Scion-based drought stress memory affects potato response to water deficit Authors: Waldemar Marczewski, Alicja Macko-Podgórni, Jarosław Plich, Dominika Boguszewska-Mańkowska, Paulina Smyda-Dajmund, Dariusz Grzebelus Published: 2026 ABSTRACT AbstractMain conclusionA scion-based stress memory signal, which was derived from drought-primed potato plants, was transmitted to new potato plants generated through vegetative reproduction. This affected potato tuber yield.AbstractDrought is one of the most significant threats to agricultural productivity worldwide. The cultivated potato (Solanum tuberosum L.) is a crop species that is sensitive to drought stress. This study investigated the impact of scion-based drought stress memory on tuber yield, physiological parameters, gene expression, and DNA methylation in the vegetative progeny of grafted plants. The tuber progeny plants remembered the drought stress signal transmitted from the drought-primed scion. Significant changes were observed in the expression of genes, primarily those related to photosynthetic metabolic pathways, as well as those associated with chromatin remodeling, DNA repair, and the plant’s response to abiotic stresses. The gene expression landscape corresponded with variability in chlorophyll fluorescence parameters. In the first and the second generation of vegetatively propagated plants, scion-based memory had a positive effect on tuber yield. This was achieved by buffering the decline in yield caused by drought, as compared to plants grown under control conditions. Whole-genome bisulfite sequencing analysis revealed no correlation between changes in DNA methylation and gene expression. Drought-induced alterations in DNA methylation were erased in the second progeny generation. We propose that there is a direct causal relationship between scion-based memory of drought stress and photosynthetic efficiency, as well as potato tuber productivity.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1007/s00425-026-05154-0.Keywords: Abiotic stress, DNA methylation, Grafting, JIP-test parameters, RNA-seq, Tuber yield Introduction Of all the effects of global warming, drought stress has the most negative impact on plants. Crop productivity losses related to drought are set to increase worldwide. The impact of drought stress depends on various factors, including stress intensity and duration, plant genotype, and growth phase (Khan et al. 2025). The cultivated potato (Solanum tuberosum L.) is a sensitive model crop species (Nasir and Toth 2022). The above- and below-ground organs of potato plants react differently to drought stress. During a period of water deficit, many physiological processes are disturbed (Huntenburg et al. 2021). For agronomic practices, the extent of the decrease in tuber yield is the main criterion for assessing potato drought tolerance (Boguszewska et al. 2010). Katahdin is an old North American potato cultivar that produces high yields of well-shaped tubers. It is relatively drought-tolerant and produces high-quality tubers even in hot and dry conditions (Clark et al. 1931). It has therefore been used as the male or female parent in the breeding of many potato cultivars worldwide (Potato Pedigree Database, www.plantbreeding.wur.nl/potatopedigree). In a previous paper, we reported on differences in the relative water content and drought-induced yield reduction among 18 Katahdin-derived potato cultivars from a half-sib family (Sołtys-Kalina et al. 2016). Our 3-year studies showed that there was a strong diversification in drought tolerance levels even among genetically closely related Katahdin-derived potato cultivars (Plich et al. 2020). We demonstrated that drought priming induced cultivar-specific profiles of differentially methylated regions (DMRs) in Katahdin and five Katahdin-derived potato cultivars. Changes in DNA methylation in these cultivars were transmitted to the first generation of vegetatively propagated progeny and then largely lost in the subsequent generation (Macko-Podgórni et al. 2025). The photosynthetic activity of Katahdin-derived cultivars propagated under non-stressed conditions was impacted by drought-stress memory (Marczewski et al. 2026). Grafting is an important horticultural technique that involves joining two plants, either of the same species or different species, together (the scion and the rootstock). Numerous molecules can move bidirectionally in a graft union as mobile signals for epigenetic modifications and gene regulation (Dong et al. 2022). Grafting experiments in potato have proven that tuberization initiation is mediated by graft-transmissible signals, such as mobile proteins, mRNA, and phytohormones (Dutt et al. 2017). MicroRNAs have been identified as potential graft-transmissible signals affecting potato plant architecture and tuberization (Odgerel and Banfalvi 2021). Inter-species grafting in Solanaceae plants has also been shown to cause extensive, heritable alterations to DNA methylation (Wu et al. 2013). This technique was commonly used in a number of fruit and vegetable crops to improve tolerance to biotic and abiotic stresses. In tobacco (Nicotiana tabacum L.), grafting onto rootstock genotypes that are tolerant to drought improves drought stress tolerance by regulating the activities of antioxidative enzymes and the expression of genes that respond to stress (Liu et al. 2014). Grafting susceptible tomato (Solanum lycopersicum L.) genotypes onto resistant rootstocks mitigates the harmful effects of drought stress by improving photosynthetic pigment levels and reducing oxidative stress (Mahapatra et al. 2025). Grafting promoted drought stress tolerance through changes in the expression of several genes, even in self-grafted tomato plants that did not use vigorous rootstocks. Several wound-healing mechanisms can be activated during the formation of a graft union, leading to increased drought tolerance (Fuentes-Merlos et al. 2022). In tomato/potato heterografting experiments, the potato yields produced by the rootstocks were significantly lower than those of non-grafted potato plants (Zhang and Guo 2019). The photosynthetic apparatus is highly susceptible to any kind of abiotic stress (Singh and Thakur 2018). JIP-test parameters enable the impact of drought stress on plant photosynthesis to be rapidly, sensitively, and non-invasively measured (Boguszewska-Mańkowska et al. 2018). These fluorescence measurements have been successfully employed to monitor and characterize a variety of photosynthetic events (Strasser et al. 2000). The present study examined the impact of drought stress memory, derived from drought-primed Katahdin scions, on tuber yield, photosynthetic parameters, gene expression, and DNA methylation in the vegetative progeny of grafted plants. Materials and methods Plant material The cultivar Katahdin was obtained from the potato collection of the Plant Breeding and Acclimatization Institute (National Research Institute, Bonin, Poland). In 2021, 12 cut tuber pieces with sprouts were planted into pots and maintained under greenhouse conditions for 4 weeks. After this period, six plants of equal height were transferred to tents in which the ground was lined with foil to prevent water soaking. The plants were planted in cylindrical plastic bags (26 cm in height, ~ 25 cm in diameter) filled with soil. To ensure the same amount of water was applied to each plant, a capillary watering system was applied. For the first four weeks, all plants were watered equally. Subsequently, the plants were divided into two groups: drought-primed plants (P) and control plants (C). P plants were not watered for the next three weeks, while C plants were still optimally irrigated. After this period, optimal irrigation was restored for both groups of plants. Tubers produced from P and C plants were harvested. In 2022, tubers from P and C plants were planted and the young plants were used for grafting. Ten 4-week-old C plants were used as rootstocks. Scions were prepared from five P and five C plants. Rootstock plants were cut approximately 8–10 cm above soil level and 1–2 cm deep slits were made. Scions were then inserted successfully into these slits. The graft junction was wrapped in plastic film. The grafted plants were kept in a mist chamber for 1–2 weeks until scions showed growth. After this time, they were acclimatized to normal growth conditions. The selection of three C/C and three P/C grafted plants was made based on plant vigor, after which they were planted in a tent with optimal irrigation. At the end of the vegetation period, tubers from C/C and P/C plants were harvested separately and stored at 6 °C. They were used as planting materials in the next-season experiment. In 2023, 18 tuber pieces were cut from C/C and P/C tubers to obtain C/C- and P/C-derived plants, respectively. They were planted in a greenhouse in accordance with the protocols outlined in our previous publications (Sołtys-Kalina et al. 2016; Plich et al. 2020). After 4 weeks, the C/C- and P/C-derived plants were transferred to a plastic tent and divided into two groups. The first group of nine plants was subjected to drought stress for 3 weeks (treatment D), while the second group of nine plants received optimal irrigation (treatment W). These plants were designated as C/C_D and P/C_D for treatment D, and C/C_W and P/C_W for treatment W. The experiment was set up in a randomized complete block design with three blocks (replications) and three plants per block. Tubers produced by each plant were harvested and weighed individually. The mean tuber yield per plant was calculated for each replication and the results were statistically analyzed. To assess the influence of the scion-based drought stress memory and the factor “treatment” on tuber yield, two-way ANOVA (analysis of variance) and planned comparisons (contrasts) were applied. The relative decrease in tuber yield under water deficit was calculated according to the following formula: 100⁎[(mean value of tuber yield per plant from control—mean value of tuber yield per plant from the drought stress)/(mean value of tuber yield per plant from control)]. In 2024, the second tuber generation of C/C and P/C plants that were vegetatively propagated under treatment D in 2023 were evaluated. These plants were designated as C/C_DW, P/C_DW, C/C_DD, and P/C_DD, respectively. To assess the effect of examined factors on the tuber yield, we applied the same statistical analyses as in 2023. Figure 1 shows the schematic diagram depicting the generation of the plant materials used for the experiments. Schematic overview of the plant material used in experiments. C/C: non-stressed Katahdin/non-stressed Katahdin graft; P/C: primed Katahdin/non-stressed Katahdin graft; C/C_D and P/C_D: plants derived from the tubers of C/C and P/C grafts, respectively, that were grown in drought conditions (D); C/C_W and P/C_W: plants derived from the tubers of C/C and P/C grafts, respectively, that were grown in control conditions (W); DD and DW: plants derived from the tubers of C/C_D and P/C_D that were subjected to repeated drought stress (D) or controlled watering (W) conditions Measurements of chlorophyll a fluorescence Forty-two JIP-test parameters were measured on 4-week-old C/C_W, P/C_W, C/C_D, and P/C_D plant samples according to the procedure described by Plich et al. (2020). For each group, 30 samples were evaluated (two treatments, three plants per treatment, and five replications per plant). A multifunctional plant efficiency analyzer (Handy-PEA fluorimeter, Hansatech Instruments Ltd., Pentney, King’s Lynn, Norfolk, UK) and Pea Plus software were employed. Leaf clips were applied to the middle region of third- and fourth-level, fully sun-exposed, mature apical leaflets. The mean values of P/C_W and P/C_D were calculated for each replicate and transformed into percentages of the respective C/C_W and C/C_D control values. Statistical differences between the scion-based primed and control plants were analyzed using the factorial analysis of variance (ANOVA). Differences were considered significant at P<0.05. Data from measurements of 15 selected parameters were also subjected to Principal Component Analysis (PCA). All these analyses were carried out using MS Excel and STATISTICA 12. Isolation of total RNA and transcriptome analysis Total RNA was isolated from leaves of C/C_W and P/C_W plants in three biological replicates, using the commercial Direct-zol RNA isolation kit (Zymo Research, Irvine, CA, USA). RNA concentration and quality was determined using a NanoDropLite Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). The quality of RNA (RIN >8.0) was defined by Bioanalyzer 2100 (Agilent Technologies, Santa Clara, CA, USA). RNA integrity of was also checked on agarose gel. RNA-seq was performed by Genomed S.A. (Warsaw, Poland) using the Illumina NovaSeq 6000 sequencing platform. Transcriptomic data were analyzed according to Machaj et al. (2018) as described in Marczewski et al. (2026). In brief, rRNA contaminations were removed by Bowtie2 v. 2.3.5.1 (Langmead and Salzberg 2012) mapping RNA-seq data to Solanum rRNA sequences downloaded from RNAcentral_v_20_0 (RNA central Consortium 2021). Cleaned reads were mapped to the Potato_dAg1_v1.0 genome assembly (Freire et al. 2021) using a 2-pass alignment approach with STAR v. 2.7.3a (Dobin et al. 2013). The reference annotation, provided by Freire et al. (2021) (Potato_dAg1_v1.0_gene_annotation.all.gff3), was merged with new isoforms identified with StringTie v. 2.1.1 (Pertea et al. 2015) using Cuffmerge v.1.0.0 (Trapnell et al. 2010). Gene expression was estimated using Kallisto v.0.44 (Bray et al. 2016). Differentially expressed genes (DEGs) were identified using three R packages: DESeq2 v.1.38.1 (Love et al. 2014), EBSeq v.1.26.0 (Leng et al. 2018), and edgeR v.3.14.0 (Robinson et al. 2010). Heatmaps were generated in R using ComplexHeatmap package v.2.14.0 (Gu 2022). Genes with false discovery rate (FDR) ≤0.05 reported by all three algorithms were selected as the final set of DEGs. All cleaned reads have been deposited in the Sequence Read Archive (SRA) database under BioProject PRJNA956351 (Table S1). GO and KEGG enrichment analysis Pathway and Gene Ontology (GO) enrichment analyses for the identified DEGs were performed using KOBAS v3.0.3 (Bu et al. 2021) with the Kyoto Encyclopedia of Genes and Genomes (KEGG) metabolic pathway database (Kanehisa and Goto 2000), the Gene Ontology database (Ashburner et al. 2000), and Solanum tuberosum background. The most enriched pathways were selected and plotted in R v4.2.1 using the dplyr v1.1.4 (Wickham et al. 2023), stringr v1.5.0 (Wickham 2022), and ggplot2 (Wickham 2016) packages. The AnnotationDbi v1.58.0 (Pagès et al. 2022) and GO.db v3.15.0 (Carlson 2022) R packages were used to map GO terms annotated by KOBAS (using -s sot; Solanum tuberosum) into Cellular Component (CC), Molecular Function (MF), and Biological Process (BP) categories. Whole-genome bisulfite sequencing (WGBS) In 2023, leaf samples were collected from C/C_W and P/C_W plants, as well as from the C/C_D and P/C_D plants three days after drought release. In 2024, leaf samples were collected from progeny of C/C_D and P/C_D, optimally watered C/C_DW and P/C_DW plants, as well as from the plants three days after drought release (C/C_DD and P/C_DD). Samples were collected in three biological replicates for each variant. DNA was extracted from 200 mg of fresh fully expanded and third-level mature apical leaves of plants from three biological replications using DNeasy Plant Mini Kit (Qiagen, Hilden, Germany). The amount of DNA and its quality were verified using NanoDrop Lite Spectrophotometer (Thermo Fisher Scientific) and gel electrophoresis. A total of 24 DNA samples were subjected to directional WGBS. Sequencing was performed by Genomed SA on the Illumina NovaSeq 6000 sequencing platform in PE150 mode (Table S1). WGBS data were analyzed as described in Macko-Podgórni et al. (2025), using the Bismark pipeline v. 0.23.1 (Krueger and Andrews 2011). Bowtie2 v. 2.3.5.1 (Langmead and Salzberg 2012) was applied to map reads to potato reference genome dAg1_v1.0 (Freire et al. 2021). The quality of the WGBS datasets generated in 2023 was lower as compared to the 2024 datasets, as indicated by the reduced fraction of mapped reads (Table S1). To assess whether the obtained sequencing depth was sufficient for downstream methylation analyses, genome coverage was estimated using the following equation: [Nr, total number of sequencing reads; Lr, read length; Fm, fraction of mapped reads; Gs, genome size (812 Mb]). The potato genome size was assumed to be 2×812 Mb, corresponding to the tetraploid genotypes analyzed in this study. Despite the lower mapping efficiency in 2023, the estimated minimum coverage of 5× was considered sufficient for the identification of 1 kb-long DMRs (Ziller et al. 2015). Furthermore, because the sequencing depth and genome coverage were higher in 2024, methylation changes established in 2023 and maintained across generations were expected to remain detectable in the 2024 datasets, as discussed below. Genome-wide methylation was inferred in three contexts (CpG, CHG and CHH) and used to identify 1 kb long DMRs with MethylKit v.1.26.0 R package (Akalin et al. 2012) using mincov = 5 and minqual = 20. DMRs shared among cultivars were assessed as the number of DMRs in the same genomic region. Results Effect of scion-based memory of drought stress on tuber yield In 2023, we compared tuber yield of the first generation of tuber progeny, which were derived from either drought-primed Katahdin/non-stressed Katahdin grafts (P/C-derived plants) or self-grafted, non-stressed Katahdin plants (C/C-derived plants). The latter served as controls. Tuber yield was evaluated in plants grown under drought stress (C/C_D and P/C_D) or with optimal watering (C/C_W and P/C_W). To determine the effect of scion-based memory of drought stress, we analyzed tuber yield using a two-way ANOVA that included the priming × treatment interaction. This analysis revealed a significant effect of the drought treatment (P = 0.003), but neither the priming factor (C/C-derived versus P/C-derived) nor the priming × treatment interaction was statistically significant (P = 0.194 and P = 0.215, respectively). The tuber yield of C/C_W and P/C_W plants was similar (Fig. 2A and B, blue vs. blue). Under drought conditions, the tuber yield of P/C_D plants was higher than that of C/C_D plants (Fig. 2A and B, red vs. red), but this difference was not statistically significant. Planned comparisons showed that drought significantly reduced tuber yield in C/C-derived plants (Fig. 2A: C/C blue vs. C/C red) by 23.3% (P = 0.004), whereas the 12.1% reduction in P/C-derived plants (Fig. 2B: P/C blue vs. P/C red) was not statistically significant (P = 0.074). Comparisons of the mean tuber yield of potato plants grown in control (blue) and drought (red) conditions. Panel 2023 compares the mean tuber yield collected from C/C_W and C/C_D plants (A), and P/C_W and P/C_D plants (B). Box-plot show results of two-way ANOVA: box, mean value; whiskers, confidence interval (CI). Factor ‘priming’, P = 0.194; factor ‘treatment’, P = 0.003; ‘priming × treatment’ interaction, P = 0.215. Other comparisons: C/C_D vs P/C_D, ns; C/C_W vs P/C_W, ns; P/C_W vs P/C_D, ns; C/C_W vs C/C_D, P = 0.004; ns stands for no significant difference. Panel 2024 shows the mean tuber yield of plants, which were grown from P/C_D tubers harvested in 2023: C/C_DW and C/C_DD plants (C), and P/C_DW and P/C_DD plants (D). Factor ‘priming’, P = 0.867; factor ‘treatment’, P = 0.005; ‘priming × treatment’ interaction, P = 0.410 In 2024, the focus was on plants derived from C/C_D and P/C_D tubers. These plants were again either drought-stressed or optimally watered (Fig. 2C, D). Two-way ANOVA revealed a significant effect of drought treatment on tuber yield (P = 0.005), whereas the priming factor (P = 0.867) and the priming × treatment interaction (P = 0.410) were not statistically significant. Planned comparisons showed a significant reduction in tuber yield in C/C-derived plants under drought (Fig. 2C, blue vs. red; P = 0.011), whereas the corresponding reduction in P/C-derived plants (Fig. 2D, blue vs. red) was not statistically significant (P = 0.074). Effect of scion-based memory of drought stress on physiological measurements In 2023, 42 JIP-test parameters were measured in C/C_W, P/C_W, C/C_D, and P/C_D plants. Exposure to drought stress caused insignificant changes in all evaluated parameters, when P/C_D plants were compared to C/C_D plants. The scion-based drought stress memory factor was statistically significant (P < 0.05) for 15 JIP-test parameters when comparing C/C_W and P/C_W plants (Table S2). They are related to fluorescence variability (Fm, Fv and Vi), light energy absorption (ABS), energy trapping (Tr), electron transport flux (Et), dissipation energy flux (Dl), electron flux reduction (Re), the reaction center (RC), and the area above the fluorescence transient (Area). For P/C plants, parameters Area, Fm, Fv, phi(Ro), Reo/CSo, ABS/CSm, Tro/CSm, Reo/CSm, γ(RC)/((1-γ(RC)), and PItotal increased, whereas Vi, ABC/RC, Dlo/RC, Tro/RC, and Eto/RC decreased (Fig. 3). The drought-stress memory values of these parameters ranged from a decrease of 20% for Dlo/RC to an increase of 170% for PItotal compared to C/C plants (Table S2). A “spider plot” showing the effect of the scion-based drought stress memory on the variability of JIP-test parameters in leaves of plants C/C_W, C/C_D, P/C_W, and P/C_D. Each dataset was normalized as percentage of changes in P/C-derived plants as compared to C/C-derived plants. Definitions of the JIP-test parameters are provided in Table S3 Transcriptome analysis and identification of candidate genes The experiments were performed in 2023 on the first tuber progeny. Analysis of expression of genes was performed on plants grown under optimal irrigation (P/C_W and C/C_W), and on those subjected to water deficit (P/C_D and C/C_D). Under control conditions (W), 297 DEGs were identified between P/C_W and C/C_W plants. Of these, 252 genes exhibited higher expression in P/C_W plants (Table S4). GO and KEGG enrichment analyses revealed significantly upregulated genes associated with photosynthesis (Fig. 4A and B, Fig. S1A and B, Table S5). Of the photosynthesis-related genes, some were found to encode chlorophyll a/b binding proteins, including both major light-harvesting complex proteins (dAg1_00025354—CAB13, dAg1_00042911 and dAg1_00018383—CAB2R, dAg1_00018773 and dAg1_00042909—LHCII, dAg1_00018381—LHCA2, dAg1_00025385 and dAg1_00022566—LHCA3, dAg1_00034464—LHCA4, dAg1_00026022—LHCB2.1) as well as minor antenna proteins (dAg1_00004712—CAP10A, dAg1_00004715—AP10B). To further characterize photosynthesis-related genes, differentially expressed between P/C_W and C/C_W plants, we examined the expression of all annotated genes that encode components and auxiliary proteins of photosystem I (PSI) and photosystem II (PSII), as well as electron transport proteins and the light-harvesting antenna proteins associated with these complexes. The DEGs identified as encoding LHCA and LHCB light-harvesting antenna proteins, as well as PSI core components, generally exhibited higher expression levels than other gene copies belonging to the same groups. Although no significant changes in the expression of genes involved in electron transport or PSII core components were observed, the mean expression of highly expressed genes within these groups was marginally higher in P/C_W plants (Fig. 5A and B). These results suggest a tendency toward improvement in light capture, which could contribute to higher photosynthetic efficiency. We also observed the overexpression of several genes typically associated with abiotic stress responses. The most prominent of these were the key genes involved in the biosynthesis of the sugar molecule raffinose: galactinol synthase 1 (dAg1_00043964, GOLS1) and probable galactinol–sucrose galactosyltransferase 1 (dAg1_00041229, RFS1). Top GO terms associated with biological processes (BP) (A, C) and KEGG pathways (B, D) for DEGs when P/C_W vs. C/C_W plants were compared (A, B), and DEGs for the C/C_D vs C/C_W comparison (C, D) Heatmaps showing expression levels (TPM) of all annotated genes related to photosynthesis (A, B) and the localization of their products within the plastid membrane, based on a modified version of the photosynthesis–antenna proteins pathway for Solanum tuberosum (potato; sot00196), as well as genes involved in DNA replication and repair (C) and chromatin remodeling and methylation (D, E). Genes with mean TPM values < 0.1 were considered not expressed and are not shown. Genes with significantly higher expression (P < 0.05) in P/C_W compared with C/C_W are highlighted in red In P/C_W plants, we further identified upregulation of genes associated with DNA replication and DNA-repair processes (GO:0000727, GO:0006267, GO:0006268, GO:0006260, GO:0006289, GO:0000724), including dAg1_00002727 (MCM7: DNA replication licensing factor MCM7), dAg1_00017709 (CDKF-1: Cyclin-dependent kinase F-1), dAg1_00032058 (CDT1A: CDT1-like protein chloroplastic), dAg1_00042888 (ROA2: DNA replication licensing factor MCM3 homolog 2), dAg1_00043675 (MCM6: DNA replication licensing factor MCM6), dAg1_00021303 (RPA1B: Replication protein A 70 kDa DNA-binding subunit B), and dAg1_00039203 (RPA2A: Replication protein A 32 kDa subunit A) (Fig. 5C). Additionally, the upregulation of genes encoding the H2AX variant of histone H2, which is involved in the recognition and repair of DNA damage, as well as dAg1_00029979 coding for RECQL5, a member of the RecQ helicase family, suggests enhanced activity of DNA maintenance mechanisms. RecQ helicases are key factors involved in multiple DNA repair and recombination pathways. Furthermore, the increased expression of dAg1_00029078 encoding ASF1 H3/H4 histone chaperone, which facilitates the removal and/or remodeling of nucleosomes around DNA damage sites, and some of H3/H4 histone coding genes, further indicates activation of processes involved in maintaining genome integrity (Fig. 5D). In addition to these genes, we also observed the upregulation of genes potentially involved in chromatin remodeling and DNA methylation. These include the ATP-dependent DNA helicase DDM1; the chromatin remodeler decrease in DNA methylation 1 (dAg1_00042113); the DNA (cytosine−5)-methyltransferase CMT1 (dAg1_00028585); methyltransferase 1 (MET1; dAg1_00014109); and the WD-40 repeat-containing protein MSI2 (dAg1_00005476) (Fig. 5E). The overexpression of three additional genes may also indicate increased activity of processes related to chromatin remodeling. One of these genes is WDR5A, coding for COMPASS-like H3K4 histone methylase component (dAg1_00044808). However, a total of 30 genes is annotated as WDR5A in the potato genome, and most of them exhibited higher expression levels than dAg1_00044808 and displayed similar expression patterns in both groups of tested plants (Fig. S2). Other core components of the COMPASS complex or Histone-lysine N-methyltransferase genes were not differentially expressed (Fig. S2). Two other genes, coding for PMT26 (dAg1_00011761), belonging to the poorly characterized Probable methyltransferase PMT family, and MED20A (dAg1_00025719), a component of the head module of the Mediator complex that plays a key role in transferring signals from transcriptional regulators to RNA polymerase II in eukaryotes (Chereji et al. 2017), also showed increased expression. However, both genes have multiple copies in the genome, and other copies exhibited higher and relatively uniform expression levels across all samples. Fewer DEGs were found when compared to the gene expression in P/C_D and C/C_D plants. Among the 38 DEGs identified under drought conditions, 12 were annotated as proteins of unknown function (Table S6). GO and KEGG enrichment analyses revealed that stress-responsive genes associated with protein folding were down-regulated in P/C_D plants compared to C/C_D plants (Table S7). We conducted two further comparisons: the first was between C/C_D and C/C_W plants, and the second was between P/C_D and P/C_W plants. A strong drought response was observed between the C/C_D and C/C_W plants, as reflected by the differential expression of 1,203 genes (Table S8). Some of the 793 genes that were upregulated were significantly enriched in photosynthesis, while others were involved in the biosynthesis of amino acids, protein processing in the endoplasmic reticulum, the elongation of fatty acids, ribosome biogenesis, the citrate cycle, DNA replication, and the repair of double-strand breaks via break-induced replication, as well as water channel activity (Fig. 4C and D, Table S9, Fig. S1C and D. This indicates the response to the applied stress. By contrast, the response to drought stress observed in P/C_D plants was less pronounced. Only 48 DEGs were revealed in the P/C_D vs P/C_W comparison (Tables S9–S10). GO and KEGG analyses did not reveal any significant enrichment of terms related to the drought stress response or photosynthesis (Table S11). However, 7 of the 17 upregulated genes were also upregulated in C/C_D plants, four of which had been reported to be drought responsive (Table S12). We also observed that 70 of the 252 genes that had increased expression in P/C_W plants compared to C/C_W plants (Table S4) were upregulated in the C/C_D vs C/C_W comparison (Tables S8 and S12). Of these genes, those associated with photosynthesis were over-represented (Fig. 6A and B). In PCA analysis, the PC1 component determined about 31% of gene expression changes in C/C_D, C/C_W, P/C_D, and P/C_W plants. The PC2 component explained 23.1% of the variance. The PCA biplot shows that P/C_W plants cluster closer to P/C_D and C/C_D than to C/C_W (Fig. 6C). A similar grouping of plants was observed in the PCA analysis of the JIP-test parameters (Fig. 6D). Top GO terms (A) and KEGG pathways (B) for 70 upregulated genes in P/C_W plants (Table S4) that were upregulated in the C/C_D vs C/C_W comparison; (C, D) PCA grouping of the tested plants based on normalized gene expression (C) and JIP-test parameters (D) Effect of scion-based memory of drought stress on DNA methylation In 2023, four groups of plants were studied: C/C_D, P/C_D, C/C_W, and P/C_W. The global DNA methylation levels were the highest in the CG context (57–66%), followed by CHG (32–36%), and CHH (9–10%) sequence contexts. The methylation levels were similar between P/C and C/C plants grown under the same conditions, except for lower methylation level in CG and CHH sequence context in P/C plants to C/C plants, both grown under stress conditions (Fig. 7A). Genome-wide analysis of DMRs was used to find differences in DNA methylation patterns between P/C and C/C plants under both well-watered (W) and drought (D) conditions. The identified DMRs were classified according to their genomic localization (upstream, gene body, downstream, and intergenic regions) and sequence context, i.e., CpG, CHG, and CHH (Fig. 7B and C, Table 1 and Tables S13-S18). Under well-watered conditions (P/C_W vs C/C_W), a total of 1,139 DMRs were identified, comprising 17% CpG, 16% CHG, and 67% CHH DMRs. In contrast, under drought conditions (P/C_D vs C/C_D), 913 DMRs were detected, including 14% CpG, 13% CHG, and 73% CHH DMRs. In both comparisons, the CHH sequence context contained the highest number of DMRs, accounting for approximately two-thirds of all identified regions, although global methylation levels were not significantly different between P/C_W and C/C_W. Analysis of CpG methylation revealed that, under well-watered conditions, the number of hypermethylated and hypomethylated DMRs was relatively balanced (Table 1). The largest proportion of CpG DMRs was located within gene bodies (50% DMRs), followed by intergenic regions (30% DMRs), upstream regions (11% DMRs), and downstream regions (9% DMRs). Under drought conditions, the total number of CpG DMRs decreased to 130, with hypomethylated regions exceeding hypermethylated regions. In this case, intergenic regions represented the predominant genomic location (42% DMRs), followed by gene bodies (36% DMRs), whereas upstream and downstream regions contained 15% and 7% DMRs, respectively. Characterization of methylation and differentially methylated regions contexts in tested plants. Global DNA methylation levels in the CpG, CHG, and CHH sequence contexts in plants grown in 2023 under drought stress (D) and control (W) conditions (A). P values from Student's t tests comparing C/C and P/C plants within each treatment are shown above the box plots. The number (B) and proportion (C) of DMRs identified between P/C vs. C/C plants grown in 2023 under drought stress and control conditions assigned to four genomic localizations. The number of DMRs shared among tested plants. D: P/C_D vs C/C_D; DD: C/C_DD vs P/C_DD; DW:C/C_DW and P/C_DW; W: P/C_W vs C/C_W, DD) in CpG (D), CHG (E), and CHH (F) contexts. DMRs shared among P/C_D vs C/C_D (2023), P/C_DW vs C/C_DW (2024), and P/C_DD vs C/C_DD (2024) are labeled with red circle Table 1DMRs identified for the C/C_W and P/C_W comparison, as well as between P/C_D and C/C_D plantsComparisonGenomic localization of DMRsCpGCHGCHHDMRsHyperHypoDMRsHyperHypoDMRsHyperHypoP/C_W vs C/C_WUpstream2281411741397861P/C_W vs C/C_WDownstream188101495884741P/C_W vs C/C_WGene body98534566273921713384P/C_W vs C/C_WIntergenic582434875532321160161P/C_W vs C/C_WTotal196931031789880765418347P/C_D vs C/C_DUpstream191099631306169P/C_D vs C/C_DDownstream9361156612338P/C_D vs C/C_DGene body47182941202117210666P/C_D vs C/C_DIntergenic552035603327299137162P/C_D vs C/C_DTotal13051791216457662327335 DMRs identified for the C/C_W and P/C_W comparison, as well as between P/C_D and C/C_D plants For the CHG context, 178 DMRs were identified in the P/C_W vs C/C_W comparison, including 98 hypermethylated and 80 hypomethylated regions. The highest number of CHG DMRs occurred within intergenic regions (49%), followed by gene bodies (37%), downstream (8%), and upstream regions (6%). Under drought conditions, the total number of CHG DMRs decreased to 121, comprising 64 hypermethylated and 57 hypomethylated regions. Similarly to plants grown under control conditions, DMRs in intergenic regions (50%) and gene bodies (34%) represented the predominant locations of CHG methylation changes. The greatest number of methylation differences was observed in the CHH context. However, no significant changes in gene expression were detected for most genes showing differential methylation between the groups studied (Table S19). Only nine DMRs were located within genes or their flanking regions when comparing C/C_W and P/C_W plants. All but one of these were associated with differences in CHH methylation (Table S19). In 2024, like in 2023, the highest methylation level was observed in CG sequence context (66–71%), followed by CHG (40–45%), and CHH (8–10%) (Fig. S3). To identify DMRs that are transmitted from 2023 plants to the next generation two types of comparison were made: C/C_DW vs. P/C_DW, and C/C_DD vs. P/C_DD (Tables S20–S25). Overlapping the identified DMRs for each group of plants showed that most DMRs were unique to each group (Table S26). Of the shared DMRs, only one located within the dAg1_00002385 gene body was hypermethylated in the CHH context in the parental plants and their progeny grown in 2023 under both conditions (Fig. 7 D–F, Tables S16–S18). The observed hypermethylation did not affect the expression of this gene in 2023. Discussion Prior exposure of plants to a transient stress is referred to as priming that can lead to a modified response of plants to subsequent stresses (Lämke and Bäurle 2017). Plants have ‘stress memory’, which can have positive or negative effects on their future growth and reproduction when exposed to the same stress factor again (Crisp et al. 2016). In plants, grafting affects morphological, physiological, and molecular responses related to environmental stresses (Razi et al. 2024). It has been suggested that grafting can cause priming drought stress due to a lack of water supply resulting from xylem disconnection (Fuentes-Merlos et al. 2023). Our study revealed a significant decrease in tuber yield in C/C_D plants compared to C/C_W plants. Prior exposure of Katahdin plants to drought stress and using them as scions reduced the decline in tuber yield in their vegetative progeny under drought conditions. A decrease in tuber yield was not statistically significant when P/C_D and P/C_W plants were compared. The similar results were observed for the second generation of plants, which were grown from P/C_D tubers harvested in 2023. In P/C_DD plants, the decrease in tuber yield was insignificant compared to that in P/C_DW plants. Although the priming × treatment interaction was not statistically significant in either experiment, the same response pattern was observed consistently in two independent experiments conducted in consecutive years. However, given the limited number of biological replicates (n = 3), these observations should be interpreted with caution. Nevertheless, the consistency of this trend across two independent experiments supports its robustness. The results are in contrast with those obtained in our previous studies of non-grafted potato plants, as we reported that the ‘primed/non-primed’ and ‘cultivar’ factors significantly affected tuber yield in control conditions (Macko-Podgórni et al. 2025). This suggests that the stress imprint has not been reset in these cultivars when grown under non-restricted water conditions. For the Katahdin cultivar, a significant influence of drought stress on tuber yield was observed for both drought-primed and non-primed plants (Macko-Podgórni et al. 2025). Our results suggest that scion-based drought stress memory may contribute to buffering drought-induced reductions in tuber yield in the vegetative progeny of grafted plants that were subsequently exposed to drought stress. Drought stress hinders photosynthesis, reducing plant growth. Although decreases in stomatal conductance reduce water loss through transpiration, they also decrease the uptake of carbon dioxide, thereby reducing the production of photoassimilates and stunting plant growth. In addition to the limitations of photosynthesis caused by stoma under severe water stress, limitations unrelated to the stoma have been reported (Rodríguez-Pérez et al. 2017). These are related to damage to the photosynthetic apparatus. Our previous studies have shown that transgenic potato plants exhibit higher tolerance to drought, exhibit ABA-hypersensitive stomatal closure and increased leaf stomata and trichome density, and have compact cuticle structures with fewer microchannels (Pieczyński et al. 2013; Wyrzykowska et al. 2022). Under drought conditions, transgenic potato plants that ectopically expressed an endogenous annexin were more tolerant of water deficiency. These plants preserved more water in their green tissues, maintained their chloroplast functions, and accumulated higher levels of chlorophyll b and xanthophylls than well-watered control plants (Szalonek et al. 2015). We found that the cell cycle in potato leaves is modulated in response to drought stress and may significantly affect the final tuber yield (Sołtys-Kalina et al. 2023). These findings demonstrate the complexity of the plant’s response to drought stress. The JIP test is a useful tool for screening plant photosynthetic performance, providing information about the structure, conformation, and function of the photosynthetic apparatus (Strasser et al. 2000). Changes in JIP parameters due to drought were observed in both the tolerant and sensitive groups, but the changes in the sensitive group were much more pronounced (Plich et al. 2020). Recently, we reported on the effect of drought-stress memory on the photosynthetic performance in non-grafted potato cultivars. Changes in JIP-test parameters corresponded with both the gene expression landscape and potato tuber productivity (Marczewski et al. 2026). In the present study, we observed significant differences in 15 JIP-test parameters between P/C_W and C/C_W plants. For P/C_D and C/C_D plants, the differences in chlorophyll fluorescence measurements were not significant. The present study revealed many more DEGs were revealed between P/C_W and C/C_W plants than in the P/C_D vs C/C_D comparison. Most of these genes exhibited higher expression in P/C_W plants and were associated with photosynthesis. Interestingly, 28% DEGs identified in the P/C_W vs. C/C_W comparison, including those associated with photosynthesis, were also found to be upregulated in the C/C_D vs. C/C_W comparison. These results may suggest that scions from drought-primed plants trigger transcriptional memory in the vegetative progeny plants, particularly in genes associated with photosynthesis. Of the genes that were down-regulated in the P/C_D vs C/C_D comparison, only two genes were annotated with GO terms including ‘response to heat’, ‘response to salt stress’, ‘protein complex oligomerisation’, ‘response to hydrogen peroxide’, ‘response to reactive oxygen species’, and ‘protein folding’. This suggests that the enrichment signal may reflect the contribution of a small number of genes that respond broadly to stress rather than a specific response to drought. In addition to the activation of photosynthesis-related genes in P/C_W plants compared with C/C_W plants, we observed increased expression of genes involved in raffinose biosynthesis, which is often considered as a drought stress marker (Li et al. 2020). In response to stress, raffinose and related oligosaccharides may contribute to membrane stabilization, antioxidant defense, and potentially signaling processes. These compounds may help maintain photosystem II stability and function as ROS scavengers, thereby playing a previously underappreciated role in protecting cellular metabolism under stress conditions (Sengupta et al. 2015). Plants exposure to various environmental stresses, including drought, can lead to the accumulation of reactive oxygen species (ROS) and indirectly promote DNA damage. To cope with these challenges, plants activate DNA damage response (DDR) pathways, which consist of both highly conserved mechanisms and plant-specific components (Pedroza‐Garcia et al. 2022). The activation of mechanisms involved in responding to replication stress is also indicated by overexpression of dAg1_00017709 (CDKF-1), dAg1_00032058 (CDT1A), dAg1_00042888 (MCM3), dAg1_00043675 (MCM6), and dAg1_00002727 (MCM7) MCM genes, which are associated with DNA replication and repair. Importantly, two of these genes, dAg1_00039203 (RPA2A) and dAg1_00021303 (RPA1B), encode components of the RPA complex, which is involved in the response to DNA damage, homologous recombination, and stabilization of damaged DNA sites (Gan et al. 2023). We also observed the upregulation of histone H2AX coding genes, whose phosphorylated form (γH2AX) is enriched at DNA double-strand break (DSB) sites and plays a crucial role in their recognition and repair by recruiting signaling and repair factors (Lu et al. 2026). Increased expression of the dAg1_00029979 gene encoding RECQL5 suggests enhanced activity of DNA repair and genome surveillance mechanisms under stress conditions as RecQ helicases are key regulators of multiple DNA repair and recombination pathways involved in maintaining genome stability. Moreover, several RecQ helicases have been shown to act as positive regulators of abiotic stress tolerance during the germinative and post-germinative phases in Arabidopsis (Dutta et al. 2024). Recent studies in humans have demonstrated that RECQL5 promotes RNA polymerase II transition toward a post-translocation state, which may facilitate the restart of Pol II elongation and reduce transcription-associated stress (Florez Ariza et al. 2025). Therefore, increased expression of RECQL5 may reflect an adaptive response aimed at maintaining genome stability and supporting efficient recovery of essential cellular processes during drought-induced stress. Because plants are sessile organisms and cannot escape unfavorable environmental conditions, they are continuously exposed to various abiotic stresses. This has led to the evolution of complex adaptive and defense strategies. Epigenetic mechanisms, including chromatin remodeling, nucleosome positioning, and DNA methylation, are key components of these responses, enabling dynamic regulation of gene expression and enhancing plant adaptation to environmental stress (Lämke et al. 2016). Histone marks are another fundamental type of epigenetic modification associated with the transcriptional response to repeated exposure to drought stress (Sadhukhanet al. 2022). Evidence for enhanced chromatin modifications may by suggested by the overexpression of MSI2. Due to its high sequence similarity to the WD40 repeat-containing protein MSI1, a core component of the Polycomb Repressive Complex 2 (PRC2), MSI2 may be associated with epigenetic silencing processes. PRC2 catalyzes the deposition of the repressive histone mark H3K27me3 and plays a central role in maintaining stable gene repression and the formation of facultative heterochromatin. However, the previous studies have demonstrated that MSI1 performs specialized functions within PRC2-mediated epigenetic regulation that cannot be fully compensated for by MSI2 (Derkacheva et al. 2013). H3K4me3 is one of the most well-characterized active chromatin marks and is often found in transcriptionally permissive genomic regions. As demonstrated by Lämke et al. (2016) in Arabidopsis, H3K4me3 can be also considered as a key histone mark for transcriptional stress memory. The COMPASS-like complex interacts with histone methyltransferases, facilitating transcription preinitiation complex (PIC) assembly and promoting H3K4me3 deposition during transcriptional activation (Temel and Gören-Sağlam 2026). In the present study, we observed increased expression of a single, low-expressed copy of the 30 annotated WDR5A, without any change in the expression of other essential COMPASS-like complex components or the catalytic H3K4 methyltransferase (e.g., SDG2/ATXR3 or related family members). In Arabidopsis, the ASH2–RBL–WDR5A complex associates with different H3K4 methyltransferases via WDR5A, forming multiple functional COMPASS-like H3K4 methyltransferase complexes (Jiang et al. 2011). In our dataset, none of the annotated histone-lysine N-methyltransferase genes were differentially expressed. However, a parental H3–H4 tetramer can be transferred to either the leading or lagging strands of DNA replication forks and memorize its position along the DNA during mitotic cell division (Escobar et al. 2019). This suggests that, even though we do not observe increased expression of these genes, H3K4me3-marked histones may still be restored in daughter cells during mitotic cell divisions (Serra-Cardona et al. 2022). The role of DNA methylation in the response to drought stress in plants has been studied in many crops. These epigenetic changes can vary according to species, genotype and tissue. For example, in maize, the methylation level under drought stress increased in roots, and the methylation is less stable in lines sensitive to drought (Wang et al. 2021), while in barley, similar number of methylation and demethylation changes under the drought stress were observed, but the roots’ response was stronger than leaves (Chwialkowska et al. 2016). Most of the available studies suggest that DNA methylation plays an important role in the regulation of expression of stress-responsive genes (Rao et al. 2024). However, in two potato genotypes that respond differently to heat, variation in constitutive DNA methylation contributed more to genetic diversity than to the direct regulation of gene expression (Yeo et al. 2026). Previously, we showed changes in DNA methylation associated with drought-induced memory in potato for eight upregulated photosynthesis-related genes (Marczewski et al. 2026). Here, we observed no association between changes in DNA methylation and alterations in the expression of DEGs, nor transmission of DMRs to the next generation. Only nine DMRs were located within genes or their flanking regions when comparing C/C_W and P/C_W plants. All but one of these were associated with differences in CHH methylation. Unlike CG and CHG methylation, CHH methylation is not transmitted through DNA replication because of its asymmetric sequence context and is very dynamic (Harris and Zemach 2020). However, their constant maintenance through the RNA-directed DNA methylation (RdDM) pathway cannot be excluded. Considering the origin of the analyzed materials (clonal propagation of the initial plant material) and the fact that, for the majority of DMRs, nearby genes do not exhibit changes in expression, this result is not unexpected. Previous studies have shown that only approximately 10–20% of DMRs display a negative correlation between DNA methylation and gene expression (Springer and Schmitz 2017). Although only a few methylation changes were associated with significant changes in gene expression in P/C_W vs. C/C_W comparison, one of these occurred in the promoter region of the CMT1 gene, whose potential role in CHG and CHH methylation is discussed below. It cannot be excluded that, due to relatively low mapping rate of WGBS data in 2023 (resulting in coverage ranging from 5× to 6 ×), some DMRs characterized by small differences in methylation were not detected. However, a reanalysis of human WGBS data demonstrated that the number of DMRs identified from datasets sampled from an initial 30× coverage, increased with sequencing depth between 1× and 8×, whereas further increases in coverage had only a limited impact on DMR detection. The authors recommended a sequencing depth of 5×–15× per sample when using two biological replicates, depending on the magnitude of methylation differences between groups and on whether DMR identification was based on smoothing approaches or single-CpG analyses. They also concluded that increasing the number of biological replicates generally has a greater impact on DMR detection power than increasing sequencing coverage. Furthermore, when the primary objective is the identification of long DMRs exhibiting substantial differences in methylation, coverage as low as 1×–2× per sample was found to be acceptable (Ziller et al. 2015). We used three biological replicates, with a minimum sequencing coverage of 5× per sample. Therefore, despite the lower quality of the sequencing data obtained from plants grown in 2023, we believe that the sequencing depth and analytical parameters employed in this study were sufficient to reliably identify the majority of biologically relevant DMRs. Consequently, the absence of detectable DMRs in the 2024 samples is unlikely to be an analytical artifact resulting from insufficient coverage which was higher than in 2023. Instead, it is more likely reflected in biological mechanisms governing DNA methylation dynamics in the studied plant material. Ongoing methylation changes might be partially explained by overexpression of the chromatin remodeler DDM1 (Decrease in DNA Methylation 1), that we observed in the analyzed P/C_W plants. DDM1 is required for efficient heterochromatic DNA methylation, particularly in the CHG and CHH sequence contexts, which is mainly mediated by CMT2 independently of the canonical RdDM pathway (Bilichak and Kovalchuk 2016). Increased DDM1 expression may enhance the accessibility of heterochromatic regions to DNA methyltransferases and thereby promote their activity. On the other hand, DDM1 has also been reported to play a role in the rapid resetting of stress-induced epigenetic states, thereby preventing their mitotic propagation and transgenerational inheritance (Iwasaki and Paszkowski 2014). Therefore, it is difficult to determine whether the observed increase in the expression of this gene represents a residual effect of their stress-induced activity or, alternatively, reflects an active process of epigenetic memory resetting aimed at restoring the original chromatin state. Additionally, DNA methyltransferase 1 (MET1) and chromomethylase 1 (CMT1), a less well-characterized member of the CMT family, were also significantly overexpressed. MET1 is responsible for the maintenance of CG methylation in repetitive sequences, gene-body regions of actively transcribed genes, and certain loci where it can also exhibit de novo methylation activity. This function enables the restoration of lost gene-body methylation patterns and may contribute not only to stabilization but also to the diversification of gene-body methylation profiles (Springer and Schmitz 2017). In contrast, the role of CMT1 remains poorly understood. Although CMT1 has been suggested to participate in CHG and CHH methylation, these functions are more clearly established for CMT2, whose expression remained stable in the analyzed plants. It is also known that small RNAs (sRNAs) required for the RdDM pathway can move between cells through plasmodesmata (Melnyk et al. 2011). This mechanism could potentially explain the observed differences in gene expression between plants grafted with a primed scion and those grafted with a control scion. However, given that we did not detect any significant differences in DNA methylation within the promoter regions or gene bodies of the differentially expressed genes, nor any increased expression of genes encoding the key components of the RdDM pathway, it is questionable whether RdDM is the primary mechanism underlying the observed transcriptional changes. Instead, other chromatin-based regulatory mechanisms, such as histone modifications or changes in chromatin accessibility, may provide a coherent mechanistic framework for the stress memory, explaining why photosynthesis-related genes remain constitutively upregulated across vegetative generations in P/C progeny. We propose that scion-derived signals may drive the establishment of epigenetically active chromatin state. However, the mechanism leading to postulated epigenetic reprogramming has yet to be elucidated. Conclusions For the first time, we have demonstrated that signals of drought memory transmitted from scion to tubers affect the vegetative progeny of potato plants. Significant increases in gene expression were observed, primarily related to photosynthetic metabolic pathways, as well as in genes associated with chromatin modeling and the plant’s response to abiotic stresses. The productivity of potato tubers was positively influenced by scion-based drought stress memory. Potato tubers are strong sink organs. Their productivity depends heavily on the functioning of their above-ground organs. Further research is needed to understand the molecular mechanisms behind the memory of drought stress in potato plants, how this memory is transmitted to tubers, and how it is passed on to non-stressed progeny. This knowledge can help to determine the molecular breeding potential of scion-based drought memory in potato breeding practice. Supplementary Information Below is the link to the electronic supplementary material. Supplementary file1 (DOCX 1387 KB) Supplementary file2 (XLSX 7217 KB) Acknowledgements This work was supported in part by the National Science Centre in Poland, under Grant No. UMO-2020/37/B/NZ9/00028. Abbreviations Control plants Decrease in DNA Methylation Differentially expressed gene Differentially methylated regions Gene Ontology Kyoto Encyclopedia of Genes and Genomes Drought-primed plants RNA-directed DNA methylation Whole-genome bisulfite sequencing Author contributions WM: conceptualization, writing—original draft, project administration, methodology, and visualization. AM-P: bioinformatics, analysis, writing—original draft, and visualization. DB-M, JP: investigation, statistical analysis, and visualization. PS-D: material preparation DG: review and editing. All authors have read and approved the final manuscript. Data availability The datasets generated during the current study are available in the Sequence Read Archive (SRA) repository under the PRJNA956351 BioProject (link: https://www.ncbi.nlm.nih.gov/sra/PRJNA956351). Declarations Conflict of interest The authors declare no conflict of interest. Ethical approval The authors have nothing to report. Footnotes Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Waldemar Marczewski, Email: w.marczewski@ihar.edu.pl. Alicja Macko-Podgórni, Email: alicja.macko-podgorni@urk.edu.pl. References Associated Data Supplementary Materials Supplementary file1 (DOCX 1387 KB) Supplementary file2 (XLSX 7217 KB) Data Availability Statement The datasets generated during the current study are available in the Sequence Read Archive (SRA) repository under the PRJNA956351 BioProject (link: https://www.ncbi.nlm.nih.gov/sra/PRJNA956351).