CYP gene sequences and structures were retrieved directly from the curated genome annotations available in the Gbrowser databases (SpottedWingFlyBase, SpottedWingFlyBase, 2013, and FlyBase, FlyBase, 1993), using Annotation Release v1 for D. suzukii and Annotation Release v3 for D. melanogaster (Supplementary Tables S1-S4). Only protein-coding CYP genes were included in the analysis. Because these genes were obtained from assembled annotations rather than from de novo searches, no overlapping or ambiguous hits were encountered. As transposable elements (TEs) in flanking regions can provide novel transcriptional regulatory signals, we extracted 10 kb upstream (5′ flanking region) and 10 kb downstream (3′ flanking region) of each CYP gene based on the annotated gene coordinates. The retrieved sequences were visually inspected in Gbrowser to compare genomic features between species.

Gene length comparisons between D. suzukii and D. melanogaster were performed and visualized in R using the genoPlotR package (Guy et al., 2010). In this package, each orthologous gene is plotted side-by-side to allow direct structural comparison, enabling visualization of local synteny among the species such as exon–intron organization, total gene length, and the positions of TE insertions. All graphical outputs were refined in Inkscape v0.92.1¹.

For phylogenetic context, we incorporated the maximum-likelihood phylogeny generated by Chiu et al. (2013) into our comparative analyses. We further expanded these analyses to include orthologous CYP genes from two sister species of D. suzukii, Drosophila biarmipes and Drosophila takahashii (Annotation Release 101 for both species), and screening these orthologs for TE insertions (Supplementary Tables S5, S6) using the same RepeatMasker-based pipeline described below to ensure robust cross-species comparisons.

To detect the presence of TEs associated with CYP genes, we analyzed each gene sequence and, separately, 10 kb of its 5′ and 3′ flanking regions using the RepeatMasker web server (RepeatMasker Open-4.0 Software, 1996). Searches were conducted with the parameters: crossmatch, fruit fly, and GC-level–based matrix. TE classification was assigned according to the highest-scoring match using the Drosophila reference library from Repbase (Jurka et al., 2005). Because the “fruit fly” RepeatMasker library is optimized for D. melanogaster, it may therefore underestimate lineage-specific or recently diverged TE families in D. suzukii.

Based on the RepeatMasker output, we classified TE fragments according to their genomic location relative to CYP genes as: (i) intronic, (ii) 5′ flanking (within 10 kb upstream of the transcription start site), or (iii) 3′ flanking (within 10 kb downstream of the transcription termination site). TE sequences located within CYP genes and 10 kb of its flanking regions were further analyzed to identify putative transcription-factor binding sites (TFBS). Strand-specific predictions were performed using the ConSite web server (ConSite Software, 2004) with the JASPAR CORE Insecta database (Bryne et al., 2008) for D. melanogaster, applying a 90% TFBS cutoff score, following the methodology of Carareto et al. (2013).

To evaluate whether the number of TE insertions in CYP genes was proportional to the overall genomic TE composition of the studied species, Illumina reads were obtained from the Sequence Read Archive (SRA): D. suzukii – SRR942805 (North American sample; Chiu et al., 2013), and D. melanogaster – SRR1738161. Graph-based clustering of NGS reads was performed using RepeatExplorer (Novák et al., 2013) on the Galaxy-based web server, following the pipeline described by Silva et al. (2016). This analysis provided genome-wide estimates of TE content and the relative contribution of different TE superfamilies, including Helitrons, which were then compared with the proportion of TE and Helitron copies overlapping CYP genes.

We first compared the size distributions of CYP genes and non-CYP genes within each species. For this purpose, a dataset of 500 additional genes was randomly selected from each genome. For D. suzukii, random genes were sampled using BEDTools v2.27.0 (Quinlan, 2014), and their orthologs in D. melanogaster were subsequently identified. Median gene lengths were used instead of means due to the asymmetrical distribution of gene sizes. Within each species, CYP gene lengths were compared with the lengths of the 500 randomly selected genes using the Mann-Whitney non-parametric test.

To account for differences in overall genome size and gene-length distributions between species, we normalized CYP gene lengths by the median size of the 500 randomly selected genes in each species. Normalized gene size was calculated as the length of each gene (in base pairs) divided by the species-specific median length of the 500 randomly selected genes. A Wilcoxon signed-rank test was then applied to compare normalized CYP sizes between orthologous CYPs of D. suzukii and D. melanogaster.

TE enrichment was compared between D. suzukii and D. melanogaster using four approaches: (1) the frequency of TEs within genes, including their flanking regions; (2) the frequency of TEs within CYP genes and flanking regions; (3) the frequency of TEs in CYP genes versus its frequency in other genes in the genome; and (4) the frequency of Helitron insertions in CYP genes versus its frequency in non CYP genes. The comparisons were performed using Chi-square tests with 1 degree of freedom and Yates’ continuity correction, which is usual in these cases. All annotated genes and intergenic regions in each genome were considered for these comparisons. Statistical analyses were conducted using SPSS^® version 18, and a significance threshold of p ≤ 0.05 was applied.

Among the 76 CYP genes annotated for D. suzukii, 42 contained transposon sequences within or near the genes (Figure 1; Table 1; Supplementary Tables S1, S2). In D. melanogaster, 41 of the 91 genes analyzed harbored such insertions (Figure 1; Table 1; Supplementary Tables S3, S4). Although D. suzukii has fewer CYP genes overall, we detected a higher number of TE fragments in these genes (140 vs. 136). Most of this difference is mainly attributable to a markedly higher number of Helitron elements in D. suzukii (118 fragments; 84% of all TE insertions, Table 1) compared with D. melanogaster (32 fragments; 24%, Table 1). Other TE subclasses also differed in proportion between species (Table 1), with D. melanogaster showing 51% of LTR retrotransposons (69 fragments). Nevertheless, the absolute difference in total TE fragment counts between the two species is primarily explained by the excess of Helitron insertions in D. suzukii (Figure 1; Table 1).

In flanking regions, TEs were detected both upstream and downstream of CYP genes, and were also detected within introns in both species (Figure 2). D. suzukii displayed more TE insertions in the 5′ flanking regions, where most promoter sequences are located. In contrast, D. melanogaster showed more TE insertions within introns, mainly due to a single gene, CYP307a2, which harbors 30 TE fragments in its intronic regions (Supplementary Table S3). In this gene, retroelements predominate across all regions (5′, 3′, and introns). The D. suzukii ortholog of CYP307a2 contains seven intronic insertions (Supplementary Table S1), where DNA transposons were the most abundant in all regions of this gene (5′, 3′, and introns). No TE insertions were detected within annotated exons.

When comparing CYP genes with TE insertions in D. suzukii and D. melanogaster, we observed that some D. suzukii genes were longer than their orthologs in D. melanogaster (Figures 3A–J). Specifically, ten of the 36 D. suzukii genes containing TE insertions had additional exons and introns, organized as repetitive conserved blocks, compared with their D. melanogaster counterparts. In general, gene organization is highly conserved among species of the same order; therefore, exon and intron annotations are well supported and consistent with the phylogenetic relationships of the species (Rewitz et al., 2007). To further explore this pattern, we also examined two sister species of D. suzukii with available genome sequences, D. biarmipes and D. takahashii. In all four species, exon and intron annotations are confirmed at the transcript level (Chiu et al., 2013; Graveley et al., 2010; Drosophila biarmipes Annotation Release 101; Drosophila takahashii Annotation Release 101).

Helitron insertions account for only a small proportion of the total length of each longer CYP gene in D. suzukii, ranging from 0.43% to 5.63%, except CYP4e2, which contains 1,403 bp of Helitron sequence in a gene of 7,998 bp (17.54%) (Supplementary Table S1). These results indicate that, although Helitrons are consistently present in longer genes, they represent only a minor contribution to overall gene length. This suggests that Helitrons may have contributed to gene lengthening through rearrangements such as exon shuffling rather than by adding new sequences. In this view, Helitron activity could have facilitated structural reorganization of CYP genes, consistent with their known role in mediating exon capture and recombination events.

Analysis with genoPlotR, which compares gene and genome maps, revealed conserved exon structures across orthologous genes in D. suzukii, D. biarmipes, D. takahashii, and D. melanogaster (Figures 3A–J). The genes CYP12a4, CYP12e1, CYP6a18, CYP6a20, CYP6a21, CYP6a23, CYP6d5, and CYP4e2 of D. suzukii contain at least one Helitron fragment in the intron region.

The CYP4e2 gene in D. suzukii is larger than its D. melanogaster ortholog but contains only one fewer exon compared with the D. biarmipes ortholog (Figure 3H). Interestingly, even with one fewer exon, the D. suzukii CYP4e2 remains longer than its D. biarmipes counterpart. A Helitron fragment located between exons six and seven is present in D. suzukii but absent from D. biarmipes, suggesting that the loss of this exon in D. suzukii may have resulted from the Helitron insertion.

In contrast, the CYP4c3 gene in D. suzukii contains two Helitron insertions in the 3′ flanking region in a positive orientation (5’-3’) (Figure 3I). Its sister species D. takahashii carries a single Helitron insertion between exons five and six in a negative orientation (3’-5’). The structural differences observed in D. suzukii CYP4c3 could be explained by the presence of these two Helitrons and the rolling-circle recombination mechanism associated with this TE superfamily, whereby exons nine to eleven may have arisen through exon shuffling involving exons six to eight.

Unlike what was previously observed in D. suzukii, the CYP12a4 and CYP12e1 orthologs in D. biarmipes and D. takahashii lack transposon insertions (Figures 3A, B). However, in D. melanogaster, the CYP12a4 ortholog contains the BARI element in the 3’ flanking region (Figure 3A), as previously annotated (Bogwitz et al., 2005). The biological functions of CYP12a4 and CYP6a20 have been reported as insecticide responses and aggressive behavior, respectively - two key traits that contribute to the ecological success and invasive potential of insects.

Little or no similarity was observed for the CYP6w1 gene annotated in scaffold 2 (Figure 3J). However, the same gene annotated on scaffold 8 showed high similarity to its orthologs. BLAST searches performed at NCBI showed high sequence identity with the CYP6d2 genes of the sister species D. biarmipes (89%) and D. takahashii (87%). The D. suzukii CYP6d2 gene is absent from Gbrowser (SpottedWingFlyBase, 2013) but is predicted by the NCBI genome browser. On the other hand, the CYP6d5 gene is annotated in two scaffolds (99 and 1273) (Figure 3G). Both paralogs display high similarity to each other and the D. biarmipes ortholog, suggesting a duplication event in D. suzukii that may have been facilitated by Helitron-mediated insertion and transposition.

To determine whether the observed genomic size increase and TE insertions were specific to CYP genes in D. suzukii or also occurred in other gene families, we randomly selected 500 additional genes from each genome for comparison. We visually inspected 124 genes that were longer in D. suzukii than in their D. melanogaster orthologs. Among these longer D. suzukii genes, 45 carried a total of 249 Helitron copies, whereas in D. melanogaster, 41 genes carried 110 Helitron copies (Supplementary Table S7).

To allow for a fair comparison not influenced by the overall genome size differences between species, we compared CYP gene lengths with the median length of 500 randomly selected genes within each species. In D. melanogaster, CYPs were significantly smaller (median = 2,117 bp) than random genes (median = 5,603 bp; p = 0.025, Mann-Whitney). In D. suzukii, however, CYPs (median = 8,032 bp) did not differ significantly in size from random genes (median = 6,325 bp; p = 0.526).

Because CYP size differences could be influenced by overall genome size rather than TE-mediated arrangements, we normalized CYP lengths to the median size of the 500 random genes from each species. After normalization, CYPs in D. melanogaster were proportionally shorter (median ratio = 0.38) than in D. suzukii (median ratio = 1.27; p = 0.002, Wilcoxon). These results indicate that, on average, CYP genes in D. melanogaster are 38% shorter than the typical genes in its genome, whereas in D. suzukii they are 27% longer. Therefore, CYPs remain proportionally larger in D. suzukii than in D. melanogaster, even after normalization for genome size differences. This normalization was necessary to account for differences in overall genome size and structure between species, allowing a direct evolutionary comparison of relative gene length rather than absolute values, which can be biased by genome expansion or contraction.

Transposable elements (TEs) often carry transcription-factor binding sites (TFBS), and these sequences are preferentially retained within genes because they can contribute to transcriptional regulation (Jordan et al., 2003; Feschotte, 2008). Such retention likely reflects a byproduct of TE transposition combined with the host-level selection. We therefore searched for putative TFBS in all sequences identified within the CYP genes of D. suzukii (88 TFBS) and D. melanogaster (140 TFBS) (Figure 4; Supplementary Table S8). Of these, 35 of the 88 TFBS in D. suzukii were located within Helitrons, whereas only 5 of the 140 TFBS in D. melanogaster were found in Helitrons. Across all TE fragments located within CYP genes and their flanking regions in D. suzukii, 51 out of 140 (36%) carried at least one predicted TFBS (Supplementary Table S8). Although no explicit false-positive control or background frequency analysis was implemented, confidence scores for each predicted TFBS are provided in Supplementary Table S8.

As different classes of TEs were present in the CYP genes (Table 1), a broad diversity of TFBS motifs was expected (Thornburg et al., 2006). However, we detected little variation in the TFBS classes across TE families (Figure 4 and Supplementary Table S8). Although D. suzukii has a larger overall proportion of TE sequences in its genome (35.94%) compared with D. melanogaster (15.96%) (Table 2), the highest number of TFBS was found in TE fragments located within CYP genes of D. melanogaster (Supplementary Table S8). This is likely explained by the greater total base pair coverage of TEs within CYP genes and their flanking regions in D. melanogaster (76,813 bp; Supplementary Table S3) compared with D. suzukii (47,421 bp; Supplementary Table S1). It is important to note that these base pair values refer only to TE content in CYP genes and should not be confused with the genome-wide TE proportions reported in Table 2.

For both species, the putative TFBS Hunchback and CF2-II (Chorion factor 2) are consistently overrepresented (Figure 4). These proteins belong to the C2H2 zinc finger class of transcription factors. Hunchback is strongly expressed early in development (Nüsslein-Volhard and Wieschaus, 1980; Lehmann, 1988), whereas CF2-II is expressed later in embryogenesis (Shea et al., 1990). TE insertions in flanking regions and introns of CYP genes may therefore influence expression by harboring putative TFBSs. In this study, flanking regions were defined as ±10 kb upstream and downstream of each gene, encompassing proximal regulatory zones where transposon-derived enhancers or silencers are most likely to act. This, in turn, suggests that TEs may play an important role in facilitating adaptation to different environments in both species (Jordan et al., 2003; Feschotte, 2008; Shea et al., 1990; Thornburg et al., 2006). However, these regulatory implications remain hypothetical. Future functional and transcriptomic experiments will be required to confirm whether these TE-associated TFBSs have effects on CYP gene regulation or adaptive phenotypes.

To place these findings in a genomic context, we next examined overall TE content in the genomes of D. suzukii and D. melanogaster. Approximately 36% of the assembled D. suzukii genome consists of TE sequences, compared with 16% in the genome of D. melanogaster (Table 2).

Three genome assemblies are currently available for D. suzukii. The first two were generated from North American samples (SRA096061; Chiu et al., 2013) – the same genomic data used in the present study – and from European samples (ERP001893; Ometto et al., 2013), both sequenced on the Illumina HiSeq2000 platform. Chiu et al. (2013) used automated homology comparison against 6,003 D. melanogaster TEs. In contrast, Ometto et al. (2013) applied a homology-based approach with RepeatMasker and the Repbase Insect library, estimating ~11% TE content in D. suzukii and ~17% in D. melanogaster. A third, near-chromosome-level assembly was produced by Paris et al. (2020) using PacBio long-read sequencing, which revealed that ~35% of the D. suzukii genome consists of repetitive sequences. However, the relative contribution of TE superfamilies was not determined. Differences among these studies likely reflect not only the methodologies used but also sequencing technology, which can affect the assembly completeness and thus TE content estimates.

In this study, we used RepeatExplorer, which integrates two complementary strategies for TE annotation: (1) homology-based searches against the Repbase library, and (2) de novo clustering to identify repetitive structures and patterns in the genome. This combined approach provides broader coverage, whereas earlier studies relying solely on homology-based annotation likely underestimated TE content. Supporting this explanation, Sessegolo et al. (2016), using the de novo pipeline dnaPipeTE (Goubert et al., 2015), estimated TE content at ~31% for D. suzukii and ~12% for D. melanogaster - values comparable to our results (~36% and ~16%, respectively). The primary difference is technical: dnaPipeTE requires local installation, while RepeatExplorer is web-based, but both rely on similar principles.

In the two species (Table 2), retrotransposons are the most abundant TE class, consistent with previous evidence that class I elements predominate in Drosophila genomes (Drosophila 12 Genomes Consortium, 2007). Among DNA transposons, Helitrons are the most abundant in both species and represent the second-largest TE category in D. suzukii. Actual TE content may be even higher, since RepeatExplorer tends to detect medium-to-high copy number and relatively recent TE insertions. In contrast, older, more diverged copies may not pass through similarity filters. Nevertheless, the clustering approach makes RepeatExplorer a fast and effective tool for initial TE analysis of Illumina data (Novák et al., 2013).

Comparing TE distribution between genes and intergenic regions revealed striking interspecific differences. In D. suzukii, TEs were found in 8.6% (10044 in 116977) of all annotated genes, whereas in D. melanogaster, 41.6% (35013 in 84181) of genes contained at least one TE insertion (p < 0.001). Moreover, in D. suzukii, 1.0% (103) of all TE copies and 1.0% (68) of all Helitron copies in the genome were located within CYP genes, while in D. melanogaster these proportions were 0.2% (87 TE copies) and 0.2% (19 Helitron copies), respectively. These differences in TE and Helitron proportions within CYP genes were statistically significant (p < 0.001).

Helitron distribution also differs between species. In D. suzukii, 95.6% of Helitrons are located in intergenic regions, whereas in D. melanogaster, 85.9% occur within genes. It should be noted that our methodology does not distinguish between complete Helitron elements and fragmented copies, which may contribute to differences observed between species. Nonetheless, fragmented copies likely represent remnants of once-intact Helitrons that were active earlier in the evolutionary history of these species. For this reason, our interpretations regarding the structural influence of Helitrons refer to their historical activity rather than current transposition or exon-shuffling events. This pattern may reflect a species-specific distribution of Helitrons, suggesting potential differences in TE dynamics, and does not alter the conclusion that D. suzukii harbors proportionally more Helitrons in intergenic regions.

Metabolic resistance mediated by cytochrome P450 monooxygenases (CYPs) is an important adaptive trait in many insect species (Scott, 1999) and a common mechanism by which insects develop resistance to pesticides (Feyereisen, 1999). Transposable elements (TEs) are often found within or near resistance genes, providing indirect evidence of their involvement in the generation of adaptive genome changes (Catania et al., 2004; Chen and Li, 2007; Chung et al., 2007; Carareto et al., 2013; Casacuberta and González, 2013). Barbara McClintock (1984) first proposed that TE activation in response to stress could induce mutations that help organisms adapt to new environmental conditions.

In this study, we examined TEs associated with CYP genes in the highly invasive D. suzukii genome. We documented CYPs with varying TE contents, including TEs carrying putative transcription-factor binding sites (TFBS) and structural changes potentially mediated by rolling-circle transposons of the Helitron superfamily. We also found that the D. suzukii genome contains roughly twice the TE content of D. melanogaster, with Helitrons representing the most abundant subclass of class II DNA transposons in both species (Table 2).

In all CYP genes analyzed, TE fragments were located exclusively in flanking regions and introns, which is consistent with the view that TEs are generally tolerated in non-coding regions. However, TE insertions near genes can also create new regulatory networks (Feschotte, 2008), and changes in a gene-regulation network are thought to be very important during adaptive evolution (Casacuberta and González, 2013). In D. suzukii, TE insertions occurred predominantly (88%) in the 5’ flanking regions of CYP genes. Previous studies have shown that TE insertions in 5’ untranslated regions confer insecticide resistance – for example, in Drosophila CYP6g1, where the upstream ACCORD retroelement carries specific transcriptional enhancers (Daborn et al., 2002; Chung et al., 2007; Schmidt et al., 2010). D. melanogaster and D. simulans CYP genes harbor multiple TE insertions, many from the Helitron superfamily, which also carries putative TFBS (Carareto et al., 2013; review in Thomas and Pritham, 2015). These findings are consistent with previous hypotheses proposing that TEs may be gradually co-opted for host gene regulation (Chung et al., 2007; Feschotte, 2008).

The acquisition of new cis-regulatory elements via TE insertions provides opportunities for adaptation to novel environmental challenges (Casacuberta and González, 2013). Several LTR retrotransposons containing cis-regulatory motifs are highly expressed in response to specific stimuli (Kumar and Bennetzen, 1999), and these motifs often match those required for the activation of stress-response genes (Grandbastien et al., 2005). In our dataset, TE fragments carried putative TFBS involved in fly development, including Hunchback (embryo patterning) and CF2-II (cell differentiation). This suggests that CYPs may be particularly permissive to TE insertions because such sequences can act as donors of transcriptional regulatory signals, potentially altering gene expression at different developmental stages. Similar TFBS have been reported in TE sequences in silico (Babu et al., 2006; Thornburg et al., 2006; Carareto et al., 2013). Thus, TEs carrying TFBS may influence gene regulation and contribute to adaptation in Drosophila (Feschotte, 2008).

Beyond regulatory effects, TEs can also mediate structural genomic changes such as insertions, excisions, retrotranspositions, and exon shuffling. These processes can lead to exonization or intronization of TE sequences, and in some cases to exaptation, where TE-derived sequences acquire new functional roles. If beneficial, such insertions can be retained in the host genome. Feyereisen (1999) proposed two mechanisms by which CYP genes can evolve insecticide resistance: (1) structural changes in specific CYPs, such as exon gain or loss, and (2) increased gene expression. Exon shuffling, as first proposed by Gilbert (1987), is one route by which novel exons can arise. In our study, ten CYP genes (Figures 3A–J) displayed structural changes involving conserved blocks of exon gain, each associated with at least one Helitron insertion.

Helitrons – subclass 2 of Class II DNA transposons (Wicker et al., 2007) – are known to mediate exon shuffling, transduplication, and the introduction of novel regulatory elements (Morgante et al., 2005; Pritham and Feschotte, 2007; Thomas et al., 2014). These elements transpose via a rolling-circle mechanism that displaces a single DNA strand. A loop is formed before cleavage and reintegration elsewhere in the genome. They have a remarkable ability to capture and duplicate gene segments, and their transposition can include flanking sequences (Kapitonov and Jurka, 2007). While Helitrons are well studied in plants – especially maize, where they have captured and redistributed numerous genes (Lal et al., 2009; Barbaglia et al., 2012) – their role in Drosophila remains less explored.

Repetitive elements within introns may act as recombination hotspots, thereby promoting exon shuffling (Gilbert, 1987). In maize, most Helitron copies have incorporated gene segments, facilitating their amplification and dispersal throughout the genome (Yang and Bennetzen, 2009). A striking example outside Drosophila comes from Palmer amaranth (Amaranthus palmeri), where Helitron-mediated amplification of the EPSPS gene cassette confers glyphosate resistance (Molin et al., 2017). Our observations in D. suzukii CYP genes are consistent with such a mechanism, suggesting that Helitron insertions are associated with increased gene length.

We propose a hypothetical example of Helitron-mediated gene capture in Figure 5, illustrating how, if a Helitron bypasses its termination signal, strand displacement could continue through adjacent gene regions until a new signal is encountered, capturing and mobilizing those sequences (Kapitonov and Jurka, 2007; Grabundzija et al., 2016). For instance, in CYP12a4 (Figure 3A) and CYP6a20 (Figure 3D), the intronic arrangement, orientation, and high sequence similarity of exons support the possibility of Helitron-mediated capture during transposition (Figure 5A). Further studies should aim to experimentally validate Helitron-mediated gene capture in D. suzukii through long-read sequencing and transcriptomic analyses to confirm the presence of chimeric transcripts.

Previous studies (Li et al., 2007) have documented genomic alterations leading to CYP overexpression in insecticide resistance. Mishra et al. (2018) reported that several longer CYP genes – CYP6w1, CYP6a20, CYP6a21, CYP6d5 – were significantly upregulated under insecticide exposure in D. suzukii, with responses varying between populations. Functional assays, such as CRISPR/Cas9-mediated knockouts, will be important for testing whether Helitron insertions affect gene expression and adaptive traits of this pest species.

In our comparative analysis of closely related Drosophila species (D. suzukii, D. melanogaster, D. biarmipes, and D. takahashii), the ten longer CYP genes revealed a largely conserved exon–intron organization across species, aligned with phylogenetic relationships (Figures 3A–J). However, D. suzukii displayed Helitron insertions absent from its sister species, contributing to gene length variation (CYP4e2 and CYP6d5). Although D. suzukii has fewer CYP genes than does D. melanogaster, previous work shows that longer genes can generate greater functional novelty than large gene families (Grishkevich and Yanai, 2014), in part because gene length is positively correlated with the number of splice variants (Kopelman et al., 2005). Gene lengthening, often driven by TE insertions (Grishkevich and Yanai, 2014), may thus contribute to adaptive structural changes. Our findings suggest that Helitrons could be a vehicle for such changes in D. suzukii CYP genes through their combined transposition and recombination activities.