Larval diapause slows adult epigenetic ageing in an insect model, Nasonia vitripennis

Diapause was induced by maintaining Nasonia vitripennis mothers at 20°C under a 16:8 h light:dark (LD 16:8) cycle. After ten days, their larval offspring are diapaused destined. Fourth instar diapaused larvae were maintained at 4°C in constant darkness (DD) for three months before adult emergence. All experiments were carried out on virgin males. Post-emergence, diapaused and non-diapaused (control) virgin males were individually housed (25°C, 40% humidity, LD 12:12), fed daily with 20% sucrose solution, and their survival monitored daily. For whole-genome bisulfite sequencing (WGBS), individuals were sampled at five timepoints post emergence (days 6, 12, 18, 24, 30). Each of the 40 libraries (2 treatments x 5 timepoints x 4 replicates) represented pooled DNA from 10 individuals and included a 1% unmethylated lambda spike-in. Sequencing and bioinformatic processing, including read trimming, alignment, methylation calling, and duplicate removal, followed our standard pipeline ( 8). Age-related differentially methylated loci (DMLs) were identified (12) and then filtered to retain CpGs strongly corre- lated (Pearson’s|r|≥0.3, uncorrected p≤0.05) with chronological age in control samples for epigenetic clock construction (13). An elastic net regression was trained on age-associated CpG sites to predict chronological age providing an epigenetic clock. Predicted epigenetic age was validated against chronological age using a linear model. Full methodological details are provided in the Supplementary Information.

and Discussion Diapause treatment extended lifespan (Figure 1A), reducing the hazard of death by approximately 65% compared to controls (Cox Proportional Hazards model; Hazard Ratio [HR] = 0.35, 95% CI: 0.26–0.49; Wald test, p = 5×10−9). This was also reflected in the median survival determined by Kaplan-Meier analysis which was 30 days (95% CI: 28–32 days) for the diapause group (n=71), versus 22 days (95% CI: 22–23 days) for controls (n=101). From 715,987 CpG sites classified as methylated (SI Dataset S1), the generalized linear model (12) identified 7,950 CpGs with significant age-related differential methylation (SI Dataset S2). We further prioritized 289 of these sites that were strongly correlated with chronological age (Pearson’s |r|≥0.3, uncorrected p≤0.05) as input features for an Elastic Net regression model (SI Dataset S3). The final optimised model (mixing parameter α= 0.5; regularization λ= 1.84207, determined by 10-fold repeated cross-validation [3 repeats] minimizing Root Mean Squared Error [RMSE]) utilized a concise panel of 27 CpGs to estimate epigenetic age (SI Dataset S4). Our resulting epigenetic clock accurately predicted chrono- logical age in control samples, explaining 91.7% of the variance (cross-validatedR2; RMSE = 2.44 days). Importantly, the clock also performed robustly when applied to diapause samples, accounting for 78.0% of chronological age variance (R2; RMSE = 3.98 days), demonstrating its potential applicability across distinct physiological conditions. Curiously, at day 6 post-eclosion, diapaused adults were epigenetically older than age-matched controls by an esti- mated 2.8 days (diapaused: 11.32 days vs. control: 8.53 days, t = -3.22, d.f. = 36, p = 0.0027, emmeans post hoc). One possible explanation is that epigenetic ageing occurs during the diapause period, albeit at a markedly reduced rate compared to adult ageing. Alternatively, the observed overshoot may reflect transient remodelling of DNA methylation during emergence from diapause. Distinguishing between these scenarios will require direct measurement of methylation dynamics during diapause itself. Despite this initial increase in epigenetic age, diapaused adults subsequently age epigenetically 29% more slowly than controls (Figure 1B; control slope = 0.78812, diapause slope = 0.55828; linear model interaction of day and treatment: t = -3.903, d.f. = 36, p = 0.000399). By day 18, both groups converge on an epigenetic age of approximately 18 days. However, by day 30, diapaused individuals are epigenetically 2.7 days younger than controls (24.71 days vs. 27.44 days; emmeans post hoc: t = 3.152, d.f. = 36, p = 0.0033). This striking deceleration mirrors the observed extension in lifespan and 65% reduction in mortality hazard in diapaused individuals. Together, these results suggest that the molecular changes captured by the epigenetic clock are closely aligned with the physiological mechanisms that promote survival post-diapause. Our findings place diapause-induced epigenetic decelera- tion within the broader context of early life programming, where environmental cues can reshape long-term molecular and physiological trajectories. The direction and magnitude of changes in the epigenome largely depend on early life environmental conditions (14). This is not unexpected, as the epigenome is highly plastic during early development. Such plasticity may be adaptive, allowing early life environments to reconfigure the epigenome in ways that enhance future fitness, 2 — www.pnas.org/cgi/doi/10.1073/pnas.XXXXXXXXXX Foley et al. .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 23, 2025. ; https://doi.org/10.1101/2025.05.22.655466doi: bioRxiv preprint DRAFT 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 consistent with the predictive adaptive response hypothesis (15). In this framework, insect larval diapause would serve as a predictive adaptive response: an overwintering strategy that anticipates a more challenging adult environment. Adults that have passed through diapause may be under selection to survive longer, facilitating reproductive success in harsher post-winter conditions. In Nasonia, our data suggest that this is mirrored at the molecular level by a long-term slowing of the epigenetic clock. Epigenetic clocks across diverse species consistently high- light key developmental gene sets as predictors of biological age (16). The CpGs comprising our Nasonia epigenetic clock are significantly enriched for gene ontology terms related to conserved developmental and nutrient-sensing pathways, including mTOR and insulin/IGF signaling (SI Dataset S5). These pathways are central regulators of growth, metabolism, and lifespan. Theoretical links between development and ageing have a long history in evolutionary biology ( 17), and experimental studies, particularly in invertebrates such as C. elegans, have shown that developmental alterations, such as entry into the dauer stage, can dramatically extend lifespan. Our findings align with recent work in Drosophila, where Kang et al. (18) demonstrated that delayed development in prothoracicotropic hormone (PTTH)-null mutants extends lifespan and postpones the onset of age-related transcriptional changes. In Nasonia, we observe a similar phenomenon, larval diapause, a naturally induced developmental delay, slows the progression of the epigenetic ageing clock. Notably, PTTH suppression has been implicated in diapause induction across diverse insect species ( 19). Together, these findings point to a conserved endocrine-epigenetic axis through which developmental timing modulates ageing trajectories. Epigenetic ageing is influenced by inflammation, cell division, metabolic state, and early-life environment ( 3). With its compact genome, short lifespan, and functional methylation system, Nasonia enables experimental dissection of these processes in vivo ( 6). Our findings demonstrate that epigenetic ageing in this model is not only measurable, but developmentally modifiable. This positions Nasonia to address a fundamental translational question: can targeted reductions in epigenetic age improve long-term health and resilience? ACKNOWLEDGMENTS. EBF was supported by a BBSRC MIBTP DTP studentships. CT was funded by a Leverhulme Trust award RPG-2020-363. EM was funded by a BBSRC Pioneer Award APP3335. For the purpose of open access, the author has applied a Creative Commons Attribution (CC BY) licence to any Author Accepted Manuscript version arising 1. K Seale, S Horvath, A Teschendorff, N Eynon, S Voisin, Making Sense of the Ageing Methylome. Nat. Rev. Genet. 23, 585–605 (2022) Publisher: Nature Publishing Group. 2. L Drew, Turning back time with epigenetic clocks. Nature 601, S20–S22 (2022) Bandiera abtest: a Cg type: Outlook Number: 7893 Publisher: Nature Publishing Group Subject term: Ageing, Society, Epigenetics. 3. CG Bell, et al., DNA methylation aging clocks: challenges and recommendations. Genome Biol. 20, 249 (2019). 4. F Lyko, R Maleszka, Insects as Innovative Models for Functional Studies of DNA Methylation. T rends genetics27, 127–131 (2011). 5. CW Hu, JL Chen, YW Hsu, CC Y en, MR Chao, Trace analysis of methylated and hydroxymethylated cytosines in DNA by isotope-dilution LC-MS/MS: first evidence of DNA methylation in Caenorhabditis elegans. The Biochem. J. 465, 39–47 (2015). 6. JH Werren, et al., Functional and evolutionary insights from the genomes of three parasitoid Nasonia species. Sci. (New Y ork, N.Y .)327, 343–348 (2010). 7. X Wang, et al., Function and Evolution of DNA Methylation in Nasonia vitripennis. PLOS Genet. 9, e1003872 (2013). 8. K Brink, CL Thomas, A Jones, TW Chan, EB Mallon, Exploring the ageing methylome in the model insect, Nasonia vitripennis. BMC Genomics 25, 305 (2024). 9. DL Denlinger, Why study diapause? Entomol. Res. 38, 1–9 (2008) eprint: https://onlinelibrary.wiley.com/doi/pdf/10.1111/j.1748-5967.2008.00139.x. 10. X Karp, Hormonal Regulation of Diapause and Development in Nematodes, Insects, and Fishes. Front. Ecol. Evol. 9 (2021) Publisher: Frontiers. 11. M Pegoraro, A Bafna, NJ Davies, DM Shuker, E Tauber, DNA methylation changes induced by long and short photoperiods in Nasonia. Genome Res. 26, 203–210 (2016). 12. Y Park, H Wu, Differential methylation analysis for BS-seq data under general experimental design. Bioinformatics 32, 1446–1453 (2016). 13. AE Teschendorff, S Horvath, Epigenetic ageing clocks: statistical methods and emerging computational challenges. Nat. Rev. Genet. 26, 350–368 (2025) Publisher: Nature Publishing Group. 14. A Vaiserman, Developmental Tuning of Epigenetic Clock. Front. Genet. 9 (2018). 15. P Bateson, P Gluckman, M Hanson, The biology of developmental plasticity and the Predictive Adaptive Response hypothesis. The J. Physiol. 592, 2357–2368 (2014). 16. D Gems, RS Virk, JP de Magalh ˜aes, Epigenetic clocks and programmatic aging. Ageing Res. Rev. 101, 102546 (2024). 17. JP de Magalh ˜aes, Programmatic features of aging originating in development: aging mechanisms beyond molecular damage? The FASEB J. 26, 4821–4826 (2012). 18. P Kang, et al., NF-κB-mediated developmental delay extends lifespan in Drosophila. Proc. Natl. Acad. Sci. 122, e2420811122 (2025) Publisher: Proceedings of the National Academy of Sciences. 19. Q Wang, AAM Mohamed, M Takeda, Serotonin Receptor B May Lock the Gate of PTTH Release/Synthesis in the Chinese Silk Moth, Antheraea pernyi ; A Diapause Initiation/Maintenance Mechanism? PLOS ONE 8, e79381 (2013). Foley et al. PNAS — May 22, 2025 — vol. XXX — no. XX — 3 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 23, 2025. ; https://doi.org/10.1101/2025.05.22.655466doi: bioRxiv preprint 1 Supporting Information for2 Larval diapause slows adult epigenetic ageing in an insect model, Nasonia vitripennis3 Erin B. Foley, Charalambos P. Kyriacou, Christian L. Thomas & Eamonn B. Mallon4 Eamonn B. Mallon.5 E-mail: [email protected] This PDF file includes:7 Supporting text8 Legends for Dataset S1 to S59 SI References10 Other supporting materials for this manuscript include the following:11 Datasets S1 to S512 Erin B. Foley, Charalambos P. Kyriacou, Christian L. Thomas & Eamonn B. Mallon 1 of 4 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 23, 2025. ; https://doi.org/10.1101/2025.05.22.655466doi: bioRxiv preprint Supporting Information Text13 Methods14 Rearing. Nasonia vitripennis used in this study were from the Leicester strain, a laboratory colony maintained at the University15 of Leicester for over nine years. This strain derives from AsymC, originally isolated in 1989 and subsequently cured ofWolbachia16 via heat shock treatment (1,2).17 To generate diapaused offspring, virgin females were housed at 20°C under a 16:8 h dark:light photoperiod. Host pupae18 were collected from Day 10 post-oviposition onwards to ensure recovery of diapaused larvae. These larvae developed to the19 fourth instar before being transferred to continuous darkness at 4°C for three months. After the diapause period, larvae were20 returned to standard rearing conditions and allowed to complete development to adulthood.21 Lifespan Experiments.Diapaused and non-diapaused virgin maleNasonia were collected in batches within 24 hours of adult22 emergence (Day 0) and housed individually in plastic tubes. All individuals were maintained at 25°C and 40% relative humidity23 under a 12:12 h light:dark cycle. Each wasp received a daily feeding of 20µL of 20% sucrose solution on filter paper, and24 mortality was recorded daily. Lifespan was defined as the number of days from adult emergence to death.25 Kaplan–Meier survival analyses were performed using the survival package (v3.7) (3) and visualized with the survminer26 package (v0.4.9) (4). Cox proportional hazards models were fitted using the survival package. All analyses were conducted in27 R v4.4.1 (5).28 Lifespan data were used to select sampling time points for whole-genome bisulfite sequencing (WGBS). Diapaused and29 non-diapaused virgin males were sampled at Days 6, 12, 18, 24, and 30, always at the same time of day. Upon collection,30 individuals were snap-frozen in liquid nitrogen and stored at –80°C until DNA extraction.31 DNA Extraction.DNA was extracted using an adapted protocol based on the AllPrep® DNA/RNA Micro Kit (Qiagen). Each32 sample consisted of ten whole-body adult virgin maleNasonia. Four biological replicates were processed per time point (Days33 6, 12, 18, 24, and 30).34 DNA quality and concentration were assessed using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific), a35 QubitTM dsDNA BR Assay Kit (Thermo Fisher Scientific), and electrophoresis on a 1% agarose gel run for 40 minutes at36 100V. Samples were sent to Novogene (Beijing) for whole-genome bisulfite sequencing (WGBS). A 1% unmethylated lambda37 DNA spike-in was included in each sample to assess bisulfite conversion efficiency.38 Whole-Genome Bisulfite Sequencing and Bioinformatic Processing. Raw sequencing data were provided in FASTQ format39 (submitted to Sequence Read Archive (SRA)). For samples sequenced across multiple lanes, read files were concatenated using40 the Unix cat command. Read quality was assessed using FastQC v0.12.1 (6).41 A custom Snakemake pipeline (Snakemake v7.32.4; Python v3.12.1) (7) was used for preprocessing. Adapter sequences were42 trimmed and the first 10 bases removed from each read using Cutadapt v4.4 (8). Paired-end reads were aligned to theNasonia43 vitripennis reference genome (Nvit_PSR1.1) (9) using Bowtie2 v2.5.1 with default parameters (10). Reads were also aligned to44 the unmethylated lambda genome (RefSeq accession: GCF_000840245.1) to assess bisulfite conversion efficiency.45 Aligned reads were deduplicated and cytosine methylation calls extracted using Bismark v0.22.3 (11). Strand ambiguity was46 resolved using thecoverage2cytosine utility in Bismark to generate destranded coverage files, which were used for downstream47 analyses.48 Methylation Analysis.All downstream analyses were performed in R v4.4.1 (5). Destranded CpG coverage files were imported49 into the MethylKit package v1.30.0 (12) using themethRead() function. CpG sites with less than 10× coverage or coverage50 above the 99th percentile were filtered out. A binomial test was applied to each sample using the lambda genome conversion51 rate as the null probability of success, with a false discovery rate (FDR) threshold ofp < 0.05 (SI Dataset S5). Only CpG sites52 showing significant methylation in at least one sample were retained. Percentage methylation at each site was calculated using53 the percMethylation() function.54 Differential Methylation Analysis.Differential methylation analysis was performed using theDSS package v2.52.0 (13), which55 models methylation proportions using a beta-binomial generalized linear model (GLM) with an arcsine link function. A design56 matrix was constructed incorporating time point and treatment (diapause vs. non-diapause) as experimental factors. Linear57 models were fitted using theDMLfit() function to evaluate main effects and interactions. Differentially methylated loci (DMLs)58 and regions (DMRs) were identified accordingly.59 Genomic features were assigned to CpG sites using a custom annotation file (GFF format) generated by Dr. Hollie Marshall60 using AGAT v0.10.0 (14).61 Epigenetic Clock Construction.An elastic net regression model was trained to predict chronological age from DNA methylation62 levels across CpG sites, using data from virgin maleNasonia vitripennis sampled at five time points (Days 6, 12, 18, 24, and63 30). Methylation data (percentage values) were filtered to retain CpG sites previously identified as differentially methylated64 across time usingDSS (13).65 To enhance model sparsity and robustness, CpG sites were pre-selected based on univariate Pearson correlation with age in66 the non-diapaused (control) group. Sites with absolute correlation≥ 0.3 and uncorrectedp-value≤ 0.05 were retained. CpG67 features were then centered and scaled.68 2 of 4 Erin B. Foley, Charalambos P. Kyriacou, Christian L. Thomas & Eamonn B. Mallon .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 23, 2025. ; https://doi.org/10.1101/2025.05.22.655466doi: bioRxiv preprint The final model was implemented using theglmnet algorithm with elastic net regularization (α = 0.5), as part of thecaret69 framework (15). Ten-fold cross-validation with three repeats was used for model tuning and performance estimation. The70 treatment group (diapause vs. control) was included as a categorical predictor, encoded as a dummy variable, and entered71 alongside the CpG methylation features. The optimal model was selected based on minimum root mean squared error (RMSE).72 Predictions were generated for all samples, and model performance was assessed using RMSE andR2, stratified by treatment73 group.74 Post-hoc analysis of predicted age trajectories was conducted using linear models with interaction terms, and estimated75 marginal means were compared between treatment groups at multiple time points using theemmeans package (16).76 Gene Ontology Enrichment Analysis.Gene Ontology (GO) enrichment analysis was performed to assess the functional signifi-77 cance of genes associated with CpG sites included in the final epigenetic clock model. The 27 clock CpGs (SI Dataset S4)78 were mapped to nearby genes, which were then tested for enrichment against a background set comprising all genes associated79 with differentially methylated loci across time (SI Dataset S1). GO annotations were derived fromNasonia vitripennis and80 formatted for compatibility withGOstats (17).81 The gene list was tested for overrepresentation of GO terms in the Biological Process (BP), Cellular Component (CC), and82 Molecular Function (MF) ontologies using a conditional hypergeometric test. Analyses were performed using theGOstats and83 GSEABase packages. For each ontology, both over- and under-representation were tested, and GO terms with an adjusted84 FDR of< 0.05 (Benjamini-Hochberg) were considered significant.85 To visualise and summarise redundant GO terms, semantic similarity clustering was performed using therrvgo package (18).86 Pairwise GO term similarities were calculated using theorg.Dm.eg.db annotation database and the “Rel” semantic similarity87 measure. Representative GO terms were identified by reducing the similarity matrix with a similarity threshold of 0.7. Enriched88 terms were visualised using treemaps, heatmaps, and scatter plots. Final results were exported in tabular form (SI Dataset S5).89 All scripts used are available athttps://tinyurl.com/5n6vcvsk90 SI Dataset S1 (SI_Datasets.xlsx/S1_Methylated_CpGs)91 CpG sites classified as methylated using a binomial test with the lambda genome conversion rate as the null probability of92 success. Sites passing a false discovery rate (FDR) threshold ofp< 0.05 and significantly methylated in at least one sample93 were retained.94 SI Dataset S2 (SI_Datasets.xlsx/S2_Age_DMPs)95 7,950 CpG sites showing significant age-associated differential methylation.96 SI Dataset S3 (SI_Datasets.xlsx/S3_Age_Correlated)97 289 age-associated CpG sites with strong correlation to chronological age (Pearson’s|r|≥ 0.3, uncorrectedp≤ 0.05).98 SI Dataset S4 (SI_Datasets.xlsx/S4_Clock_Coefficients)99 Elastic net model coefficients for the 27 CpG sites and intercept used to estimate epigenetic age.100 SI Dataset S5 (SI_Datasets.xlsx/S5_GO_terms)101 Enriched gene ontology terms associated with the 27 CpG sites comprising the epigenetic clock.102 References103 1. JH Werren, DW Loehlin, The parasitoid wasp Nasonia: an emerging model system with haploid male genetics.Cold104 Spring Harb. Protoc. 2009, pdb–emo134 (2009) Publisher: Cold Spring Harbor Laboratory Press.105 2. DC Darling, JH Werren, Biosystematics of Nasonia (Hymenoptera: Pteromalidae): two new species reared from birds’106 nests in North America.Annals Entomol. Soc. Am. 83, 352–370 (1990) Publisher: Oxford University Press Oxford, UK.107 3. Terry M. Therneau, Patricia M. Grambsch,Modeling Survival Data: Extending the Cox Model. (Springer, New York),108 (2000).109 4. A Kassambara, M Kosinski, P Biecek,survminer: Drawing Survival Curves using ’ggplot2’ . (2021).110 5. R Core Team,R: A Language and Environment for Statistical Computing. (R Foundation for Statistical Computing,111 Vienna, Austria), (2024).112 6. S Andrews, et al., FastQC (2012) Place: Babraham, UK Published: Babraham Institute.113 7. J Köster, S Rahmann, Snakemake—a scalable bioinformatics workflow engine.Bioinformatics 28, 2520–2522 (2012)114 _eprint: https://academic.oup.com/bioinformatics/article-pdf/28/19/2520/48879301/bioinformatics_28_19_2520.pdf.115 8. M Martin, Cutadapt removes adapter sequences from high-throughput sequencing reads.EMBnet.journal 17, 10–12116 (2011).117 9. E Dalla Benetta, et al., Genome elimination mediated by gene expression from a selfish chromosome.Sci. Adv. 6, eaaz9808118 (2020) Publisher: American Association for the Advancement of Science.119 10. B Langmead, SL Salzberg, Fast gapped-read alignment with Bowtie 2.Nat. methods 9, 357–359 (2012) Publisher: Nature120 Publishing Group.121 Erin B. Foley, Charalambos P. Kyriacou, Christian L. Thomas & Eamonn B. Mallon 3 of 4 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 23, 2025. ; https://doi.org/10.1101/2025.05.22.655466doi: bioRxiv preprint 11. F Krueger, SR Andrews, Bismark: a flexible aligner and methylation caller for Bisulfite-Seq applications.bioinformatics122 27, 1571–1572 (2011) Publisher: Oxford University Press.123 12. A Akalin, et al., methylKit: a comprehensive R package for the analysis of genome-wide DNA methylation profiles.124 Genome biology 13, 1–9 (2012) Publisher: Springer.125 13. H Feng, H Wu, Differential methylation analysis for bisulfite sequencing using DSS.Quant. Biol. 7, 327–334 (2019)126 Publisher: Springer.127 14. J Dainat, AGAT: Another Gff Analysis Toolkit to handle annotations in any GTF/GFF format.(Version v0. 7.0).Zendo.128 doi 10 (2023).129 15. M Kuhn, Building Predictive Models in R Using the caret Package.J. Stat. Softw. 28, 1–26 (2008).130 16. RV Lenth, emmeans: Estimated marginal means, aka least-squares means, manual (2022).131 17. S Falcon, R Gentleman, Using GOstats to test gene lists for GO term association.Bioinformatics 23, 257–258 (2007)132 Publisher: Oxford University Press.133 18. S Sayols, rrvgo: a Bioconductor package for interpreting lists of Gene Ontology terms.microPublication Biol. (2023)134 Publisher: Caltech Library.135 4 of 4 Erin B. Foley, Charalambos P. Kyriacou, Christian L. Thomas & Eamonn B. Mallon .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 23, 2025. ; https://doi.org/10.1101/2025.05.22.655466doi: bioRxiv preprint