A review of the genetic, physiological, and agronomic factors influencing secondary dormancy levels and seed vigour in Brassica napus L.

Abstract Dormancy in canola (Brassica napus L.) is a complicated process due to many overlapping and interacting factors affecting the absolute dormancy levels. It is unknown if seed dormancy plays a role in the poor stand establishment of planted canola but given that germination and dormancy are two ends of the same continuum, it has been suggested that dormancy may be a culprit of poor establishment. This review synthesizes literature pertaining to secondary dormancy in B. napus and the interaction of genetic, physiological, environmental, and agronomic factors. Seed germination and vigour and the interaction with dormancy are also addressed. The persistence of volunteer canola seed in the soil seedbank is a result of the induction of canola seed into secondary dormancy under adverse environmental conditions such as low temperature or low light. Genetics is a major influencing factor on absolute secondary dormancy (∼50%) in canola. Plant hormones abscisic acid and gibberellic acid and their interactions also influence dormancy with highly dormant genotypes having increased abscisic acid concentration in the seed. Seed sugars, seed storage proteins, glucosinolate content, and growth habit are all additional factors affecting absolute dormancy in B. napus. Furthermore, maternal environmental conditions affect dormancy levels. In addition to genetic, physiological, and environmental factors, farming practices such as harvest timing, and tillage regimes can influence secondary dormancy of canola seed that has entered the seedbank unintentionally. Given the documented high heritability of secondary dormancy, it is feasible to reduce secondary dormancy in canola cultivars; however, consideration of all interacting factors must be given.


Introduction
Canola seed germinating and emerging subsequent years after the seed has entered the soil seedbank results in volunteer canola. The relative abundance of volunteer canola, measured based on frequency, uniformity, and density, ranks as the fifth most occurring weed on the Canadian prairies (Beckie 2015). Persistence of the seed in the soil seedbank results from the ability of canola seed to be induced into secondary dormancy under adverse environmental conditions (Baskin and Baskin. 1998). Two types of dormancy exist in canola seed depending on the stage of maturity of the seed. Primary dormancy occurs while the seed is maturing on the mother plant or shortly thereafter. Primary dormancy, due to increased levels of abscisic acid (ABA), prevents seed germination in the pod known as precocious germination or vivipary. Secondary dormancy can be present in mature seed and is induced when environmental conditions are not favourable for germination. An additional temporary seed state is quiescence, which is the temporary delay of germination in non-dormant seeds until all germination conditions are met. These three processes overlap on the germination continuum as demonstrated in Fig. 1.
Volunteer canola can be an issue if the volunteers and the planted crop have the same herbicide resistance such as glyphosate-resistant soybean making the control of volunteers very difficult (Geddes and Gulden 2017). There is also the potential for gene stacking between different herbicide modes of action causing the subsequent volunteer populations to be resistant to more than one herbicide and very difficult to control (imidazolinone and glyphosate resistance) (Hall et al. 2000). Furthermore, the potential for crosscontamination of oil profiles due to pollen-mediated gene flow or volunteer emergence in a seeded canola field is a possibility. Of greatest concern for contamination is high erucic acid volunteers contaminating a "double-low" field, or the contamination of identity preserved canola crops with speciality oil profiles. A more recent concern is the perpetuation of the soil-borne pathogen clubroot (Plasmodiophora brassicae) Fig. 1. Canola seed development continuum from late seed maturation to seed germination. Primary dormancy and vivipary reach the lowest level as the seed is after ripened. In after ripened seed, secondary dormancy levels increase and can cycle for many years before the seed eventually germinates or dies.
as the volunteers act as host plants resulting in no rotational break from canola (Strelkov and Hwang 2014). One way to minimize the persistence of B. napus seed in the soil seedbank that can then become volunteers may be to reduce the genetic potential of the seed to be induced into secondary dormancy. Harvest losses may still occur, but the seed will lack the potential to be induced into a prolonged dormancy period if poor environmental conditions persist.
Western Canadian agronomists have postulated that the poor emergence of canola may, in part, be due to seed dormancy. Typical field emergence of canola after seeding is around 50% of what is planted, with the non-emerged seed having many possible fates such as attack by disease, seed disappearance due to predation by pests or natural death sometimes accompanied by limited moisture for germination (Baskin and Baskin 1998;Harker et al. 2003Harker et al. , 2015. Direct selection in canola breeding programs for low secondary dormancy does not occur; therefore, the effect of reducing dormancy on seed vigour and other seed traits is unknown (Bus et al. 2011;Schatzki et al. 2013a).
The largest contributing factor affecting secondary dormancy is the genetic makeup of the seed (Gulden et al. 2004a). Other factors such as maternal environment, physiological characteristics, seed components, and agronomic practices affect absolute secondary dormancy levels within the seed. This review will provide a current overview of secondary dormancy research on B. napus and the potential effects on seed vigour, which has not been conducted since the early 2000s (Pekrun et al. 1998;Gulden and Shirtliffe 2009).

Dormancy in Brassica napus
Brassica napus is a non-deep physiological dormant species meaning an unknown prohibiting mechanism within the embryo prevents radicle protrusion under continuous darkness and conditions not favourable for germination such as cold temperature or low moisture (Pekrun et al. 1997). Non-deep dormancy is broken by a short period of imbibed cold stratification or after-ripening (Baskin and Baskin 1998). Under nondeep physiological dormancy, the seed cycles between fully dormant, conditionally dormant, and non-dormant states. When B. napus seedbank populations germinate under a narrow range of environmental conditions, the seed has broken conditional dormancy. When the seed is non-dormant, germination can occur under a wider range of conditions (Baskin and Baskin 1998). Seed that has entered this cycling pattern and does not germinate in the current season can persist in the soil for up to 7 years (Beckie and Warwick. 2010), but typically persist for 2-5 years in Canada (Légère et al. 2001;Simard et al. 2002).
Brassica napus seeds directly off the mother plant are classified as non-dormant and do not require light as a germination condition (Baskin and Baskin 1998;Soltani et al. 2018). B. napus has a low potential for primary dormancy, induced during seed development on the mother plant because of elevated ABA, one of two major hormones governing seed germination and dormancy (Baskin and Baskin 1998). It is unclear if primary dormancy is a prerequisite for secondary dormancy in B. napus (Soltani et al. 2018). The correlation between primary and secondary dormancy is variable with r-values between the two traits ranging from 0 to 0.62 (Schatzki et al. 2013b;Soltani et al. 2018). The range of results is believed to be due to variation in how the experiments were performed as well as the post-harvest conditions prior to testing of seed. If the seed was "fresh" but stored in ambient temperatures for a few months before testing it, there is opportunity for after-ripening to occur which will degrade primary dormancy (Soltani et al. 2018). In other plant species, a clear association between primary and secondary dormancy exists, such as wild oat (Avena fatua L.), where the seed cannot enter secondary dormancy without some degree of primary dormancy (Symons et al. 1987;Soltani et al. 2018). Furthermore, a positive correlation between primary and secondary dormancy also exists in Arabidopsis, a relative of B. napus (Soltani et al. 2018).
Another consideration when screening for dormancy is the cycling between dormant and non-dormant states that can occur. Typically, primary dormancy assays are performed under a constant temperature in the dark; however, these conditions may not break conditional dormancy. What is more fitting for testing primary dormancy is the use of a wide range of temperatures in both light and dark conditions, therefore identifying conditionally versus primarily dormant seed (Soltani et al. 2018). Clarifying the relationship between primary and secondary dormancy is an important step in defining strategies for canola breeders to select for low primary and secondary dormancy while preventing precocious germination on the mother plant.

Genetic factors contributing to secondary dormancy in Brassica napus
3.1. Heritability and genetic control Secondary dormancy in B. napus is a highly heritable, yet quantitative trait controlled by multiple genes. A wide distribution of dormancy values across the species has been observed. Spring-type B. napus genotypes absolute values range from 0% to 90% dormant (Gulden et al. 2004a) and winter-type B. napus range from 0% to 60% dormant (Momoh et al. 2002). The reported heritability for secondary dormancy in wintertype B. napus ranges from 0.96 to 0.97 (Weber et al. 2013;Schatzki et al. 2013a). While direct selection for secondary dormancy has never occurred before, it is possible given the high heritability of the trait. Five quantitative trait loci (QTL) were identified in a winter DH B. napus mapping population accounting for 42% of the phenotypic variance in total seed dormancy (primary + secondary) (Schatzki et al. 2013b). Furthermore, four QTLs were identified specifically controlling secondary dormancy in the same population accounting for 35% of phenotypic variance. Limited QTL studies on spring B. napus have been done, but it is the expected QTL detection would be quantitative as with the winter type.

Delay of Germination 1 (DOG1) gene
The discovery of the Delay of Germination (DOG) family has provided insight into the genetic control of seed dormancy in plant species (Graeber et al. 2014). First found in an Arabidopsis recombinant inbred population, Delay of Germination 1 (DOG1) was specifically involved in increased primary and secondary seed dormancy (Alonso-Blanco et al. 2003). DOG1 codes for a protein of unknown function, which is highly expressed during seed maturation, detectable in dry seeds and decreases following seed after-ripening (Bentsink et al. 2007). The Arabidopsis mutant, dog1, has no dormancy thus, confirming the importance of the gene and protein in seed dormancy (Nakabayashi et al. 2012;Graeber et al. 2014).
The recent discovery of DOG1 involvement in ABA signalling has concluded the two pathways are not only parallel to each other but epistatic in nature (Nonogaki. 2019). In Arabidopsis, ABA HYPERSENSITIVE GERMINATION1 (AHG1) and AHG3, clade A protein phosphatases (PP2Cs), negatively regulate ABA signalling and seed dormancy (Nonogaki 2019). The PP2Cs function downstream of DOG1 and promote germination, while DOG1 increases dormancy due to increased ABA sensitivity. However, when an ABA-deficient mutant with the functional DOG1 was tested for dormancy, DOG1 alone could not induce primary dormancy (Nonogaki 2014). Expression of DOG1 is temperature sensitive and determines the optimum temperature for germination to occur. Greater accumulation and storage of the DOG1 protein in the seed inhibits germination, whereas lower seed content of DOG1 protein permits germination to proceed (Graeber et al. 2014).
Seed longevity describes the viability of seed following dry storage with greater longevity extending the lifespan of the seed (Rajjou and Debeaujon 2008). Decreased longevity has been observed in Arabidopsis dog1 mutants, which are non-dormant (Carrillo-Barral et al. 2020). Primary metabolites essential for seed storage and survival were less abundant in dog1 mutants during seed maturation, thus seed energy reserves required to prolong their lifecycle are reduced (Dekkers et al. 2016). It is hypothesized that different genetic mechanisms control longevity and dormancy because QTL for both traits has not been found to co-locate (Carrillo- Barral et al. 2020). Regardless, these findings suggest that DOG1 may not only play a role in seed dormancy but other seed characteristics including but not limited to seed maturation on the mother plant (Dekkers et al. 2016;Nonogaki 2019;).
While most of the DOG1 work is in the model species Arabidopsis, three homologues in B. napus have been identified with high homology to DOG1 in Arabidopsis (BnaADOG1.a, BnaCDOG1.a, and BnaDOG1.b) (Nee et al. 2015). The combined BnaDOG1 transcripts are detected in seed tissues and higher DOG1 transcript levels, and resulting proteins are detected after secondary dormancy is induced (Nee et al. 2015). There currently is a gap in the understanding of the relative contribution of each homologue to secondary dormancy as well as the diversity in DOG1 sequences across B. napus. However, a proposed avenue for screening for genetic secondary dormancy could be examining polymorphisms between high and low secondary dormancy genotypes in the identified B. napus DOG1 sequences and creating markers for use in marker-assisted breeding (Nee et al. 2015). Relative transcript levels should also be quantified and associated with corresponding dormancy levels.

Mother of FT and TFL1 (MFT) gene
In recent years, more genes associated with seed dormancy have been identified. The Mother of FT and TFL1 (MFT) gene is conserved across plant species and is involved in seed dormancy and germination regulation in Arabidopsis and wheat (Xi et al. 2010;Nakamura et al. 2011;Lui et al. 2020). Four homologues of this gene have been identified in B. napus (BnaMTFs) with high similarity to Arabidopsis (AtMFT) (Lui et al. 2020). The BnaMFTs transcripts were found to increase 3.5fold following secondary dormancy induction using Polyethylene glycol 6000 compared to untreated seed (Lui et al. 2020). Transcripts were only detectible in seed and transcription levels peaked 35-42 days after pollination (Lui et al. 2020). BnaMFTs and BnaDOG1 have similar expression patterns; however, BnaDOG1 transcripts were greater in seeds than BnaMFT transcript levels. The two genes are thought to function independently from each other. BnaMFTs were shown to be ABA dependent, whereas BnaDOG1 functions parallel to ABA pathways (Lui et al. 2020). As proposed for BnaDOG1, to further our understanding of BnaMFTs it would be beneficial to look at differential gene expression and sequence polymorphisms between high and low dormancy genotypes.

Maternal environment
Secondary dormancy in spring B. napus is affected by maternal environmental conditions. The absolute dormancy of a genotype differs when produced in different environmental locations and (or) years (Gulden et al. 2004a). Ideal environmental conditions producing high yield and high oil content were found to increase secondary dormancy in winter B. napus produced in Germany (Schatzki et al. 2013b). Huang et al. (2016) also determined winter B. napus seed produced under ideal conditions exhibited higher dormancy than seed produced in stressful environments. Weber et al. (2013) tested spring and winter B. napus cultivars in two different locations in Germany over several years and the absolute dormancy varied across years and locations, but the general ranking of the cultivars remained consistent. Regardless of the growth habit, these studies highlight the variability as well as the influence the maternal environment has on secondary dormancy levels.
Maternal environmental conditions such as high heat or moisture stress can lower secondary dormancy in winter B. napus (Weber et al. 2013;Brunel-Mugnet et al. 2015). Hormone balance, age of the mother plant, and seed position on the mother plant all impact dormancy in other plant species (Baskin and Baskin 1998). Higher nitrate levels in the soil have been found to reduce dormancy by affecting ABA synthesis and degradation (Matakiadis et al. 2009;He et al. 2014). Specifically, ABA gene expression is regulated, in part, by nitrate and higher nitrate decreases ABA levels, thereby decreasing dormancy (Matakiadis et al. 2009). Similarly, a green-house study found a weak inverse correlation between available soil nitrogen and secondary dormancy in spring-type B. napus. Seed produced under high nitrogen had lower secondary dormancy when compared to seed produced under ideal or limiting soil nitrogen (Charles Geddes, personal communication). The correlation was non-significant in the corresponding field study but does warrant further study to examine the possible relationship.
An association between field and greenhouse-produced seed for dormancy emphasizes the importance of maternal environment and dormancy. Field-produced Arabidopsis seed possessed higher dormancy compared to greenhouse seed, an important consideration when conducting seed dormancy experiments (Postma and Âgren 2015). The greenhouse likely is a more consistent and less stressful environment than the field, and as a result may be why the seed produced in it has lower dormancy. Results from controlled environments, like a greenhouse, do not always translate to the field where conditions are much more variable.

Seedbank conditions
Abiotic stresses within the soil seedbank impacting secondary dormancy of the seed include light, temperature, anoxia (low oxygen), and osmotic stress (Pekrun et al. 1997;Momoh et al. 2002). Osmotic stress combined with constant temperature was the most effective way to induce secondary dormancy in B. napus seeds (Momoh et al. 2002). Constant temperature was necessary for dormancy induction as fluctuating temperatures are a dormancy-breaking signal within the seed. In the experiments establishing this, all dormancy testing was done under green light not detectable by seeds to ensure dormancy was not broken by a light signal. Flooding or anoxia did not induce secondary dormancy, and the seed either rotted or germinated immediately once the conditions were removed (Pekrun et al. 1997;Momoh et al. 2002). Seed lots tested under higher temperatures (20 • C) and osmotic stress were not only artificially induced into dormancy faster, but the seed populations also had greater overall dormancy compared to lower temperatures (12 • C) in winter-type accessions (Momoh et al. 2002). The soils in western Canada during early to mid-May are cold and often saturated, reducing the propensity of the seed to be induced into secondary dormancy. Rather those conditions can signal germination or increase the chance of the seeds rotting due to the low oxygen conditions. Conversely, when seedbank additions occur in the fall with harvest seed losses, the soil is dry creating ideal secondary dormancy-inducing conditions.

Post-harvest seed storage conditions
Post-harvest seed storage conditions influence secondary dormancy with increased length of storage and higher temperatures decreasing secondary dormancy and seed longevity of B. napus (Gulden et al. 2003a;Nguyen et al. 2015). Artificially preserving seed dormancy is most effective at a temperature of −70 • C (Gulden et al. 2004a). Momoh et al. (2002) tested winter seed lots for secondary dormancy from a variety of storage conditions ranging from room temperature to low-temperature storage. It was observed that the longer the seed lots were stored in cold conditions, the greater the secondary dormancy. The study was not designed to specifically test storage conditions, only the length and type of storage were compared across seed lots. Nevertheless, the results are in congruence with what is observed among the species, which is that cold storage temperatures maintain the degree of secondary dormancy.
A study on spring B. napus genotypes and the effect of storage conditions on secondary dormancy found the greatest change in dormancy levels was observed when temperature cycling between 25 and 45 • C occurred (Vail and Headley 2018). Other temperature treatments that led to a loss of dormancy occurred when the temperature increased from −20 to 25 • C both quickly and more gradually (Vail and Headley 2018). The genotype, temperature treatment, and temperature treatment × genotype interactions were all statistically significant in the loss of dormancy. These results highlight the importance of seed storage conditions prior to and during secondary dormancy screening to prevent dormancy loss due to drastic temperature changes. From a commercial seed storage standpoint, these results may be favourable for dormancy loss in seed that is being planted in the soil that season. It should be clarified though that the seed farmers are planting is not the seed of concern for volunteer perpetuation, rather the seed entering from pod shattering is the seed of concern.

Seedbank addition
The largest addition of seed into the soil seedbank occurs at the seed maturation stage and during seed harvest. As the crop dries down, pods become brittle and prone to shattering which releases the seed into the seedbank. In addition, canola seed exits the combine harvester at harvest and enters the seedbank (Haile and Shirtliffe. 2014). Not all seeds entering the seedbank survive, but seed losses of up to 3000 viable seeds/m 2 or 5.8% of total yield have been observed (Gulden et al. 2003b). These seed losses contribute to the continuation of seedbank input which can eventually become volunteers.
Harvest methods commonly practiced in western Canada include swathing the crop at approximately 60% seed colour change and allowing the crop to further dry down in windrows before it is combined (Haile and Shirtliffe 2014). The second method, which has become more common over the past 10 years, leaves the crop standing longer in the field to dry down before it is directly combine harvested. Haile and Shirtliffe (2014) studied different harvest techniques in western Canada and the relationship with seed dormancy. Seed harvested and screened for dormancy before full maturity (10%-60% seed colour change) were more likely to exhibit primary rather than secondary dormancy, while seed screened at maturity (colour change complete) had higher secondary dormancy levels. Primary dormancy at the early harvest timing ranged from 13% to 16% and decreased to near zero in mature seeds. It was concluded that seed harvested by swathing then combining ensured that the seed is at proper maturity while reducing some of the seedbank input caused by pod shattering if the crop is left in the field longer. Harvesting the seed any earlier than at 60% seed color change may reduce seedbank inputs; however, the subsequent quality of the seed may be compromised and is not a recommended method to reduce seedbank input (Haile and Shirtliffe 2014).
A study evaluating on-farm harvest losses of canola seed found good farm management practices that led to increased yield such as timely fungicide application, proper swathing timing, and lower combine harvester speed led to reduced harvest losses (Cavalieri et al. 2016). Wind speed was a contributing factor that increased seed losses due to pod shatter; however, that factor is beyond the control of the producer (Cavalieri et al. 2016).
The recent introduction of pod shatter resistant varieties along with proper farm management on the Canadian prairies will help with some of the seed loss. Pod shatter resistant varieties will also make direct combining a viable option due to reduced seed loss (Raman et al. 2014). However, the complete reduction of harvest losses is not possible, and some seedbank input will occur which can then become volunteers in subsequent seasons. While the primary focus in preventing volunteers should be on reducing the genetic potential of the seed to be induced into secondary dormancy, farmer education, and outreach should also be a priority because these factors are within farmers control and strategies can be employed to limit seedbank introduction.

Tillage regime
Tillage regimes have been studied to determine which methods reduce the persistence of B. napus seed in the soil. In Germany, delayed tillage or zero-till was the most effective methods for volunteer control in winter B. napus when compared to immediate tillage following harvest. Immediate tillage (<15 cm) incorporated the seeds on the soil surface into the top layers of soil and resulted in higher volunteer populations in the following year following a vernalization period (Gruber et al. 2005). Tillage treatments of 15 and 25 cm incorporated the seeds on the surface into greater depths where the seeds were likely to be induced into secondary dormancy or germinated deeply in the soil and were unable to reach the surface. Over time with continued tillage, the deeply buried seeds were brought back to the surface and germination could occur even several years after burial (Gruber et al. 2005). It should be noted that in this study up to 75% of seeds that did not germinate or became deeply buried were never accounted for, and it is believed that the seed was either removed by animals or rotted due to disease, another factor contributing to the seedbank dynamics (Gruber et al. 2005).
A more recent study, also on winter B. napus in Germany, also found the timing of tillage had a more significant effect on volunteer occurrence compared to the depth at which the seed was buried (Huang et al. 2018). Like the previous study, immediate tillage caused higher volunteer populations the following season (Huang et al. 2018). Three weeks post-harvest was the recommended time delay before tillage deeper than 6 cm should be applied to reduce the volunteer populations the following year. In the 3 weeks from harvest to tillage, the seed was exposed to fluctuating environmental conditions and lost dormancy potential before being buried too deep to emerge (Gulden et al. 2003a;Huang et al. 2018). Although increased depth resulted in greater dormancy in the previous German study, the depth factor alone at which the seed was buried was found to not have a significant effect on volunteer occurrence when investigated by Huang et al. (2018).
Conversely, in spring B. napus in western Canada tillage immediately following harvest in early autumn was the most effective method to reduce volunteer populations in subsequent years (Geddes and Gulden 2017). Post-harvest tillage signalled the seeds in the seedbank to germinate (seedbank recruitment) followed by winterkill. Seed persistence in the seedbank after autumn tillage was reduced by half compared to either zero-till or spring tillage regimes (Geddes and Gulden. 2017). These results contrast with Gruber et al. (2005) and Huang et al. (2018), possibly because of differences in climatic conditions and growth habit differences of the crop. In western Canada, the recruitment of volunteers the same season the seed entered the soil is desired because the volunteers cannot survive the harsh winters and ultimately die before contaminating the current crop, whereas, in Europe, the immediate recruitment of volunteers is not desired because fall-germinated volunteers are vernalized over the winter, like the winter seeding practice and continue their lifecycle to following spring.
Another factor affecting B. napus volunteer presence is cultivar choice and the corresponding dormancy levels. Volunteers were identified in farmers' fields and genotyped through polymerase chain reaction and cluster analysis to determine the cultivar. Most of the identified volunteers were from high dormancy cultivars . Volunteer plant populations of 0-7 plants/m −2 for the high dormancy compared to 0-1.3 plants/m −2 for the low dormancy cultivars were found. The greatest volunteer plant populations were seen the year following cultivation. This study concluded that cultivar choice contributed more to volunteer populations via shatter than tillage regime. In terms of contributing factors to volunteer occurrence, the environment is not controllable or predictable; however, farm management decisions can be controlled by farmers and are recommended to reduce volunteer populations ).

Spring versus winter growth types
A marked difference exists between winter and spring genotypes for secondary dormancy, with the spring types having higher dormancy potential. In the study establishing this, 26 winter-types and three spring genotypes were screened for secondary dormancy. The spring types had higher absolute values when compared to the winter types under conditions of osmotic stress and darkness (Momoh et al. 2002). The development of secondary dormancy in the spring types ranged from 0% to 85% dormant, while the winter types ranged from 0% to 60% dormant (Momoh et al. 2002). Similarly, 75% of the spring genotypes screened by Gulden et al. (2004a) had high dormancy (>60% dormancy). Differences in secondary dormancy between the two growth types are due to the physiology of each growth type (Momoh et al. 2002). Spring types may exhibit greater dormancy to avoid autumn growth when germination conditions are ideal but result in subsequent winterkill, whereas winter genotypes must germinate in the fall to fulfil the vernalization requirement needed to initiate flowering the following spring. Thereby winter types may be less sensitive to dormancy-inducing signals from the environment.

Plant hormones
The hormones ABA and gibberellic acid (GA) work antagonistically in the control of seed dormancy and germination, respectively. ABA prohibits germination by preventing radicle elongation, cell wall loosening, as well as preventing the degradation of seed storage proteins (SSP; Schopfer and Plachy. 1985;Bewley. 1997). Gibberellic acid counteracts the effect of ABA and weakens the testa, thereby allowing radicle protrusion and the completion of germination signalled by light and temperature (Weitbrecht et al. 2011;Dekkers et al. 2016).
Crosses in Arabidopsis between ABA-deficient and wild-type lines showed maternally inherited tissues such as the endosperm produce ABA, but only ABA produced in the embryo can induce dormancy (Karsen et al. 1983;Baskin and Baskin 1998). ABA levels spike during early seed development, thereby preventing germination of the seed while still on the mother plant. As the seed dries down, levels decrease with the lowest concentrations in dry seed (Hilhorst. 1995;Baskin and Baskin. 1998). Low temperatures during seed maturation cause an increase in ABA, resulting in greater dormancy as seed germinating in cold temperatures make survival less likely (He et al. 2014).
A study between high (AC Excel) and low (DH12075) dormancy spring B. napus genotypes examined the changes in gene expression related to ABA when osmotically stressed (Fei et al. 2007). Three highly upregulated genes at the fullsize embryo stage were identified in both the high and low dormancy genotypes (Fei et al. 2007). All three genes were either induced by or responsive to ABA (Fei et al. 2007). A further study on the same genotypes found ABA was three times higher in the high dormancy genotype when compared to the low dormancy genotype. ABA content of the high dormancy genotype tripled again following dormancy induction (Fei et al. 2009). After induction, 28 genes were commonly upregulated in both genotypes with the majority involved in metabolism and abiotic stress detection. Furthermore, 158 genes were solely upregulated in the high dormancy genotype and involved in primary and secondary metabolism, protein biosynthesis, and metabolism compared to only 10 genes in the low dormancy genotype (Fei et al. 2009). This study highlights the effect of ABA not only on dormancy but other seed processes that are indirectly affected by dormancy.
In another study, high dormancy genotypes were found to be more sensitive to ABA application than low dormancy genotypes after being osmotically treated (Gulden et al. 2004b). Applied ABA did not affect germination of the low dormancy genotypes but did for the high dormancy genotypes (Gulden et al. 2004b). The total ABA concentration within the osmotically treated seeds of the high dormancy genotype was approximately two times that of the low dormancy genotype. The reduced sensitivity to ABA of the low dormancy genotype may be attributed to the germination process already being initiated within the seed and the resulting insensitivity to ABA (Schopfer and Placy 1985;Gulden et al. 2004b). Whereas the greater dormancy in the high dormancy genotype may be due to increased ABA sensitivity (Gulden et al. 2004b). The literature has established the important role both ABA and GA plays in dormancy and germination. Future studies evaluating the upstream gene signalling that results in the hormone changes would be of great value to better understand how dormancy and germination are initiated.
The main plant hormones controlling seed dormancy are ABA and GA; however, ethylene and auxin also play a role. Primary dormancy in Arabidopsis is affected by auxin levels; however its direct role remains unknown (Bai et al. 2018;Lui et al. 2019). To further investigate the role auxin plays in dormancy, two winter-type B. napus genotypes were examined for differential gene expression using RNA sequencing (Lui et al. 2019). Between the two genotypes, 998 differentially expressed genes were detected. The upregulated genes related to the indole glucosinolate (GLS)-linked auxin biosynthesis pathway (Lui et al. 2019). This study also detected upregulation in ABA biosynthesis in the high dormancy genotype, like previous studies (Fei et al. 2009;Lui et al. 2019). In the work by Lui et al. (2019), exogenously applied indole-3-acetic acid (IAA) increased dormancy level in the high dormancy genotype. The results from this study provide further insight into the physiological mechanisms of secondary seed dormancy in B. napus.

Seed sugar and protein
Seed components have been examined for possible relationships with secondary dormancy, and two of interest are sugar and protein. Following induction, higher concentrations of glucose were found in a high dormancy B. napus genotype when compared to a low dormancy genotype (Fei et al. 2009). Post induction, the high dormancy genotype had greater cell wall strength, likely playing a role in restricting the radicle from emerging, as well as increased sugar movement and metabolism (Finch-Savage and Leubner-Metzger 2006;Fei et al. 2009). In Arabidopsis, exogenously applied glucose prohibited germination at concentrations as low as 1% (Dekkers et al. 2004). This study was the first to find inhibitory effects of glucose at such low concentrations. The exact reasons behind the inhibitory effects of glucose remain unknown. No interactions were found between sugar and GA synthesis nor did the glucose cause osmotic stress-reducing germination (Dekkers et al. 2004). When the seed coats of the seeds were removed, germination was still restricted indicating that glucose affected the growth potential of the embryo rather than the seed coat. ABA-deficient mutants were screened in conjunction with glucose application and germi-nation was not affected suggesting that glucose was more prohibitive when ABA is present in seeds. It is hypothesized that glucose affects ABA signalling, but the two pathways are separate (Dekkers et al. 2004). Moving forward, further exploration of the effects of glucose on maintaining seed dormancy should be an area of focus.
The two major components within canola seed are oil and protein. Within the protein fraction of the seed, the SSP comprise 80% (Mieth et al. 1983;Aider and Barbana 2011), with the remaining proteins being mainly oil body and lipid transfer proteins (Schatzki et al. 2014). Seed storage proteins contribute to the germination of the seed and seedling vigour, dormancy, and seed longevity (Schatzki et al. 2014;Nguyen et al. 2015). The two main SSP in B. napus are 12S cruciferin, which is known to buffer oxidative stress during seed storage, and the smaller 2S napin protein. Both SSP provide nutrients to the germinating seed and seedling before photosynthesis occurs, and seedlings become autotrophic. Cruciferin makes up 60%, and napin makes up 20% of the SSP within the seed (Schatzki et al. 2014). The relative proportions of SSP are highly influenced by the environment and genetics, and the heritability of each protein ranges from 0.77 to 0.79 (Schatzki et al. 2014). In total, five QTLs have been identified with three QTLs associated with napin and two for cruciferin accounting for 47% and 35% of phenotypic variation, respectively (Schatzki et al. 2014).
A negative correlation between cruciferin and napin exists with higher cruciferin found in modern winter B. napus genotypes (Schatzki et al. 2014). Greater cruciferin content is due to co-localization between napin content and seed glucosinolate content. Thus, selection for low glucosinolate varieties has indirectly resulted in the reduction in napin content (Schatzki et al. 2014). In a winter-type mapping population generated from a canola quality parent and high glucosinolate/high erucic acid parent, a significant negative correlation was found for napin content and total seed dormancy (Schatzki et al. 2014). If this negative correlation between napin content and dormancy is found across the B. napus species, it is expected that modern spring-type lowglucosinolate canola will have higher dormancy levels. When selection for one SSP occurs the total protein content does not decrease, instead the other SSP compensates to result in the same total protein (Schatzki et al. 2014). A closer look at the association between napin and dormancy revealed the production of napin is regulated, in part, by the gene ABI3 which also plays a role in ABA production, and ABA is known to promote dormancy (Schatzki et al. 2014). In summary, modern varieties with low napin content due to reduced glucosinolate content may have higher dormancy due to the negative correlation between napin and dormancy (Schatzki et al. 2014). This association requires further study, and the possibility for indirect selection exists whereby dormancy level could be inferred from the SSP fractions.

Relationship with seed dormancy
Low seed vigour has often been incorrectly classified as seed with higher dormancy levels when that may not be the case. The differentiation between seed dormancy or low vigour is difficult because seeds in both cases have low germination percentages (Finch-Savage and Bassel. 2016). Seed vigour, like seed dormancy, is quantitative in nature with a large environmental influence on the resulting phenotype (Finch-Savage and Bassel. 2016).
Seed vigour describes the physiological processes related to seed germination and plant growth prior to emergence (van de Venter. 2008). Seed vigour is defined as "the sum total of those properties of the seed that determine the potential level of activity and performance of the seed during germination and seedling emergence" (Perry. 1978;Finch-Savage and Bassel. 2016). Germination speed and uniform plant emergence are qualities contributing to the classification of seed vigour. Variation in seed vigour can be found among the same genotype produced in different maternal environments and under different seed storage conditions (van de Venter. 2008;Finch-Savage and Bassel. 2016). At physiological maturity and shortly thereafter, seed vigour peaks with levels decreasing with subsequent seed handling and storage (Finch-Savage and Bassel. 2016). Factors decreasing vigour include precocious germination, seed age, and storage conditions. Crop domestication has resulted in the selection for fast germinating, large seeded, non-dormant variants of crops, thereby, increasing the vigour of the crop over wild relatives and improving the germination over a wide range of environmental conditions (Finch-Savage and Bassel. 2016). Studies examining seed germination, a component of seed vigour, and seed quality have found weak correlations between seed protein, oil, and germination performance (Hatzig et al. 2018). These results agree with the associations that have been identified between seed dormancy and seed quality components.

Germination
The antagonist process to seed dormancy is seed germination, and, like seed dormancy, germination is highly influenced by genetic, environmental, and the interactions between the two factors (Schatzki et al. 2014). Germination can be described as the absence of dormancy and vice versa. To better understand germination and the genetics governing it in B. napus, a diversity panel of over 200 winter genotypes was screened using a high-throughput phenotyping platform. A genomic region of interest influencing time to 50% germination and final germination rate at 72 h contained a gene known to also play a role in ABA signalling (Schatzki et al. 2014). Further QTL analysis of germination and seed vigour in relation to plant hormone profiles was tested in a winter population formed from a cross between a high vigour and a low vigour parent (Nguyen et al. 2018). Germination traits were recorded 7 and 14 days after sowing, and profiling of ABA, and its metabolites and auxin was performed 5 and 12 days after sowing. Auxin is known to play a role in germination and dormancy and was found to maintain dormancy in seeds due to interactions with the ABA pathway (Nguyen et al. 2018). Sixty-four per cent of phenotypic variation in seed vigour observed in the population was attributed to genetic composition (Nguyen et al. 2018). In total, 12 QTLs overlapped between germination and hormone profiles. Moving forward, the development of molecular markers to select for rapid germination and high vigour breeding lines should be a focus for canola breeders.

Relationship with seed quality profiles in Brassica napus
Intense plant breeding to reduce erucic acid and glucosinolate content (anti-nutritionals) has occurred over the last 50 years in canola to improve the taste and utility of the meal and health of the oil. However, with such intense breeding, a genetic bottleneck has resulted causing low genetic diversity within canola breeding lines. A diversity panel of 215 winter B. napus genotypes with different seed quality profiles, (low erucic; low glucosinolates, low erucic; high glucosinolates, etc.) grown in two contrasting environments were screened for seed vigour traits. The breeding for reduced antinutritionals within the seed has led to an inadvertent reduction in seed vigour and an increase in germination time (Hatzig et al. 2018). Varieties released after the 1980s and classified as canola quality had significantly slower germination time than varieties released pre-1980s (Hatzig et al. 2018). The unintentional reduction in germination time resulting from the selection for reduced anti-nutritionals has been caused by linkage drag. A QTL identified for mean germination time was found to be in the same region as a minor QTL for both erucic acid and glucosinolates (Hatzig et al. 2018). To avoid a similar occurrence from happening when selecting against high dormancy levels, the relationships between dormancy and other seed traits must be fully understood.

Precocious germination
Precocious germination (pre-harvest sprouting or vivipary) is germination of seed while on the mother plant during the maturation phase. Primary dormancy in maturing B. napus seed is prevalent to prevent precocious germination, and primary dormancy levels decrease to negligible in mature seed (Feng et al. 2009). Precocious germination can be a problem in certain genotypes and environments and result in a decrease in yield due to the seed germinating on the mother plant. Precocious germination is a quantitative trait and is influenced by both the environment and genetics as well as the interaction between the two (Ren and Bewley. 1998). Environmental conditions resulting in greater precocious germination levels include excess moisture and high humidity (Feng et al. 2009). Lower concentrations of ABA and changes in osmotic potential are commonly found in precociously germinated seed (Johnson-Flanagan et al. 1991). Heat stress can be another environmental factor that increases precocious germination. This phenomenon is likely caused by the decrease in the ABA: GA ratio, resulting in lower primary dormancy and increased germination (Brunel-Muguet et al. 2015). While precocious germination is low in mature B. napus seed, the modification of secondary dormancy may lead to greater precocious germination prevalence if primary and secondary dormancy are controlled by the same physiological mechanism.

Secondary dormancy screening methodology for Brassica napus
The protocol to screen for secondary dormancy in B. napus is time consuming and labour intensive. Screening for secondary dormancy can take anywhere from 14 to 46 days to complete, depending on the protocol used (Pekrun et al. 1997;Gulden et al. 2004a;Weber et al. 2010). The general procedures are similar across protocols beginning with the dormancy induction, followed by germination and viability steps; however, the duration of the steps varies. Currently, no standardized method exists for secondary dormancy screening in B. napus (ISTA 2008;Weber et al. 2010). The most involved protocol (Pekrun et al. 1997) includes an induction period with the osmoticum polyethylene glycol 6000 extending for 4 weeks at 20 • C in darkness. Followed by germination testing for 14 days at 20 • C, then temperature stratification for any non-germinated seeds, and finally another germination test for viability on any remaining non-germinated seeds. Gulden et al. (2004a) followed the same protocol with the only modification being a tetrazolium test for viability following the temperature stratification period.
The Hohenheim standard dormancy test was developed to reduce the time of the induction period from 28 to 14 days (Weber et al. 2010). The same osmoticum and 20 • C temperature as above were used during the 14-day induction. Following the induction, a germination test for 14 days at 20 • C in darkness was performed. Any non-germinated seeds following the germination test were tested for viability at alternating temperatures of 30 and 3 • C for 12 h/12 h and alternating light and darkness for 12 h/12 h. To further increase the efficiency of screening, the rapid dormancy test was developed from modifications to the Hohenheim standard dormancy test (Weber et al. 2010). The rapid dormancy test reduced the induction and germination period by half to 7 days for each test. When absolute dormancy levels were compared between the Hohenheim standard dormancy test and rapid dormancy test, a correlation coefficient of 0.96 resulted. The Hohenheim standard dormancy test had greater absolute dormancy values compared to the rapid dormancy test, but both tests ranked the genotypes in the same order for dormancy. The rapid dormancy test also eliminated the viability screen and found no effect in doing so on the resulting dormancy values. The sample size was also decreased from 100 to 50 seeds, which resulted in a significant difference between tests for dormancy. The 50 seed test had lower absolute dormancy, but again the ranking of the genotypes between tests remained the same. It was speculated that lower absolute values were due to fewer seeds to imbibe water resulting in a more diluted polyethylene glycol solution for the 50 seed test compared to the 100 seed test. Overall, the rapid dormancy test was determined to be the most efficient screening process.

Conclusion
A combination of genetic, environmental, and agronomic factors influences secondary dormancy within the canola seed. Due to the interactions between the several factors, it is difficult to predict dormancy in any given maternal or seedbed environment. Secondary dormancy can be induced in seed experiencing unfavourable environmental conditions in darkness, whereby a physiological mechanism within the embryo prevents seed germination when conditions may not support a complete life cycle. While this trait is beneficial from an evolutionary perspective, in cultivated crops it can lead to issues with volunteer weeds. Concerns with multiple herbicide resistance, oil profile contamination, and clubroot perpetuation are issues arising from the seed remaining dormant in the seedbank and then germinating and emerging in later years.
The reduction of the genetic potential secondary dormancy in B. napus, while still maintaining current seed vigour and germination standards and preventing increased precocious germination should be a breeding objective for canola breeders. Given that such a significant proportion of the phenotype is controlled by the genetics, and it has been shown to be a heritable trait, selection for reduced dormancy breeding lines within commercial canola breeding programs should be feasible. Furthermore, variation in secondary dormancy across B. napus has been demonstrated in both spring and winter growth habit types. It remains unclear if seed dormancy is the culprit for poor germination, low vigour, and low stand emergence often observed in canola in western Canada. As such, any associations between secondary dormancy and seed germination, vigour, and quality traits should be evaluated to avoid unintentionally affecting other important traits. Another consideration when reducing secondary dormancy in the seed is the effect on primary dormancy levels. It will be important to reduce secondary dormancy while maintaining some level of primary dormancy during the early seed maturation phase to prevent precocious germination. Lastly, the consideration of land management strategies to decrease the presence of the seed in the soil entering through pod shattering and then the management of the seed once it is in the seedbank can also be practised to reduce volunteer concerns. Ultimately, low dormancy genotypes in combination with proper agronomic management are a promising way to decrease volunteer populations in western Canadian fields.
Canada. This work is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.

Data availability
The article does not report data