale for a connection between KEupstream and KEdownstream. This field can also incorporate additional mechanistic details that help inform the relationship between KEs, this is useful when it is not practical/pragmatic to represent these details as separate KEs due to the difficulty or relative infrequency with which it is likely to be measured. More help

The biological plausibility of this KER is high. DNA strand breaks are recognized by DNA damage response pathways. DSBs are sensed by the MRN complex (MRE11–RAD50–NBS1), which recruits and activates ATM; ATM phosphorylates H2AX (forming γH2AX) and CHK2 and contributes to p53 stabilization. SSBs and stalled replication forks generate RPA-coated single-stranded DNA that activates ATR and its effector CHK1. CHK1 and CHK2 inhibit the CDC25 phosphatases, which prevents activation of cyclin-dependent kinases and causes arrest at G1/S, in S phase, or at G2/M, while p53-dependent induction of p21 maintains prolonged arrest when damage persists (Hartwell and Weinert, 1989; O'Connell et al., 2000; Cuddihy and O'Connell, 2003; Jackson and Bartek, 2009; Ciccia and Elledge, 2010). This relationship is canonical knowledge in DNA damage response biology and is the basis of the DNA damage checkpoint concept.

Mechanistically, the response depends on the damage load and the cell-cycle context. A limited number of DSBs is sufficient to activate the checkpoint in human cells: a single Cas9-induced DSB delayed cell cycle progression, and G1-to-S progression decreased in a graded manner as the number of breaks increased (van den Berg et al., 2018). In non-mammalian taxa, DNA-damage-induced cell cycle responses have been described in green algae, although the effectors and the dependence on cell-cycle phase differ from those in animals (Hlavová et al., 2011). The KER is therefore biologically plausible across eukaryotes but conditional on lesion type, damage persistence, cell-cycle phase, and checkpoint competence.

Empirical Evidence

Provides specific (citable) evidence that supports the idea of a change in the upstream KE (KEupstream) leading to, or being associated with, a subsequent change in the downstream KE (KEdownstream), assuming the perturbation of KEupstream is sufficient. More help

Empirical support is moderate. Several studies show co-occurrence of DNA strand breaks or DNA damage responses with cell cycle disruption after exposure to genotoxic stressors, but the evidence is stressor- and system-specific, relies on proxy endpoints in some taxa, and does not include direct manipulation of strand breaks as an isolated upstream event. In Scenedesmus quadricauda, zeocin (a DSB-inducing radiomimetic) applied before S phase caused S-phase delay and a permanent G2 block, whereas zeocin applied during G2 did not arrest most cells (Hlavová et al., 2011). In embryonic zebrafish cells exposed to silver nanoparticles and ionic silver, cell cycle arrest that prevented entry into S phase occurred together with high percentages of DNA breaks, and the smaller nanoparticles and ionic silver triggered the strongest responses (Quevedo et al., 2021). In the copepod Tigriopus japonicus, gamma irradiation increased expression of DSB repair genes (DNA-PK, Ku70, Ku80) and p53, which was used here as an indirect indicator of checkpoint activation (Han et al., 2014). In Pacific oyster embryos and larvae, DNA strand breaks measured with the comet assay were positively related to larval abnormalities, which are a distal developmental endpoint rather than a direct measure of cell cycle disruption (Wessel et al., 2007; Mai et al., 2012).

Biological system	Stressor	Evidence for KE 1635 (DNA strand breaks)	Evidence for KE 1505 (cell cycle disruption)	Interpretation	Reference
Green alga (Scenedesmus quadricauda)	Zeocin (DSB-inducing radiomimetic)	DNA damage induced by zeocin, a radiomimetic that causes DNA strand breaks.	Zeocin applied from the beginning of the cell cycle: S-phase delay and permanent G2 block. Zeocin applied during G2: no arrest in about two thirds of the cells.	Supports the KER when damage occurs before S phase; shows that the response depends on cell-cycle phase.	Hlavová et al. (2011)
Embryonic zebrafish cells (ZF4)	Silver nanoparticles (10, 30, and 100 nm) and ionic silver (AgNO₃)	High percentages of DNA breaks after exposure to the smaller nanoparticles and ionic silver.	Cell cycle arrest preventing entry into S phase; the 100 nm nanoparticles led to activation of G1 and G2 phases, suggesting that cells could overcome the damage.	DNA breaks and cell cycle arrest co-occurred, and responses were strongest for the smaller nanoparticles and ionic silver.	Quevedo et al. (2021)
Marine copepod (Tigriopus japonicus)	Gamma radiation	Increased expression of DSB repair genes (DNA-PK, Ku70, Ku80), indicating DSB formation.	Increased p53 expression (indirect indicator of checkpoint activation).	Concurrent DNA damage response and p53 activation; cell-cycle phase was not measured.	Han et al. (2014)
Pacific oyster embryos and larvae (Crassostrea gigas)	Copper, cadmium, metolachlor, irgarol; benzo[a]pyrene and endosulfan	DNA strand breaks (comet assay): significant increases after 24 h from 0.1 µg/L Cu, 10 µg/L Cd, and 0.01 µg/L irgarol and metolachlor (Mai et al., 2012).	Percentage of abnormal D-larvae (developmental endpoint; not a direct measure of cell cycle disruption).	Strong positive relationship between DNA strand breaks and larval abnormalities; indirect support only.	Mai et al. (2012); Wessel et al. (2007)

Uncertainties and Inconsistencies

Addresses inconsistencies or uncertainties in the relationship including the identification of experimental details that may explain apparent deviations from the expected patterns of concordance. More help

The main uncertainty is that not all strand breaks trigger cell cycle disruption. Transient SSBs generated during normal DNA repair may be resolved without checkpoint activation, and the probability of checkpoint induction depends on lesion type (SSB or DSB), the number and clustering of breaks, the cell-cycle phase at the time of damage, and p53 and checkpoint pathway status.

A second uncertainty is that checkpoint activation is reversible when damage is repaired. Transient G1/S or G2/M arrest followed by recovery would not cause a persistent reduction in proliferation; arrest becomes prolonged or permanent only when strand break accumulation exceeds repair capacity. The response also differs among taxa and cell-cycle contexts. In Scenedesmus quadricauda, zeocin caused a permanent G2 block when applied before S phase but did not arrest most cells when applied during G2 (Hlavová et al., 2011). In Chlamydomonas reinhardtii, the genotoxicant N-OH-2-AAF produced DNA strand breaks (David et al., 2009), but cell-cycle endpoints were not measured, so that study cannot be used to assess this KER.

A third uncertainty is that the available empirical evidence comes from a small number of stressors and systems, uses proxy endpoints for cell cycle disruption in invertebrates, and was generated with genotoxic stressors in which oxidative strand breaks are not clearly separated from other genotoxic mechanisms. Evidence from oxidative-stress-specific stressors with concurrent measurement of strand breaks and cell-cycle endpoints in the same system is limited. Finally, comet assay endpoints detect strand-break-like DNA migration that may include alkali-labile sites and repair intermediates, so results must be interpreted in light of assay design and exposure duration.

Known modulating factors

This table captures specific information on the MF, its properties, how it affects the KER and respective references.1.) What is the modulating factor? Name the factor for which solid evidence exists that it influences this KER. Examples: age, sex, genotype, diet 2.) Details of this modulating factor. Specify which features of this MF are relevant for this KER. Examples: a specific age range or a specific biological age (defined by...); a specific gene mutation or variant, a specific nutrient (deficit or surplus); a sex-specific homone; a certain threshold value (e.g. serum levels of a chemical above...) 3.) Description of how this modulating factor affects this KER. Describe the provable modification of the KER (also quantitatively, if known). Examples: increase or decrease of the magnitude of effect (by a factor of...); change of the time-course of the effect (onset delay by...); alteration of the probability of the effect; increase or decrease of the sensitivity of the downstream effect (by a factor of...) 4.) Provision of supporting scientific evidence for an effect of this MF on this KER. Give a list of references. More help

Modulating factor	Details	Effect on this KER	References
DNA repair capacity	Capacity of base excision repair, single-strand break repair, homologous recombination, and non-homologous end joining.	Higher repair capacity reduces the persistence of strand breaks and the probability of checkpoint activation; impaired repair prolongs checkpoint signaling.	Jackson and Bartek, 2009; Ciccia and Elledge, 2010
Cell-cycle phase at the time of damage	G1, S, G2, or M phase when strand breaks occur.	Determines which checkpoint is engaged and whether arrest occurs. In Scenedesmus quadricauda, zeocin before S phase caused a permanent G2 block, whereas treatment during G2 did not arrest most cells. S-phase cells are vulnerable to replication-associated conversion of lesions into DSBs.	Hlavová et al., 2011; Cuddihy and O’Connell, 2003
p53 and checkpoint pathway integrity	p53, p21, ATM/ATR, CHK1/CHK2, and related checkpoint signaling.	Loss of function reduces checkpoint stringency and can allow cells to continue dividing despite damage; functional p53 maintains prolonged arrest when damage persists.	Jackson and Bartek, 2009; O’Connell et al., 2000
Number, type, and clustering of strand breaks	SSB or DSB; isolated or clustered lesions; repeated or chronic exposure.	A higher DSB burden and clustered lesions increase the probability and duration of checkpoint activation; in human cells even a single DSB can delay cell cycle progression.	van den Berg et al., 2018; Jackson and Bartek, 2009
Species and cell type	Conservation and effectors of the checkpoint pathway; proliferative state of the cells.	Checkpoint stringency and effectors vary among taxa and cell types; for example, plant lineages lack CHK1/CHK2 kinases. Non-proliferating cells show little cell cycle response.	Hlavová et al., 2011

Quantitative Understanding of the Linkage

Captures information that helps to define how much change in the upstream KE, and/or for how long, is needed to elicit a detectable and defined change in the downstream KE. More help

Quantitative understanding of this KER is moderate. The qualitative relationship is strongly supported and threshold-like behavior is biologically expected. However, quantitative data linking the number of strand breaks to the magnitude of cell cycle disruption exist for only a limited set of systems, mainly human cells, and a general model across taxa, stressors, and exposure durations is not yet available.

Response-response Relationship

Provides sources of data that define the response-response relationships between the KEs. More help

The best data defining the response-response relationship come from human cells, in which Cas9 was used to introduce a defined number of double-strand breaks. A single break at one genomic site was sufficient to trigger a checkpoint response that delayed cell cycle progression, and G1-to-S progression decreased in a graded manner as the number of breaks increased (van den Berg et al., 2018). The response also depends on repair: breaks that are repaired cause a transient delay, whereas persistent damage can lead to prolonged arrest or permanent exit from the cell cycle (Jackson and Bartek, 2009). In other taxa only semi-quantitative information is available. In Scenedesmus quadricauda, the G2 block after zeocin depended on the cell-cycle phase at the time of damage (Hlavová et al., 2011), and in embryonic zebrafish cells the co-occurrence of DNA breaks and cell cycle arrest varied with silver nanoparticle size (Quevedo et al., 2021). No mathematical function or computational model describing the relationship across taxa has been defined, and the precision with which cell cycle disruption can be predicted from a measured level of strand breaks has not been quantified. Quantitative interpretation should therefore be system-specific.

Time-scale

Information regarding the approximate time-scale of the changes in KEdownstream relative to changes in KEupstream (i.e., do effects on KEdownstream lag those on KEupstream by seconds, minutes, hours, or days?). More help

Checkpoint signaling is activated within minutes of break formation (rapid ATM/ATR activation), and delay or arrest of cell cycle progression develops over hours. The duration depends on damage severity and repair kinetics. When damage persists, permanent exit from the cell cycle can develop over days (Jackson and Bartek, 2009). Because strand breaks and checkpoint activation both occur within hours, studies that measure both events only at one late time point cannot resolve their temporal order.

Known Feedforward/Feedback loops influencing this KER

Define whether there are known positive or negative feedback mechanisms involved and what is understood about their time-course and homeostatic limits. More help

Negative feedback operates through repair: once the breaks are repaired, the activating signal disappears, checkpoint signaling is switched off, and cells re-enter the cycle, so a transient disruption causes only a delay (Cuddihy and O'Connell, 2003; Jackson and Bartek, 2009). Positive feedback and feedforward mechanisms amplify and sustain the response. ATM-dependent formation of γH2AX at breaks recruits further DNA damage response proteins and amplifies signaling (Ciccia and Elledge, 2010), and persistent p53-p21 signaling can convert a reversible arrest into permanent cell cycle exit (Jackson and Bartek, 2009). The homeostatic limit of the negative loop is the repair capacity of the cell: when strand breaks exceed it, checkpoint recovery fails and the disruption becomes prolonged. The kinetics of these loops have not been quantified for this KER. They mainly shape the time course and duration of the response, so they are described here and not represented as separate KEs.

Domain of Applicability

A free-text section of the KER description that the developers can use to explain their rationale for the taxonomic, life stage, or sex applicability structured terms. More help

The domain of applicability of this KER is broad but conditional. It is applicable to eukaryotic cells and organisms in which DNA strand breaks can accumulate and the DNA damage checkpoint is functional. ATM/ATR-dependent checkpoint signaling is highly conserved, and cell-cycle responses to DNA damage have been described in yeast (Hartwell and Weinert, 1989), green algae (Hlavová et al., 2011), invertebrates, fish cells, and mammalian cells; the downstream effectors differ among lineages. Applicability is highest in actively proliferating systems, including dividing cells in embryonic and juvenile organisms, cell cultures, and proliferating tissues, and is lower in post-mitotic tissues. The KER is not sex-specific. Empirical support is strongest for mammalian cell systems, while evidence from other taxa comes mainly from green algae, a fish embryonic cell line, copepods, and oyster embryos and larvae. The KER is not stressor-specific: any stressor that generates DNA strand breaks, including ionizing radiation, oxidative stress, redox-active chemicals, metals, and nanoparticles, falls within its stressor domain. Taxonomic extrapolation is supported by the conservation of the DNA damage response across eukaryotes, but quantitative predictions should be made cautiously because repair capacity, checkpoint stringency, and cell-cycle structure vary among taxa and life stages.

References

List of the literature that was cited for this KER description. More help

AOP-Wiki. 2026. Relationship 3480: Increase, DNA strand breaks leads to Cell cycle disruption. Available at: https://aopwiki.org/relationships/3480. Accessed 14 May 2026.

Ciccia A, Elledge SJ. 2010. The DNA damage response: making it safe to play with knives. Molecular Cell 40:179-204. https://doi.org/10.1016/j.molcel.2010.09.019.

Cuddihy AR, O’Connell MJ. 2003. Cell-cycle responses to DNA damage in G2. International Review of Cytology 222:99-140. https://doi.org/10.1016/S0074-7696(02)22013-6.

David RM, Winter MJ, Chipman JK. 2009. Induction of DNA strand breaks by genotoxicants in the alga Chlamydomonas reinhardtii. Environmental Toxicology and Chemistry 28(9):1893-1900. https://doi.org/10.1897/08-349.1.

Han J, Won EJ, Lee BY, Hwang UK, Kim IC, Yim JH, Leung KMY, Lee YS, Lee JS. 2014. Gamma rays induce DNA damage and oxidative stress associated with impaired growth and reproduction in the copepod Tigriopus japonicus. Aquatic Toxicology 152:264-272. https://doi.org/10.1016/j.aquatox.2014.04.005.

Hartwell LH, Weinert TA. 1989. Checkpoints: controls that ensure the order of cell cycle events. Science 246(4930):629-634. https://doi.org/10.1126/science.2683079.

Hlavová M, Čížková M, Vítová M, Bišová K, Zachleder V. 2011. DNA damage during G2 phase does not affect cell cycle progression of the green alga Scenedesmus quadricauda. PLoS ONE 6(5):e19626. https://doi.org/10.1371/journal.pone.0019626.

Jackson SP, Bartek J. 2009. The DNA-damage response in human biology and disease. Nature 461:1071-1078. https://doi.org/10.1038/nature08467.

Mai H, Cachot J, Brune J, Geffard O, Belles A, Budzinski H, Morin B. 2012. Embryotoxic and genotoxic effects of heavy metals and pesticides on early life stages of Pacific oyster (Crassostrea gigas). Marine Pollution Bulletin 64(12):2663-2670. https://doi.org/10.1016/j.marpolbul.2012.10.009.

O’Connell MJ, Walworth NC, Carr AM. 2000. The G2-phase DNA-damage checkpoint. Trends in Cell Biology 10(7):296-303. https://doi.org/10.1016/S0962-8924(00)01773-6.

Quevedo AC, Lynch I, Valsami-Jones E. 2021. Cellular repair mechanisms triggered by exposure to silver nanoparticles and ionic silver in embryonic zebrafish cells. Environmental Science: Nano 8(9):2507-2522. https://doi.org/10.1039/D1EN00422K.

van den Berg J, Manjón AG, Kielbassa K, Feringa FM, Freire R, Medema RH. 2018. A limited number of double-strand DNA breaks is sufficient to delay cell cycle progression. Nucleic Acids Research 46(19):10132-10144. https://doi.org/10.1093/nar/gky786.

Wessel N, Rousseau S, Caisey X, Quiniou F, Akcha F. 2007. Investigating the relationship between embryotoxic and genotoxic effects of benzo[a]pyrene, 17α-ethinylestradiol and endosulfan on Crassostrea gigas embryos. Aquatic Toxicology 85(2):133-142. https://doi.org/10.1016/j.aquatox.2007.08.007.