In recent years, the origin of the hexaploid species Camelina, C. sativa, has been of great interest as this species is defined as an extremely attractive plant platform for genetic engineering research, including research related to lipid metabolism (Nguyen et al., 2013; Faure and Tepfer, 2016; Ghidoli et al., 2023). In addition, the fact that C. sativa is the only allohexaploid in the tribe Camelineae and that this species is a close relative of A. thaliana (both are Brassicaceae Lineage I representatives) makes C. sativa a very attractive model for studying plant polyploidy. Previously, the view of the origin of C. sativa was largely limited to the understanding that this species originated in southeastern Europe and southwestern Asia (Radatz and Hondelmann, 1981), most likely in the regions adjacent to the Black Sea.

The first attempt to understand the structure of the genome and thus to suggest the evolutionary path of C. sativa was made by Gehringer et al. (2006). These authors noted for the first time that amplification of linked SSR markers linked to fatty acid desaturase gene in C. sativa results in the amplification of multiple alleles, suggesting the polyploid nature of this crop. Next steps were made by Hutcheon et al. (2010), who studied FAD2 and FAE1 genes of Camelina species. The authors found that C. sativa and C. microcarpa were potentially hexaploid species, whereas Camelina rumelica was tetraploid. The authors also found that C. sativa and C. microcarpa have two out of the three almost identical copies of genes of common origin, whereas the origin of the third copy was intended for different species with lower ploidy. In addition, the polyploid status of C. sativa was confirmed by Galasso et al. (2011) by comparing the number of β-tubulin gene orthologs in C. sativa and in A. thaliana.

The next major advance in understanding the evolution of Camelina came after the whole-genome sequencing and assembly of the C. sativa genome by Kagale et al. (2014), who confirmed the hexaploid status of this species. Three subgenomes were identified in the C. sativa genome, which had a highly undifferentiated structure and significantly rearranged chromosomes in each of the subgenomes, compared with the ancestral cruciferous karyotype (ACK) and the diploid ancestral karyotype of Camelina (CAM) (Kagale et al., 2014; Mándaková et al., 2019). Additionally, the authors found a phylogenetic relationship between C. sativa subgenomes and lower-ploidy species. The C. sativa subgenome 3 (CsG-3) was found to have higher collinearity with Camelina hispida, whereas the other two subgenomes were more similar to different C. rumelica accessions (Kagale et al., 2014). Later, Galasso et al. (2015) identified a specimen of C. microcarpa with significantly reduced ploidy (2n=12, USDA-NPGS accession ID: PI 650135) and a distinct profile of amplified β-tubulin-based markers. Taking into account the results of molecular analysis and the combination of morphological features, a new diploid C. neglecta was described based on this specimen (Brock et al., 2019), which allowed to include this new species in the genome analysis of representatives of the genus Camelina.

Mándaková et al. (2019) proposed for the first time a detailed model of genome evolution of Camelina species based on the data of genomic hybridization in situ and other methods of comparative genomics. In terms of the proposed model, it was found that C. neglecta contributed to the formation of CsG-1 and Cs-G2, whereas Cs-G3 is more collinear with C. hispida, as previously shown by Kagale et al. (2014). This evolutionary scheme indicates that the karyotype of C. neglecta (n = 6) originated from the ancestral C. neglecta-like karyotype (n = 7). The reduction in the number of chromosomes in modern C. neglecta was achieved by the association (end-to-end translocation) of chromosomes CAM6 and CAM7 (CAM—the ancestral Camelina karyotype), which led to the origin of the Cn6 chromosome of C. neglecta, which is the largest chromosome in C. sativa genome—Cs11 (Cs-G1) ( Table 1 ). Moreover, C. rumelica was found to be an allotetraploid, resulting from hybridization of C. neglecta and C. hispida (Mándaková et al., 2019), which may explain the previously observed by Kagale et al. (2014) high karyotype collinearity between C. rumelica and the third subgenome of C. sativa. It should be noted that for the majority of Camelina polyploids, no signs of translocations between subgenomes were found, which indicates a high stability of the karyotype after interspecific hybridization (Mándaková et al., 2019).

These findings were later confirmed by Chaudhary et al. (2020), who performed genotyping by sequencing of 193 accessions of different representatives of the genus Camelina. The authors found that the subgenomes of tetraploid cytotype of C. microcarpa could only be related to CsG-1 and CsG-2 of C. sativa, which are believed to be the result of hybridization between C. neglecta and ancient C. neglecta-like karyotypes, originally described by Mándaková et al. (2019). Furthermore, the origin of C. rumelica was confirmed by Chaudhary et al. (2020), but the authors noted that its subgenomes did not share the genomic pattern with CsG-1 and CsG-3 of C. sativa, indicating that this species might not be a direct ancestor of C. sativa, which was consistent with previous findings of Mándaková et al. (2019). The karyotype evolution of Camelina species, reported by Mándaková et al. (2019), was also confirmed by Zhang et al. (2020), in which the distribution of ACK genomic blocks in C. sativa chromosomes was reconstructed de novo using A. lyrata as internal reference. The relatively recent genome sequencing of diploid Camelina species (Camelina laxa, C. neglecta, C. hispida var. hispida, and var. grandiflora) confirmed previous findings on the evolution of Camelina genomes and also highlighted the crucial role of these species in the formation of polyploid taxa within the genus (Martin et al., 2022; Chaudhary et al., 2023). Research of Čalasan et al. (2019) and Brock et al. (2020) greatly contributed to understanding of the C. microcarpa distribution and cytotype/ribotype composition of this taxon (hexaploid C. microcarpa is represented by 2n = 40 and 2n =38 cytotypes, types 1 and 2, respectively). A study by Brock et al. (2022a), (2022b) shed light on cytoplasmic inheritance within the Camelina genus, showed the executive role of C. neglecta in the origin of polyploid taxa, and confirmed the close relationship of C. microcarpa (2n = 40) with domesticated C. sativa (2n = 40). Finally, a recent study by Mandáková and Lysak (2022) revealed the nature of the “missing link” in Camelina genus—the tetraploid Camelina intermedia nom. prov. (2n = 26, formerly known as C. microcarpa tetraploid or C. neglecta-like tetraploid).

Dating the evolutionary events that took place within the genus Camelina turned out to be a non-trivial task. First attempts to estimate divergence time of Camelina species were made using 18S ETS sequences to reconstruct phylogeny and date evolutionary events (Čalasan et al., 2019). However, this approach faced several obstacles. In particular, sequencing of a single copy of the ETS from polyploid Camelina species may result in extraction of sequences from a specific subgenome, which may affect the topology of the inferred tree. Čalasan et al. (2019) reported C. microcarpa as a polyphyletic group consisting of several ribotypes, which may, however, be influenced by sequence samples isolated from highly similar N⁶ and N⁷ subgenomes. Similarly, C. hispida was placed together with C. sativa, which was most likely due to extraction of ETS sequences corresponding to the H⁷ subgenome. In addition, Kwiatek et al. (2021) later reported that the number of different rDNA loci may vary within different genotypes of C. sativa and C. microcarpa, which may further complicate such studies. However, to date, Čalasan et al. (2019) reported the most comprehensive study of nuclear marker-based divergence dating of Camelina species. Thus, their result should be taken into account but with the caveat that the datings rather represent the divergence time of (sub)genomes, especially for polyploid taxa.

Another recent study by Brock et al. (2022a) attempted to date evolutionary events within the Camelina genus, based on chloroplast genome phylogeny. The authors used a wider panel of species and achieved dates accurate enough to distinguish cytoplasmic lineages. The majority of the evolutionary events were dated without “subgenome bias”; however, some aspects still remain unresolved. For example, the phylogeny reconstruction, proposed by Brock et al. (2022a) identifies C. rumelica as a polyphyletic group in relation to C. hispida. Similarly, C. microcarpa cytotypes and C. sativa often shared common clades. In view of the described questions, the results of both studies by Čalasan et al. (2019) and Brock et al. (2022a) were taken into account in the evolutionary model, described below.

Based on the recently described findings, here we propose an integrated evolutionary model for the Camelina genus ( Figure 1 ). The proposed model relies on the most recent research on Camelina genetics, genomics, and phylogenetics (Kagale et al., 2014; Kim et al., 2017; Čalasan et al., 2019; Mándaková et al., 2019; Brock et al., 2020; Chaudhary et al., 2020; Zhang et al., 2020; Chaudhary, 2021; Brock et al., 2022a; Brock et al., 2022b; Martin et al., 2022; Chaudhary et al., 2023).

It has been suggested that the divergence of the ancestral Camelina karyotype (the last common ancestor for the genus) occurred approximately 2.5 ± 1 Mya (Čalasan et al., 2019; Brock et al., 2022a) as the time of Camelina speciation among other Camelineae representatives. The next major evolutionary event, the divergence of C. laxa species from the other Camelina lineage, occurred approximately 2 ± 1 Mya, according to the ETS-based dating (Čalasan et al., 2019). Chloroplast-based dating suggests that the diversification of C. laxa cytoplasmic lineage took place approx. 1.62 Mya (estimated dating for most recent common ancestor of Camelina spp.) (Brock et al., 2022a) ( Figure 1 ). The basal status of C. laxa and its significant genetic differences from other Camelina sp. have also been confirmed by numerous studies (Hutcheon et al., 2010; Kagale et al., 2014; Brock et al., 2018; Mándaková et al., 2019; Chaudhary et al., 2020; Martin et al., 2022).

Most studies show that C. hispida was the next species to diverge from the genus along with others (Hutcheon et al., 2010; Kagale et al., 2014; Brock et al., 2018; Mándaková et al., 2019; Brock et al., 2022a; Martin et al., 2022). Instead, results of Čalasan et al. (2019) showed that C. hispida belongs to the same clade as C. sativa and Camelina alyssum, which is most likely caused by the above-described “subgenomic bias” of the analyzed ETS sequences. Thus, we suggest that the diversification event of C. hispida ( Figure 1 ) occurred approximately 1.5 ± 1 Mya, if dated by the divergence of ETS sequences (Čalasan et al., 2019). Chloroplast-based dating also suggests that speciation of C. hispida occurred 1.2–1.1 Mya (Brock et al., 2022a). It is also worth noting that C. hispida has highly shattered chromosomes, if compared with CAM (Mándaková et al., 2019; Martin et al., 2022). Furthermore, this species is the only Camelina diploid that has infraspecific taxa, namely, var. hispida, var. grandiflora, var. lasiocarpa, and var. steifelhagenii.

A recent study of Martin et al. (2022) revealed genomic differences between C. hispida var. hispida and var. grandiflora. Divergences between these infraspecies may be a potential explanation for the polyphyly of C. hispida, observed by Brock et al. (2022a) in their study. The ploidy and phylogeny of the other two infraspecies (var. lasiocarpa and var. steifelhagenii) remain a mystery; therefore, the divergences between these varieties are not shown in Figure 1 . It should be noted that C. laxa and C. hispida are the only self-incompatible species of the genus Camelina (the status of Camelina anomala is currently unknown) (Čalasan et al., 2019). At the same time, another diploid taxon, C. neglecta, is a facultatively self-pollinated plant, similar to C. sativa, C. alyssum, and C. microcarpa.

The evolutionary pathway of C. anomala cannot currently be adequately described due to the lack of experimental data. This species was not found for a long time because of its very limited geographical distribution (Brock et al., 2018). Consequently, the lack of fresh plant material or seeds of C. anomala significantly limits its use in genomic studies. The chromosome number and ploidy status of this species are currently unknown. Previously, C. anomala was discussed as an ancestral species for the whole Camelina genus due to its atypical morphology (Mirek, 1984). On the contrary, the results of Čalasan et al. (2019) suggest that C. anomala may be the closest relative of C. hispida, which diversified probably less than 0.3 Mya, whereas the data of Brock et al. (2022a) suggest that C. anomala is basal species for C. hispida–C. rumelica lineage and its speciation event occurred much earlier—approx. 0.7–0.5 Mya ( Figure 1 ). C. anomala has a robust habitus, particularly elongated silicles (alike Brassica sp.) (Mirek, 1984), which makes this species a very attractive target for research, given the possibility of introducing these traits into the cultivated C. sativa to promote its seed productivity and overall yields. However, due to the unknown status of ploidy, it is not yet possible to draw a definitive conclusion on the possibility of interspecies hybridization of C. anomala.

As shown in Figure 1 , according to chloroplast-based datings, the lineages of C. neglecta and C. hispida diverged approximately 1.2–1.1 million years ago (Brock et al., 2022a). Unfortunately, Čalasan et al. (2019) did not include C. neglecta in their ETS-based phylogeny reconstruction, so no nuclear marker-based dating is available for this species. The ancestor of C. neglecta is thought to have contained seven chromosomes (C. neglecta-like, n = 7, N⁷). To date, no living populations of the C. neglecta-like (n = 7, N⁷) ancestor have been found in nature (Mándaková et al., 2019; Brock et al., 2022a; Mandáková and Lysak, 2022).

Unlike other Camelina diploids, the date of origin of C. neglecta from its presumed C. neglecta-like (N⁷) ancestor (Mándaková et al., 2019) is currently unknown ( Figure 1 ). C. neglecta (n = 6; N⁶) is thought to have descended from its ancestor of a higher chromosome number as a result of descending dysploidy (Mandáková and Lysak, 2018), sometimes called “mixoploidy” in Brassicaceae species (Kunakh et al., 2008). Signs of this evolutionary event are evident in the presence of the so-called fusion chromosome (Cn6, homolog in C. sativa—Cs11), which is one of the largest chromosomes among all Camelina sp. karyotypes (Mándaková et al., 2019; Chaudhary et al., 2020; Zhang et al., 2020). It is noteworthy that C. neglecta is characterized by the smallest genome size within the genus (Chaudhary et al., 2020; Martin et al., 2022; Chaudhary et al., 2023).

As reported by Brock et al. (2019), C. neglecta is currently found in one location, so only little is known about its distribution and wild population structure. Potentially, reported findings of n = 6 C. rumelica in Hungary and the USA and 2n = 12 C. sativa (Warwick and Al-Shehbaz, 2006) may correspond to unidentified specimens of C. neglecta (Brock et al., 2019), which needs to be confirmed. It was later supposed that C. neglecta may be distributed throughout the Eurasian steppe (Brock et al., 2022a). In addition, when collecting Camelina sp. in Turkey, Brock et al. (2018) identified a potentially diploid accession (JRB30 in original paper), which had a genome size of 370 Mbp, which is very close to C. laxa and C. hispida. However, this species is grouped in the C. rumelica clade, as a basal branch, apart from the diploid Camelina clades, so potentially this accession may correspond to a C. neglecta-like species (morphological data did not support this) or another unidentified diploid species, probably, close to C. hispida ( Figure 1 ).

As noted above, C. hispida is known to have several infraspecies, including var. hispida and var. grandiflora. As in the case of C. rumelica, the status and recognition of these varieties are still a matter of debate. However, it has been recently shown that these varieties differ in genome size: smaller in C. hispida var. grandiflora—0.65 ± 0.03 pg—and greater in C. hispida var. hispida—0.73 ± 0.02 pg (Wu, 2016). This opens a valid reason for further revision of the taxonomic structure within the C. hispida, as well as to identify which of the known C. hispida varieties is closer to the H⁷ subgenomic component of C. sativa ( Figure 1 ).

For instance, Chaudhary et al. (2020) investigated only C. hispida var. grandiflora (given the taxonomy of the accession, used in the mentioned study), which could potentially have a smaller genome size (0.59 ± 0.02 pg in this study), and found that its genome structure is not as highly collinear with Cs-G3, as previously reported by Mándaková et al. (2019) for C. hispida. On the contrary, Mándaková et al. (2019) and Brock et al. (2020) most likely used samples of C. hispida var. hispida in their genomic studies (taking into account the taxa recognition, made by the authors) collected in the center of origin (Turkey). Thus, there is a very high probability that C. hispida var. hispida is the exact progenitor of Cs-G3, but not the C. hispida var. grandiflora.

Recent results by Martin et al. (2022) strongly support these assumptions. It was demonstrated that C. hispida var. hispida maintains a more similar genome structure to the third (H⁷) subgenome of C. sativa than C. hispida var. grandiflora. The latter infraspecies shows signs of several unique translocations and has a significantly lower percentage of transposable elements (TE), than the genome of C. hispida var. hispida. This fact may be a potential explanation for the previously observed difference in relative genome size between these two taxa (Wu, 2016). At the same time, the subgenome H⁷ of C. sativa and the genome of C. hispida var. hispida have a similarly large number of TEs and also share a TE location in syntenic regions almost twice as often as H⁷ and C. hispida var. grandiflora do (Martin et al., 2022).

Moreover, such evidence of genome-wide differences between C. hispida var. grandiflora and C. hispida var. hispida could be an additional issue for revision of their taxonomic status and probably the further rehabilitation of these taxa at the species rank. The taxonomy of two additional varieties of C. hispida (var. lasiocarpa and var. stiefelhagenii) is also discussed by Mutlu and Karakuş (2019). Acceptance of these two varieties of C. hispida as separate species is also under debate (Al-Shehbaz and Barriera, 2019), as well as their role in the genus evolution. Further involvement of molecular genetic analysis could help facilitate this investigation.

The vast majority of studies indicate extremely limited genetic variability in C. sativa accessions (Galasso et al., 2015; Singh et al., 2015; Luo et al., 2019; Blume et al., 2020; Chaudhary et al., 2020). This is well explained by the evolutionary history of C. sativa, as this crop has encountered at least three major bottlenecks limiting its genetic diversity. The first of such events of narrowing genetic diversity occurred during the formation of tetraploid C. intermedia, the second during its hybridization with C. hispida, which led to the emergence of hexaploid C. microcarpa type 1, and, finally, the third, when domesticating wild C. microcarpa and further breeding C. sativa crop. During all of these three stages, only a limited number of genotypes or/and populations have participated in allopolyploidization, crucially narrowing the diversity of descendant species. These bottlenecks are marked in the evolutionary diagram, shown in Figure 1 .

A significant narrowing of the genetic variability of C. sativa was also caused by self-pollination of this species and its ancestors (except C. hispida). Kim et al. (2017) noted that the paucity of genetic diversity of cultivated C. sativa was caused not only by strong directional selection in the past (purifying selection, including wild progenitors), which may have occurred during domestication of this crop and its subsequent selection, but also by the high level of self-compatibility of C. sativa (and its ancestors, as we know now), combined with high selective pressure during the evolution this taxa. Most extant C. sativa cultivars have a very low heterozygous rate, indicating that this is due to the inbred nature of this crop, resulting from self-compatibility and common self-pollination (Manca et al., 2012; Blume et al., 2020).

Finally, the abandonment of C. sativa cultivation during the twentieth century (due to the predominant cultivation of rapeseed, sunflower, etc., in Europe) led to loss of the existing varietal diversity of C. sativa (European Commission, 2017). However, regions of previously active cultivation of C. sativa may have contained a fraction of varietal/landraces in the wild. For instance, in Ukraine, which was the place of widespread cultivation of C. sativa until the 1940s–50s (Iljinska et al., 2007; Blume et al., 2022), wild (escaped from fields) populations of this crop and of C. alyssum can be found (Brock et al., 2020). Previously, Ghamkhar et al. (2010) reported that the Eastern European region (in particular, Ukraine) may be a hotspot for the hexaploid Camelina genetic diversity. Sometimes, such wild populations of C. sativa can be found across the European Plain, the Caucasus, the Caspian region, and the eastern part of Central Asia (Čalasan et al., 2019).

The limited genetic variability among C. sativa accessions is apparently a result of the gradual narrowing of genetic diversity described above. This fact along with the limited number of accessions available in gene banks significantly limits the efficient propagation of this emerging biofuel crop. However, there are some possibilities to overcome genetic paucity of C. sativa, particularly introgression of genes from wild populations of the species or from its wild relatives, as well as resynthesis of this crop from its potentially ancestral species, enabled by the understanding of the genus evolution. These possible approaches will be discussed below.

Wild Camelina species can be an extremely valuable source of germplasm for overcoming the limited genetic variability of allohexaploid C. sativa. Although it has been described above that the limited diversity of C. sativa is largely due to the self-pollination nature of this plant, it can be expected that self-incompatible diploids may show much higher levels of genetic polymorphism. Galasso et al. (2015) observed no genetic variation in individual plants of a particular lineage/affinity of C. sativa, C. microcarpa, C. alyssum, and C. rumelica. In contrast, C. hispida (var. grandiflora) had the highest level of genetic polymorphism within one line, whereas C. laxa had a lower, but still very significant, variation among individuals within a particular line. Such difference in levels of genetic polymorphism is well consistent with the self-incompatible nature of these two species. Additionally, Chaudhary et al. (2020) noted that wild tetraploid and, especially, diploid Camelina species have greater genetic diversity than any of the extant hexaploids, even more than any of C. microcarpa cytotypes. From this perspective, Camelina diploids appear extremely attractive as germplasm donors for C. sativa improvement.

Unfortunately, none of the diploid species can be successfully crossed directly with cultivated C. sativa. Martin et al. (2019) reported the inability to form hybrids between C. neglecta (referred to as diploid C. microcarpa) and C. sativa. Crossing of C. laxa and C. hispida with C. sativa resulted in production of several silicles on hybrids containing no seeds (Zhang and Auer, 2020). These results suggest that the direct hybridization between Camelina wild diploids and cultivated hexaploids is rather impossible. Potentially, wild-type alleles from these species could be introgressed into C. sativa using the tetraploid species as an intermediate acceptor for such alleles of interest. A diagram, showing the cross-compatibility of different Camelina species is presented in Figure 2 .

Using Camelina tetraploids for gene introgression may be a more meaningful option, as they can hybridize with C. sativa more successfully. The most obvious candidate for such crossing is a direct maternal progenitor of this crop—C. intermedia (tetraploid form of C. microcarpa). To date, only one study has investigated the possibility of such interspecific crossing (Martin et al., 2019). It was established that tetraploid C. intermedia is able to form stable hybrids with C. sativa, but at low rates (0.009 hybrids per flower pollinated) ( Figure 2 ). Such hybrids showed reduced pollen fertility and produced limited seeds when backcrossed or selfed. Interestingly, backcrossing C. sativa × intermedia with C. intermedia tended to produce more seeds (Martin et al., 2019). Such low fertility of the resulting hybrids can reduce the efficiency of gene introgression into cultivated Camelina. In addition, it has been reported that the tetraploid form of C. microcarpa (≡C. intermedia) can coexist with hexaploid C. microcarpa in the wild, suggesting that there is a natural gene flow between these two taxa in nature (Martin et al., 2017). These facts suggest that potential production of C. sativa × intermedia hybrids may be somewhat successful, albeit potentially labor-intensive, in restoring normal fertility.

Although C. rumelica did not play an important role in the formation of C. microcarpa/sativa lineage ( Figure 1 ), this tetraploid species is still of great interest for further research. This is especially true, taking into account that two C. rumelica infraspecies, subsp. rumelica and subsp. transcaspica, demonstrate different abilities to hybridize with C. sativa ( Figure 2 ). Seguin-Swartz et al. (2013) reported that C. rumelica subsp. rumelica was able to hybridize with female plants of C. sativa and form fertile progeny. Such hybrids were reported to be highly sterile but viable and able to self-produce. At the same time, C. rumelica subsp. transcaspica failed to hybridize with C. sativa or the hybrid seeds formed were non-viable. In contrast, Zhang and Auer (2020) state that C. rumelica (most likely subsp. rumelica) failed to produce any hybrid seeds after crossing with C. sativa and in backcrosses.

A study by Zhang et al. (2022) confirmed findings of Seguin-Swartz et al. (2013). Authors of the latter study reported that hybrids of C. rumelica × sativa and C. sativa × rumelica can be obtained with almost similar success rates (1 hybrid per 217–220 ovules pollinated). It was also shown that the pollen of such hybrids had reduced viability (less than 2%), whereas the F₂ progeny could only be obtained by selfing, but not backcrossing, which mirrors and is consistent with the results of Seguin-Swartz et al. (2013) on C. rumelica subsp. rumelica. Zhang et al. (2022) also reported that hybrids with reduced seed productivity could inherit the winter-type life cycle from C. rumelica with a probability of 0.5%.

Potentially, such differences in the viability of hybrid progeny can be caused by different levels of unequal pairing of chromosomes and different chromosomal organization within subgenomes of the mentioned subspecies (e.g., the N⁶ subgenome of C. rumelica was completely rearranged). Nothing is known about the genome organization of C. rumelica subsp. transcaspica, so its complete inability to hybridize with C. sativa can hardly be explained. Additionally, nothing is known about the ability of C. anomala to hybridize with other species, nor about its genome and ploidy. Summarizing the given above, hybridization of C. sativa with tetraploid relatives may be an acceptable option but has limited success and may require further arduous breeding for fertility restoration.

The possibility of gene introgression from C. microcarpa is more often discussed, since this species is the closest relative of cultivated camelina ( Figure 1 ). Crossing of C. sativa with its closest ancestor may allow the restoration of higher genetic diversity that was lost during domestication. Recently, C. sativa has been confirmed to lack a number of alleles present in C. microcarpa type 1 and type 2 genotypes during domestication (Chaudhary et al., 2020). However, since this wild species consists of two distinct cytotypes with the different genome organization, hybridization success may depend on the ploidy/chromosomal number of the C. microcarpa sample used.

The first records of interspecific hybridization of C. microcarpa and C. sativa were published by Tedin (1925), who suggested that cross was likely to be successful but that the progeny had a high level of sterility and possessed a dwarf phenotype. At that time, the chromosome count of C. microcarpa was obviously not taken into account. More recently, Seguin-Swartz et al. (2013) reported that hybridization success with C. sativa greatly depends on the particular C. microcarpa genotype used. The number of chromosomes of the accessions used in the study was not reported, but the authors noted that the genotypes differed in genome size. Later, Chaudhary et al. (2020) reported that C. microcarpa type 1 and type 2 may have differences in relative genome size.

Zhang and Auer (2020) reported that C. sativa and C. microcarpa (and vice versa) could produce hybrid progeny, suitable for backcrossing. The authors used C. microcarpa accessions PI633186 and PI633190, the first of which had been reported earlier to belong to type 2 (2n = 38), according to Chaudhary et al. (2020). Unfortunately, Zhang and Auer (2020) did not report any chromosomal rearrangements or fertility reduction in the hybrid progeny.

Recently, Tepfer et al. (2020) reported that C. microcarpa (accessions collected in France) were not able to form stable and fertile hybrids with C. sativa. F₁ C. microcarpa × C. sativa hybrids had various meiotic abnormalities (presence of univalents, bridges, and chromosome fragments), which led to an abnormal phenotype in the hybrid progeny and led to a decrease in the fertility of interspecific hybrids. Tepfer et al. (2020) investigated the approximate ploidy of the studied samples and concluded that the number of chromosomes was “approximately” n = 20 (the authors did not determine the exact number of chromosomes). However, provided by Tepfer et al. (2020), microphotographs may indicate that C. microcarpa most likely had n = 19 chromosomes. Taking into account the results of Chaudhary et al. (2020), it can be assumed that the studied C. microcarpa of French origin, used by Tepfer et al. (2020), could belong to type 2, which experienced multiple genome restructuring events.

Recently, Chaudhary (2021) confirmed that hybridization of type 2 C. microcarpa with C. sativa (followed by C. sativa genome-type species, C. alyssum) leads to unequal recombination between chromosomes and significant chromosomal rearrangements, including the loss of certain genomic regions or resulting chromosomes. However, the example of this study shows that backcrossing such hybrids with a 40-chromosomal relative could help to restore fertility and produce stable hybrids that may be used in breeding.

On contrary, hybridization between 2n = 40 C. microcarpa type 1 and C. sativa can be highly successful ( Figure 2 ). Martin et al. (2019) reported that such a cytotype can be crossed with C. sativa and viable and fertile hybrid progeny can be obtained. However, such hybrids will still show reduced pollen viability (only 17%) and decreased seed productivity. Seed size, plant height, and thousand-seed weight in hybrids were decreased and represented the average of two parental species. The possibility of the gene flow between C. microcarpa and C. sativa is supported by findings of mixed populations in the nature (Martin et al., 2017).

C. microcarpa type 2 was found to have the highest number of minor alleles in the first subgenome (derived from C. neglecta) and the highest number of unique minor alleles (over 1000), compared with C. microcarpa type 1 and C. sativa. At the same time, the third subgenome (H⁷) of C. microcarpa type 1 has more such minor alleles than the third subgenome (N⁶-type) of type 2 does. Some of these minor alleles may be of interest for C. sativa breeding, so introducing them from both types of C. microcarpa appears to be a promising approach to overcome the limited genetic diversity of false flax crops. Hopefully, there are a number of studies that reveal the association of agronomically important traits with certain loci in the C. sativa genome, which has allowed the identification of genes, involved in the shaping of the productivity of this crop (Gehringer et al., 2006; Li et al., 2021).

In addition, populations of C. microcarpa and wild C. sativa can be screened for useful allelic variants for further gene introgression. Brock et al. (2020) found that C. microcarpa type 2 populations of Ukrainian origin may contain more genetic diversity than type 1 representatives, collected from the Caucasus (potentially the entire range of type 1 distribution). The latest results also indicate that the Ukrainian population of C. microcarpa has a certain genetic heterogeneity (Sakharova et al., 2023). This is consistent with previous findings that reported high genetic diversity of hexaploid Camelina sp. in Eastern Europe (Ghamkhar et al., 2010). However, the cytotype of C. microcarpa should always be taken into account because of the potential difficulties, associated with crossing of hexaploid Camelina with different chromosome numbers. Currently, no tools have been proposed for efficient and high-throughput recognition of C. microcarpa cytotypes (e.g., DNA-barcoding tools), but some studies have been conducted to address this problem (Galasso et al., 2015; Brock et al., 2022a; Mandáková and Lysak, 2022; Sakharova et al., 2023).

Improvement of C. sativa via interspecific crossing may not be limited to hybridization with C. microcarpa, but other hexaploid species/forms should be considered ( Figure 2 ). For instance, C. alyssum has several beneficial agronomic traits, including significantly enlarged seeds and seed pods. In addition, all available studies show that C. alyssum is fully interfertile with C. sativa (Tedin, 1922; Seguin-Swartz et al., 2013; Zhang and Auer, 2020; Chaudhary, 2021). Moreover, the intermediate phenotype of C. alyssum x sativa hybrids might not have the reduction in seed or pod size that would be expected in the case of C. microcarpa × sativa hybrids. This would significantly simplify further breeding and selection of hybrids for higher plant productivity.

Another option could be the crossing of cultivated C. sativa with its winter forms or isolated populations of potentially separate infrataxa—subsp. pilosa—as such winter plants were found to be genetically distinct from majority of existing spring varieties of C. sativa (Kim et al., 2017; Chaudhary et al., 2020). In this case, no difficulties are expected, associated with chromosome instability in hybrid progeny or with reduction of plant productivity. In addition, it has recently been shown that hybrids between winter and spring forms of C. sativa will have an intermediate phenotype with prolonged vegetation cycle (terms are intermediate between those of the parental genotypes) and often do not require vernalization (Chaudhary, 2021). Crossing genetically distant varieties or obtaining doubled haploid lines from hybrids can be an additional considerable option to increase C. sativa genetic diversity (Zelt and Schoen, 2016; Sadeghikian et al., 2022).

However, the use of these approaches is limited by the number of Camelina sp. accessions, available in major gene banks (USDA-ARS, IPK Gatersleben, PGRC, etc.). For example, the diversity of C. alyssum is represented with only a few samples (1–3, depending on the gene bank), which may not be enough for effective breeding. From this point of view, the possibility to use plant germplasm (including wild populations of C. sativa) collected from the wild nature should be considered. Luckily, recent research on the diversity of hexaploid Camelina species offer information on the distribution of these species and the possible hotspot of the diversity of these species. The considerable candidates for interspecific hybridization with C. sativa are summarized in Figure 3 , as well as a proposed potential approach for the resynthesis of C. sativa which is described in the next section.

Another interesting approach that is currently being widely discussed is the resynthesis of allohexaploid C. sativa (Brock et al., 2019; Brock et al., 2022a; Martin et al., 2022). The proposed integrated evolutionary model of the Camelina genus ( Figure 1 ) made it possible to clarify the most important questions of the origin of hexaploid species, in particular the origin of C. sativa, enabling further research toward the successful resynthesis of this complex allohexaploid from its diploid and tetraploid relatives ( Figure 3 ). At the current stage, the creation of a synthetic hexaploid Camelina based on the existing tetraploid form of C. intermedia (N⁶N⁷) (Mandáková and Lysak, 2022) and diploid species C. hispida (H⁷) seems to be the most promising in particularly using var. hispida, as it was defined as the most likely exact progenitor of the third subgenome (Mándaková et al., 2019; Martin et al., 2022). There is a high probability that the use of another var. of C. hispida for resynthesis may lead to instability of hybrid progeny of synthetic and natural C. sativa. C. hispida var. grandiflora maintains much lower collinearity with the karyotype of Cs-G3, which may lead to problems similar to those observed in hybridization between C. microcarpa type 2 and C. sativa. It is also important that during such resynthesis the tetraploid C. intermedia should be a maternal plant to keep proper cytoplasmic inheritance, as it was during the evolution of C. sativa ( Figure 1 ). Cytoplasmic inheritance can affect further hybrid progeny ability to self-pollinate or be compatible with the wild natural C. sativa, since, for example, mitochondrial genes play an extremely important role in self-pollen acceptance or rejection (García-Valencia et al., 2017).

A successful example of creating synthetic hexaploids within the Brassica genus by crossing tetraploid and diploid species suggests that the proposed approach for Camelina may also be highly realistic (Gaebelein et al., 2019). In addition, the use of synthetic Brassica polyploids proved to be an extremely efficient tool for rapid boost of phenotyping variation, which can significantly promote breeding (Zhang et al., 2016).

Since diploids cannot be directly crossed with hexaploid C. sativa, the only possible way to exploit the diversity of the highly polymorphic germplasm of these species is to use intermediate acceptor of either a resynthesized hexaploid, or naturally occurring tetraploid (Eeckhaut et al., 2006). As the C. neglecta cannot hybridize with C. sativa directly, it could be firstly crossed with tetraploid C. microcarpa, which can then be hybridized with hexaploid C. sativa. The same approach can be applied to C. hispida and C. rumelica. However, in it is currently unknown whether this approach will be successful, since no study has reported the ability of diploid and tetraploid Camelina species to form hybrids. Tetraploid C. microcarpa and C. rumelica subsp. rumelica may be of particular interest as donors of the wild germplasm diversity ( Figure 3 ). The recently discovered wide range of geographical distribution of tetraploid C. microcarpa (≡C. intermedia) (Čalasan et al., 2019; Brock et al., 2022a; Mandáková and Lysak, 2022) suggests that some populations of this cytotype may be sufficiently genetically distinct to be considered as a potential source of allelic variation. The reported ability of tetraploid C. microcarpa to form stable hybrids with C. sativa (Martin et al., 2019) greatly expands the possibilities of introgression of the wild gene pool into cultivated C. sativa. However, such hybrids will have an intermediate phenotype between the parental species, which may result in reduced seed productivity of such hybrids.

C. rumelica is also able to form fertile hybrids with C. sativa, but less successfully: only as a paternal plant when crossed with C. sativa. Another problem is the high sterility of such hybrid progeny, which leads to a poor ability to produce seeds in subsequent back- or self-crossing. It is likely that separate evolutionary histories (different allopolyploidy trajectories and distinct cytoplasmic inheritance) set barriers to efficient interspecific hybridization of C. rumelica with C. sativa. However, this research direction is of considerable interest. Since C. rumelica plants are often found in regions with arid climate (Brock et al., 2018; Brock et al., 2020) or on the shores of seas or salt lakes (Iljinska et al., 2007), they may have a high level of tolerance to abiotic stresses, associated with increased salinity or drought. The introgression of such traits in C. sativa is of considerable interest, as the yields of this crop are significantly limited by water scarcity in arid regions (Estakhr and Ranjbar, 2021; Gore and Kurt, 2021; Neupane et al., 2022). However, this topic requires further research, especially regarding the incompatibility of C. sativa and C. rumelica subsp. transcaspica, which may be caused by genes involved in self-incompatibility reactions, as observed in other species (Morimoto et al., 2020). Crossing the two tetraploids 4x C. microcarpa and C. rumelica should also be considered, as a potential hybrid could benefit from crossing wild germplasm with C. sativa. However, according to the reported results of such crossing within Brassica genus (Katche et al., 2021), such a hybrid can potentially be tetraploid, carrying complete N⁶ genome and the N⁷-H⁷ genome association, which may cause chromosomal instability in progeny from crossing of such a hybrid with conventional C. sativa. This approach requires additional research but can still be viewed as an additional opportunity for improvement of genetic variation within C. sativa.

Intergeneric hybridization of C. sativa with other species can also be considered as a possible approach to overcome the genetic paucity. However, to date, no significant success toward the obtaining of such intergeneric hybrids has been achieved. Crossing of C. sativa with Neslia paniculata—the closest relative of Camelina genus, resulted in an extremely low number of F₁ hybrids. When C. sativa was a maternal plant in the crossing, C. sativa x N. paniculata hybrids showed very low fertility and were unable to form viable progeny in self- or back-crosses (Martin et al., 2020). If N. paniculata was a maternal plant in the cross, neoautotetraploids of N. paniculata were formed.

Hybridization of C. sativa with more distant Brassicaceae species was also unsuccessful. Several studies reported that C. sativa was able to hybridize with Capsella bursa-pastoris (Séguin-Swartz et al., 2011; CFIA, 2012; Julié-Galau et al., 2014; Martin et al., 2015; Zhang and Auer, 2020). Obtained hybrids were viable and had intermediate phenotype between parental species, especially the shape of seedpods. However, hybrid progeny was sterile and unable to produce any seeds in self- and back-crosses (Julié-Galau et al., 2014; Martin et al., 2015). A similar outcome was observed for hybrids between C. sativa and Capsella rubella (Séguin-Swartz et al., 2011; CFIA, 2012). Potentially, such hybrids could be used for further research, including breeding, but only if the fertility is restored (Martin et al., 2015). Attempts to obtain crosses between C. sativa and even more distant relative A. thaliana brought no seeds, whereas hybridization with Cardamine hirsuta appeared to be more successful—several hybrid seeds were produced, albeit non-viable (Julié-Galau et al., 2014).

Crossing of C. sativa with other Brassicaceae, like Brassica rapa, Brassica nigra, B. napus, B. juncea, and Thlaspi arvense, was also unsuccessful (Séguin-Swartz et al., 2011; CFIA, 2012). Only little progress was achieved in somatic hybridization of C. sativa with B. napus, which resulted in producing hybrids of intermediate phenotype, increased content of linolenic fatty acid, and an ability to backcross with B. napus (Jiang et al., 2009). Somatic hybridization between C. sativa and Brassica carinata or with Brassica oleracea was unsuccessful (Narasimhulu et al., 1994; Hansen, 1998; CFIA, 2012).