# Source paper archive - Source title: Temporal genomic imprinting regulates grain filling in barley - Author(s): Kovacik, M., Lampar, A., Nowicka, A., Peirats-Llobet, M., Saly, P., Trafford, K., Korchanova, Z., Kaduchova, K., Strejckova, B., Rasmussen, M. W., Knudsen, S., Sreeprakash, A., Cervenkova, M., Dobisova, T., Dockter, C., Whelan, J., Lafon Placette, C., Pecinka, A. - Source landing URL: https://www.biorxiv.org/content/10.64898/2026.11.04.764152v1 - PDF URL: https://www.biorxiv.org/content/10.64898/2026.11.04.764152v1.full.pdf - PDF SHA-256: 452ddbad2c3f6a2996a1e2fd5134fea2de805efa4563089c7bbb253d0437ed11 - Converter: pymupdf4llm 0.0.27 - Limits: Automated PDF text extraction; reading order, equations, tables and formatting may be incomplete or incorrect. Images are omitted, and no OCR was performed. Markup and links are neutralized for safe downloading. Consult the original PDF for the authoritative version. --- [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764152; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.764152) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.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 [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 1 **Letter** 2 3 **Temporal genomic imprinting regulates grain filling in barley** 4 5 **Martin Kovacik** **[1]** **, Adam Lampar** **[1]** **, Anna Nowicka** **[1]** **, Marta Peirats-Llobet** **[2,3]** **,** **Peter Šály** **[1,4]** **, Kay** 6 **Trafford** **[5]** **,** **Zuzana Korchanová** **[1]** **, Kateřina Kaduchová** **[1]** **, Beáta Strejčková** **[1]** **, Magnus** 7 **Wohlfahrt Rasmussen** **[6]** **, Søren Knudsen** **[6]** **, Aparna Sreeprakash** **[7,8]** **, Marie Červenková** **[1]** **,** 8 **Tereza Dobisová** **[7]** **, Christoph Dockter** **[6]** **, James Whelan** **[9]** **, Clément Lafon Placette** **[10]** **, & Ales** 9 **Pecinka** **[1,*]** 10 11 1 Centre of Plant Structural and Functional Genomics, Institute of Experimental Botany, Czech 12 Acad Sci, Olomouc, Czech Republic. 13 2 La Trobe Institute for Sustainable Agriculture and Food, Department of Plant, Animal and Soil 14 Sciences, La Trobe University, AgriBio Building, Bundoora, VIC 3086, Australia. 15 3 Australian Research Council Centre of Excellence in Plants for Space, AgriBio Building, La Trobe 16 University, Bundoora, VIC 3086, Australia. 17 4 Department of Cell Biology and Genetics, Faculty of Science, Palacký University, Olomouc, 18 Czech Republic. 19 5 The John Innes Centre, Norwich Research Park, Norwich, NR4 7UH, UK. 20 6 Carlsberg Research Laboratory, Copenhagen, Denmark. 21 7 Labdeers Ltd., 680-01 Boskovice, Czech Republic 22 8 National Centre for Biomolecular Research, Faculty of Science, Masaryk University, Brno, Czech 23 Republic 24 9 State Key Laboratory of Plant Environmental Resilience, College of Life Science, Zhejiang 25 University, Hangzhou, Zhejiang 310058, P.R. China. 26 10 Department of Botany, Faculty of Science, Charles University, Praha, Czech Republic. 27 [*e-mail: pecinka@ueb.cas.cz](mailto:pecinka@ueb.cas.cz) 28 29 **Short title:** Temporal imprinting in barley endosperm 30 31 **One-sentence summary:** We identified imprinted genes in developing barley endosperm and 32 characterized transcription factor _LYS3_ as a temporally maternally expressed gene, uniparentally 33 regulating barley grain filling. 34 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764152; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.764152) (which 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 [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 35 **Keywords:** barley, grain, genomic imprinting, endosperm, LYS3, grain filling 36 37 **Abstract** 38 In cereals, the persistent endosperm accumulates starch and storage proteins, making their grains 39 a major source of food and feed [1–3] . Early grain development is epigenetically controlled through 40 the parent-of-origin-specific expression of imprinted genes in the endosperm that regulate 41 endosperm proliferation, nutrient accumulation and embryo nourishment [4–6] . Imprinted genes have 42 been identified in several major cereals, and a few have been functionally characterized [7–14] . 43 However, it is unclear how imprinted genes regulate the accumulation of storage compounds in 44 cereal grains. Through spatiotemporal transcriptome analysis of 255 imprinted genes in barley 45 ( _Hordeum vulgare_ ), we identify endosperm transfer cells in grains as a hotspot of expression of 46 imprinted genes. Furthermore, we show that the transcription factor _LYS3_ dynamically regulates 47 grain filling in barley through early expression of the maternal allele, followed by partial activation 48 of the paternal allele. 49 50 **Results and Discussion** 51 To identify imprinted genes in barley at the genome-wide level, we sequenced endosperm 52 transcriptome at 8 and 12 days after pollination (DAP), corresponding to cellularizing and 53 cellularized stages, respectively, from two different reciprocal F 1 hybrid grains: (i) cultivar Morex 54 (MOR) crossed with Ethiopian landrace (ETH), and (ii) MOR with wild barley (HS) (Fig. 1a). We 55 identified 255 candidate imprinted genes, including 150 maternally expressed genes (MEGs) and 56 105 paternally expressed genes (PEGs) (Fig. 1b, c, Supplementary Fig. 1a, Supplementary Data 57 Sets 1-5). The parent-of-origin specific expression was independently validated with highly 58 consistent results (Supplementary Fig. 2). Substantial numbers of unique MEGs and PEGs in 59 individual crosses were mostly due to hybrid-specific diagnostic SNPs (Supplementary Fig. 3). We 60 observed similar numbers of imprinted genes between 8 and 12 DAP timepoints (104 versus 124 61 MEGs, 52.0% shared, and 87 versus 70 PEGs, 49.5% shared, respectively; Fig. 1c, 62 Supplementary Data Set 5). Hence, we identified barley imprinted genes active predominantly 63 during its endosperm cellularization and filling. 64 In terms of functions associated with imprinted genes (KEGG database; Fig. 1d, 65 Supplementary Data Set 6), MEGs were enriched in ubiquitin-mediated proteolysis via the Skp1 66 Cullin-F-box complex (Supplementary Fig. 4a). Both MEGs and PEGs were enriched in plant 67 hormone signal transduction, with MEGs involved in auxin and abscisic acid (ABA)-based 68 regulation, and PEGs in ABA and cytokinin signaling (Supplementary Fig. 4b). Cytokinin signaling 2 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764152; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.764152) (which 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 [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 69 associated PEGs included four authentic _HISTIDINE-CONTAINING PHOSPHOTRANSFER_ 70 _PROTEINs_ and a _B-type_ _TWO-COMPONENT RESPONSE REGULATOR_ transcription factor. 71 This points to parent-of-origin-specific effects in hormonal regulation and proteolysis during mid 72 to-late stages of barley endosperm development. 73 To assess the degree of evolutionary conservation, we identified orthogroups for imprinted 74 genes of barley, maize, rice and wheat [7–9,14,15] . A total of 42 orthogroups with 120 genes were 75 conserved in two species, and 2 orthogroups (8 genes) in three species (Fig. 2a, Supplementary 76 Data Set 7). Surprisingly, only 31 imprinted barley genes (12.2%) were orthologous to imprinted 77 genes in any other cereal and only 17 to hexaploid wheat ( _Triticum aestivum_ ). Comparison of 78 barley with the wheat progenitors, _Aegilops tauschii_ and _Triticum turgidum_, revealed 21 and 79 22 conserved imprinted genes, respectively (Fig. 2b, Supplementary Data Set 8). However, 14 80 and 15 imprinted genes shared between barley and _A. tauschii_ or _T. turgidum_, respectively, were 81 not conserved between barley and _T. aestivum_ . Among the 31 conserved imprinted barley genes, 82 14 genes, including DOF, MYB, and NAC transcription factors, were expressed in endosperm but 83 silent in embryo and maternal tissues (MTs), while the rest were widely expressed in embryo 84 and/or grain maternal tissues (Fig. 2c, Supplementary Data Set 9). These data suggest a low 85 degree of conservation of the imprinted genes in cereals, thus supporting data from dicots [16] . We 86 hypothesized that the parent-of-origin-specific effects might be better conserved at the pathway 87 level. In total, 121 biological processes (BPs), 73 molecular functions (MFs) and 35 cellular 88 components (CCs) were common for the imprinted genes in maize, rice, wheat, and barley 89 (Fig. 2d, Supplementary Data Set 10). Exclusion of general and narrow terms revealed a core 90 conserved set of 8 BPs, 4 MFs and 1 CC (Fig. 2e). The hormonal phosphorelay signal 91 transduction system was imprinted across cereals and significantly enriched in barley (FDR 92 adjusted p < 0.001, 6.85-fold enrichment), with 10 genes being specifically involved in the two 93 component system (TCS) pathway of cytokinin signaling (Supplementary Data Set 10). Activity of 94 the TCS genes was previously localized to barley endosperm transfer cells (ETCs), suggesting 95 an important role for imprinting in endosperm development and grain filling [17] . 96 To explore the temporal expression pattern of imprinted genes, we performed k-means 97 clustering of them using an endosperm transcriptomic data series from 4 to 32 DAP [18] and 98 identified three co-expression clusters (CL). The 56 CL1 genes (36 MEGs, 20 PEGs) and 97 CL2 99 genes (53 MEGs, 44 PEGs) peaked at 4 and 8 DAP, corresponding to the coenocytic and 100 cellularizing stages, respectively (Fig. 3a, Supplementary Data Set 5). This indicated that although 101 our samples represented cellularizing and cellularized endosperm, they contained also some 102 imprinted genes peaking in the coenocytic endosperm. The expression profile of 102 CL3 genes 3 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764152; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.764152) (which 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 [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 103 (61 MEGs, 41 PEGs) increased from 4 to 8 DAP and remained high until 32 DAP. These genes 104 may remain imprinted beyond 12 DAP (as reported for some genes in wheat) or may transition to 105 biased or biallelic expression (as found in maize) [7–9,14,15] . 106 The expression of imprinted genes in Arabidopsis coenocytic endosperm is concentrated 107 at the chalazal pole [19] . The absence of a chalaza in cereals let us also analyze the expression of 108 imprinted gene clusters in a spatially-resolved dataset of developing barley grains [20] (Fig. 3a). The 109 expression of CL1 genes localized mainly to the periphery of the coenocytic endosperm at 6 DAP 110 (Fig. 3b; Supplementary Fig. 5). The most expressed CL1 genes were 5 members of the _non-_ 111 _specific LIPID TRANSFER PROTEINS_ ( _nsLTPs_ ) involved in cuticle formation [21] . Interestingly, CL2 112 genes were highest in ETCs at 12 DAP and included ABA and cytokinin signaling and regulatory 113 genes (Fig. 3a and c, Supplementary Fig. 6). ETCs represent nutrient-absorbing tissue directly 114 adjacent to the nucellar projection, a tissue of maternal origin with a key role in transporting 115 nutrients to developing endosperm [22] . The analogous functions of the chalazal endosperm and 116 ETCs suggest convergent evolution of expression of imprinted genes at the interface between the 117 maternal and the filial bodies. Such patterning may reflect parent-of-origin-specific competition for 118 maternal resources and supports the parental conflict hypothesis [23] . 119 The expression of CL3 genes was scattered across the entire endosperm (Fig. 3a, 120 Supplementary Fig. 7). CL3 included the prolamin box binding transcription factor _HIGH LYSINE_ 121 _3_ ( _LYS3_ ; LOC123395703), known to positively regulate hundreds of genes involved in endosperm 122 filling and maturation [24] . Grains of its loss-of-function mutants have shrunken endosperm, reduced 123 starch and HORDEINs, and higher lysine content [25] . However, _lys3_ phenotypes segregate in a 3:1 124 recessive Mendelian manner [25,26], thus deviating from 1:1 or 2:1:1 ratio typical for imprinted 125 genes [27] . To explain this discrepancy, we analyzed heritability of _lys3_ phenotypes over grain 126 development using molecular and histological methods. Hierarchical clustering of genome-wide 127 endosperm transcriptomic samples from all F 1 hybrid combinations of wild-type (L) and _lys3.c_ (l; 128 E37*) plants revealed a 1:1 segregation at 12 DAP, separating the samples with maternally 129 inherited _LYS3_ wild-type (LL-L and LL-l) and mutant (ll-L and ll-l) endosperm constitutions (Fig. 4a, 130 Supplementary Data Set 11). Interestingly, the 20 DAP segregation pattern corresponded to 3:1, 131 with separated homozygous mutant samples (Fig. 4b). The maternal allele effects were obvious 132 also when comparing the number of genes downregulated in crosses with maternal wild-type (LL 133 l) or maternal mutant (Il-L and ll-l) endosperm constitution relative to homozygous wild-type (LL 134 L) at 12 DAP (Fig. 4c). By comparing genes downregulated in ll-L and ll-l relative to LL-L 135 endosperm over time, we identified 90 _LYS3_ targets downregulated in 12 DAP ll-L (Fig. 4d; 136 Supplementary Fig. 8a, b). These genes showed up to 5.6-fold enrichment for the GO terms 4 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764152; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.764152) (which 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 [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 137 _endopeptidase regulator activity_ and a striking 40-fold enrichment for _nutrient reservoir activity_ 138 (Fig. 4e, Supplementary Data Set 12). The _endopeptidase regulator activity_ included genes 139 coding for various enzyme inhibitors and SERPINs Z4/Z7, representing the major components of 140 endosperm albumin with storage and serine protease inhibitory functions [28,29] . The _nutrient_ 141 _reservoir activity_ genes coded for the major storage proteins HORDEINs and GLUTENINs [30,31], 142 showing 1:3 expression pattern at 8 DAP, 1:1 maternal allele-dependent pattern at 12 DAP, and 143 3:1 homozygous mutant-dependent pattern (particularly for _B1-hordein-like_ genes) at 20 DAP 144 (Fig. 4f, Supplementary Data Set 13). Therefore, _LYS3_ temporally regulates its target genes in a 145 non-Mendelian, maternal allele-dependent manner. 146 Since LYS3 regulates both storage protein and starch synthesis [32], we analyzed the genetic 147 basis of their accumulation during grain filling. HORDEINS were abundant in LL-L and LL-l and 148 significantly reduced in ll-L and ll-l endosperms at 16 DAP (two-way ANOVA with Tukey's HSD, 149 p < 0.05), resembling a 1:1 segregation (Fig. 5a-b, Supplementary Fig. 9). At 20 DAP, the ll-L 150 cross endosperm resembled the wild-type sample, suggesting a shift towards a 3:1 ratio. Starch 151 accumulated from 12 DAP with similar numbers of starch granules per endosperm, but with 152 different granule morphology (Fig. 5c-d, Supplementary Fig. 10, Supplementary Data Sets 14 and 153 15). Starch granules of LL-L and LL-l crosses were larger and more uniform in size and shape, 154 while the ll-L and ll-l combinations contained smaller and less uniform granules, corresponding to 155 maternal allele-dependent 1:1 segregation. The segregation pattern changed to 3:1 at 20 DAP, 156 with more uniform and lens-shaped starch granules in all combinations except ll-l. Interestingly, 157 the number of starch granules per endosperm correlated negatively with the number of _LYS3_ 158 alleles (R [2 ] = 0.9, p < 0.001) at 20 DAP, suggesting a quantitative effect of _LYS3_ expression 159 (Fig. 5d). Finally, we performed high-throughput quantitative analysis of the shrunken endosperm 160 phenotype by measuring and scoring dry grain thickness profiles of self-pollinated wild-type, 161 homozygous _lys3.c_, and self-pollinated heterozygous _LYS3_ / _lys3.c_ plants (Fig. 5e, Supplementary 162 Fig. 11a-b, Supplementary Data Sets 16 and 17). A Gaussian mixture model fitted to the observed 163 score distribution of the heterozygote-delivered F 2 seeds strongly preferred two components over 164 three (ΔBIC = 19.0), indicating no detectable intermediate phenotypic class and excluding semi 165 dominant 2:1:1 segregation (Supplementary Fig. 11c). Simulating 1000 synthetic F 2 datasets 166 under 3:1 segregation and comparing them to the observed distribution confirmed that the 167 observed F 2 score distribution matches the 3:1 expectation (Kolmogorov-Smirnov test, p = 0.165; 168 Supplementary Fig. 11d). 169 To explain the shift in expression patterns of LYS3 targets, we analyzed expression of the 170 parental _LYS3_ alleles. Based on RNA-seq of F 1 MOR-ETH and MOR-HS endosperms, only the 5 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764152; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.764152) (which 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 [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 171 maternal _LYS3_ was expressed at 8 DAP, but the paternal allele became partially active (up to 172 18%, compared to 33.3% expected reads) at 12 DAP (Supplementary Fig. 12, Supplementary 173 Data Sets 1-4). This was confirmed on an extended timescale (8, 12, 16 and 20 DAP) by reverse 174 transcription quantitative PCR (Fig. 5f), which revealed an increasing amount of paternal _LYS3_ 175 transcript over time, reaching 10-23% of the amplicons. Hence, the initial maternal expression of 176 _LYS3_ at 8 DAP was followed by partial transcriptional activation of the paternal allele at later 177 stages of endosperm development, consistent with a progressive compensation for the early 178 maternal allele deficiency and the subsequent transition toward a recessive Mendelian 179 segregation frequency. Taken together, the 12-16 DAP transcriptomes of _LYS3_ / _lys3_ crosses, their 180 HORDEIN phenotypes, and the dynamics of paternal _LYS3_ allele expression explain why _lys3_ 181 shows an early maternal 1:1 effect during grain filling but a recessive 3:1 phenotype in mature F2 182 grains. 183 Asking whether this changing expression pattern of paternal genes is common in barley, 184 we analyzed parent-of-origin specific expression of 102 genes from CL3 (Fig. 5g, Supplementary 185 Data Set 5). In total, 52.5% MEGs and 46.3% PEGs retained their character at both 8 and 12 DAP 186 (Fig. 5g). On the contrary, 13.1% MEGs and 29.3% of PEGs were parent-of-origin-expressed at 187 8 DAP but parent-of-origin biased or biallelically expressed at 12 DAP. The opposite case of 188 biparental expression at 8 DAP and parent-of-origin specific expression at 12 DAP was 189 represented by 34.4% MEGs and 24.4% PEGs. The latter case suggested that profiling at late 190 stages of endosperm development may reveal novel imprinted genes possibly affecting the final 191 stages of grain maturation. Indeed, 83 imprinted genes were detected in fully developed and 192 desiccating hexaploid bread wheat endosperm at 25 DAP [9] . 193 Our combined transcriptome and genetic data show that genomic imprinting, previously 194 well documented during early endosperm development across different species, has also an 195 important role during grain filling and maturation in barley. The transition from initial maternal 196 expression of transcription factor gene _LYS3_, regulating nutrient and energy supply in multiple 197 ways, to later activation of the paternal allele, suggests a tight maternal genome-exerted control 198 over the developing grain, in line with the parental conflict theory [23] . 199 200 **Online content** 201 Any methods, additional references, Nature Portfolio reporting summaries, source data, extended 202 data, supplementary information, acknowledgements, peer review information; details of author 203 contributions and competing interests; and statements of data and code availability. 204 6 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764152; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.764152) (which 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 [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 205 206 **References** 207 1. Cocucci, A. E. Morphogenetic seed types of Spermatophyta. _Plant Syst. Evol._ **250**, 1–6 208 (2005). 209 2. Lu, J. & Magnani, E. Seed tissue and nutrient partitioning, a case for the nucellus. _Plant_ 210 _Reprod._ **31**, 309–317 (2018). 211 3. Baroux, C. & Grossniklaus, U. Seeds—An evolutionary innovation underlying reproductive 212 success in flowering plants. in _Curr. Top. Dev. Biol._ vol. 131 605–642 (Academic Press, 213 2019). 214 4. Sato, H. & Köhler, C. Genomic imprinting regulates establishment and release of seed 215 dormancy. _Curr. Opin. Plant Biol._ **69**, 102264 (2022). 216 5. Köhler, C. & Weinhofer-Molisch, I. Mechanisms and evolution of genomic imprinting in 217 plants. _Heredity 2010 105:1_ **105**, 57–63 (2009). 218 6. Feil, R. & Berger, F. Convergent evolution of genomic imprinting in plants and mammals. 219 _Trends Genet._ **23**, 192–199 (2007). 220 7. Zhang, M. _et al._ Extensive, clustered parental imprinting of protein-coding and noncoding 221 RNAs in developing maize endosperm. _Proc. Natl. Acad. Sci. U.S.A._ **108**, 20042–20047 222 (2011). 223 8. Chen, C. _et al._ Characterization of imprinted genes in rice reveals conservation of regulation 224 and imprinting with other plant species. _Plant Physiol._ **177**, 1754–1771 (2018). 225 9. Yang, G. _et al._ Genomic imprinting was evolutionarily conserved during wheat 226 polyploidization. _Plant Cell_ **30**, 37–47 (2018). 227 10. Zhang, M. _et al._ Genome-wide screen of genes imprinted in sorghum endosperm, and the 228 roles of allelic differential cytosine methylation. _Plant J._ **85**, 424–436 (2016). 229 11. Dai, D. _et al._ Paternal imprinting of dosage-effect defective1 contributes to seed weight 230 xenia in maize. _Nat. Commun._ **13**, 1–11 (2022). 231 12. Li, Q. _et al._ The maize imprinted gene FLOURY3 encodes a PLATZ protein required for 232 tRNA and 5S rRNA transcription through interaction with RNA polymerase III. _Plant Cell_ **29**, 233 2661–2675 (2017). 234 13. Gutiérrez-Marcos, J. F. _et al._ maternally expressed gene1 Is a novel maize endosperm 235 transfer cell-specific gene with a maternal parent-of-origin pattern of expression. _Plant Cell_ 236 **16**, 1288–1301 (2004). 237 14. Luo, M. _et al._ A genome-wide survey of imprinted genes in rice seeds reveals imprinting 238 primarily occurs in the endosperm. _PLoS Genet._ **7**, (2011). 7 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764152; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.764152) (which 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 [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 239 15. Waters, A. J. _et al._ Parent-of-origin effects on gene expression and DNA methylation in the 240 maize endosperm. _Plant Cell_ **23**, 4221–4233 (2011). 241 16. Hatorangan, M. R., Laenen, B., Steige, K. A., Slotte, T. & Köhler, C. Rapid evolution of 242 genomic imprinting in two species of the Brassicaceae. _Plant Cell_ **28**, 1815–1827 (2016). 243 17. Hertig, C. W. _et al._ Genome-wide characterization of two-component system elements in 244 barley enables the identification of grain-specific phosphorelay genes. _BMC Plant Biol._ **25**, 245 209- (2025). 246 18. Kovacik, M. _et al._ The transcriptome landscape of developing barley seeds. _Plant Cell_ **36**, 247 2512–2530 (2024). 248 19. Picard, C. L., Povilus, R. A., Williams, B. P. & Gehring, M. Transcriptional and imprinting 249 complexity in Arabidopsis seeds at single-nucleus resolution. _Nat. Plants_ **7**, 730–738 250 (2021). 251 20. Peirats-Llobet, M. _et al._ A four-dimensional spatial transcriptome atlas of barley caryopsis 252 development and germination. _Plant Cell_ **38**, (2026). 253 21. Zhang, M. _et al._ Genome-wide analysis of the barley non-specific lipid transfer protein gene 254 family. _Crop J._ **7**, 65–76 (2019). 255 22. Sreenivasulu, N. _et al._ Barley grain development. in _Int. Rev. Cell Mol. Biol._ vol. 281 49–89 256 (2010). 257 23. Trivers, R. L. Parent-offspring conflict. _Am. Zool._ **14**, 249–264 (1974). 258 24. Mena, M., Vicente-Carbajosa, J., Schmidt, R. J. & Carbonero, P. An endosperm-specific 259 DOF protein from barley, highly conserved in wheat, binds to and activates transcription 260 from the prolamin-box of a native B-hordein promoter in barley endosperm. _Plant J._ **16**, 53– 261 62 (1998). 262 25. Doll, H. Inheritance of the high-lysine character of a barley mutant. _Hereditas 1973, 74_ **74**, 263 293–294 (1973). 264 26. Orman-Ligeza, B. _et al._ LYS3 encodes a prolamin-box-binding transcription factor that 265 controls embryo growth in barley and wheat. _J. Cereal Sci._ **93**, 102965 (2020). 266 27. Kermicle, J. L. Dependence of the R-mottled aleurone phenotype in maize on mode of 267 sexual transmission. _Genetics_ **66**, 69–85 (1970). 268 28. Dahl, S. W., Rasmussen, S. E. & Hejgaard, J. Heterologous expression of three plant 269 serpins with distinct inhibitory specificities. _J. Biol. Chem._ **271**, 25083–25088 (1996). 270 29. Rasmussen, S. K., Klausen, J., Hejgaard, J., Svensson, B. & Svendsen, I. Primary structure 271 of the plant serpin BSZ7 having the capacity of chymotrypsin inhibition. _Biochim. Biophys._ 272 _Acta_ **1297**, 127–130 (1996). 8 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764152; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.764152) (which 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 [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 273 30. Vinje, M. A. & Simmons, C. H. Characterization of barley (Horduem vulgare) lys3 mutants 274 identifies genes under the regulation of the prolamin-box binding transcription factor and 275 elucidates its role in endosperm promoter methylation during grain development. _Mol._ 276 _Genet. Genom._ **299**, (2024). 277 31. Vinje, M. A., Walling, J. G., Henson, C. A. & Duke, S. H. Comparative gene expression 278 analysis of the β-amylase and hordein gene families in the developing barley grain. _Gene_ 279 **693**, 127–136 (2019). 280 32. Ingversen, J., Køie, B. & Doll, H. Induced seed protein mutant of barley. _Experientia_ **29**, 281 1151–1152 (1973). 282 283 **Data availability** 284 All data supporting this study are included in the article and its Supplemental material. Sequence 285 data from this article can be found in the Gene Expression Omnibus under accession numbers 286 GSE349961 and GSE350207. 287 288 **Code availability** 289 The central code repository for the GATK output processing and post-processing analyses is 290 available via GitHub at https://github.com/pecinkagroup/barley-imprinting/. Post-processing 291 analyses were implemented with basic R (v.4.5.2) functionality using publicly available R 292 packages. 293 294 **Acknowledgments** 295 This work was supported by the Ministry of Education Youth and Sports of the Czech Republic 296 grants LUC24011 and “TowArds Next GENeration Crops (TANGENC)” (No. 297 CZ.02.01.01/00/22_008/0004581) and MSCA Doctoral Network HeatDDR (Funded by the 298 European Union under Grant Agreement No. 101169110). Computational resources were 299 supplied by the project "e-Infrastruktura CZ" (e-INFRA CZ LM2018140) supported by the Ministry 300 of Education, Youth and Sports of the Czech Republic. Barley FIND-IT screening and variant 301 isolation was performed with funding from the Carlsberg Foundation (grant CF15-0236). We thank 302 H. Tvardíková, K. Lakomčíková and M. Krautsova for assistance, Z. Bursová for plant care, P. 303 Navrátil for IT support, JIC Bioimaging facility and staff for their contribution to this publication and 304 O. Mittelsten Scheid for comments. 305 306 **Author contributions** 9 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764152; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.764152) (which 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 [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 307 A.P., M.K. and C.L.P conceived and designed the project. A.P. supervised the project, M.K., A.L., 308 A.N., P.Š., Z.K., K.K., K.T., B.S., T.D., M.Č. performed experiments, M.P.-L. and J.W. provided 309 spatial transcriptomics data; M.W.R, S.K., C.D. provided the barley FIND-IT variant; A.P and M.K. 310 wrote the manuscript with contribution of all authors. All authors have seen and agreed upon the 311 final version of the manuscript. 312 313 **Competing interests:** T.D. is the founder and owner of Labdeers Ltd., which produces the 314 Boxeed robot. The authors declare no competing interests. 315 316 **Additional information** 317 318 **Supplementary information** The online version contains supplementary material available at 319 320 **Correspondence and requests for materials** should be addressed to Ales Pecinka. 10 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764152; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.764152) (which 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 [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 321 **Figures and legends** 322 323 **Fig. 1.** Identification of 255 imprinted genes in barley endosperm. ( **a** ) Experimental workflow. ( **b** ) 324 Visualization of the candidate-imprinted and genotype-specific genes in F 1 MOR-ETH reciprocal 325 crosses 8 and 12 days after pollination (DAP). Dots: pink/blue – maternally/paternally expressed 11 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764152; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.764152) (which 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 [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 326 genes (MEGs/PEGs, respectively); yellow/dark green – genes with genotype (MOR/ETH)-specific 327 expression, respectively; grey – non-imprinted genes with biallelic expression. ( **c** ) Venn diagrams 328 of the MEGs (red) and PEGs (blue) 8 and 12 DAP. ( **d** ) Bubble plot of KEGG pathway enrichment 329 among MEGs and PEGs. The color scales indicate FDR-adjusted p-values, and the sizes of 330 bubbles indicate gene counts. 331 332 **Fig. 2.** Genomic imprinting operates via conserved biological processes in cereals. ( **a** ) Venn 333 diagram of imprinted orthogroups in cereals. Non-barley data are from [7–9,14,15] . ( **b** ) Venn diagram 334 comparing imprinted genes found in _Triticeae_ species. Non-barley data are from [9] . ( **c** ) Expression 335 based hierarchical clustering of barley maternally and paternally expressed genes (MEGs and 336 PEGs, respectively) found imprinted in more than one species. The total expression is shown at 337 different days after pollination (DAP) in barley endosperm (END), embryo (EMB) and maternal 338 tissues (MTs) [18] . Genes identified as imprinted in different cereals are indicated by the color circles 339 on the left. ( **d** ) Gene ontology-based biological processes represented by imprinted genes shown 340 in (a). ( **e** ) Numbers of imprinted genes per gene ontology-based biological process (BP), molecular 341 function (MF) and cellular component (CC) found in all four cereal species filtered for term size 342 and similarity. 12 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764152; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.764152) (which 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 [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 343 344 **Fig. 3.** Spatio-temporal dynamics of expression of barley imprinted genes. ( **a** ) K-means-based 345 temporal expression clusters of 255 barley imprinted genes. Number of genes per cluster is given 346 in the top right corner. Mean expression (colored line) and standard deviation (grey area) in 4 to 347 32 days after pollination (DAP) endosperm are based on [18] . Parental expression graphs give the 348 numbers of maternally/paternally expressed genes (MEGs/PEGs) per k-means cluster. Spatial 349 transcriptomics plots indicate average expression of imprinted genes in 6 and 12 DAP grain 350 sections. Circle sizes - percentage of expression; color - z-score expression; upper row – MEGs; 351 lower row - PEGs. The x-axis represents individual tissues as indicated in the representative 352 sections on the far-right based on [20] . ( **b-c** ) Spatial cumulative expression of 70/84 MEGs, and 353 69/71 PEGs at (b) 6 and (c) 12 DAP, respectively. Color scales represent the normalized counts 354 with the cut-offs individually specified on the image. Insets show detailed views of the spatio 355 temporal hotspots. 13 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764152; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.764152) (which 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 [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 356 357 **Fig. 4.** Non-Mendelian effects of _LYS3_ on endosperm transcriptome. ( **a-b** ) Hierarchically clustered 358 sample dissimilarity heatmaps of endosperm transcriptomes of the _LYS3_ (L) and/or _lys3.c_ (l) 359 crosses at (a) 12 and (b) 20 days after pollination (DAP). Maternal alleles are given first. R – 360 biological replicates. ( **c** ) Venn diagram of significantly downregulated (FDR-adjusted p < 0.05, 361 log2FoldChange < -1) genes in individual crosses at 12 DAP compared to LL-L endosperm. ( **d** ) 362 UpSet plot of significantly downregulated genes in ll-L and ll-l crosses at 12 and 20 DAP. The x 363 axis set size defines a total number of differentially expressed genes (DEGs) per cross. The y 364 axis shows the number of DEGs (intersection size) unique (single dots) or shared (connected 365 dots) among crosses. ( **e** ) Bubble plot of GO-term enrichment of common significantly 366 downregulated genes in ll-L and ll-l crosses relative to LL-L at 12 DAP. The x-axis shows fold 367 enrichment, the color scale indicates FDR-adjusted p-value, and size of the bubbles indicate gene 368 count. ( **f** ) Hierarchically clustered expression of GO term nutrient reservoir activity genes from (e) 369 in different cross combinations at 8, 12 and 20 DAP. Color scale indicates z-score. 14 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764152; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.764152) (which 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 [available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 370 371 **Fig. 5.** Maternal _LYS3_ allele controls accumulation of storage proteins and starch. ( **a** ) SDS-PAGE 372 of HORDEINs isolated from grains of _LYS3_ (L) and _lys3.c_ (l) crosses at given days after pollination 373 (DAP). Equivalent fractions of total extracted proteins were loaded. ( **b** ) Difference in band 374 intensities in individual crosses. Values marked with the same letter do not differ within the 375 respective group according to two-way ANOVA with Tukey's HSD (p < 0.05). ( **c** ) Accumulation of 376 starch granules (brown) based on Lugol staining of transverse grain sections of individual crosses 377 at given DAPs, and morphology of starch granules obtained by scanning electron microscopy at 378 12 and 20 DAP. ( **d** ) Quantification of starch granules per endosperm of individual _LYS3_ and _lys3.c_ 379 crosses at given DAPs. ( **e** ) Machine learning classification of thickness profiles of individual F 2 380 grains from self-pollinated _lys3.c_ heterozygotes. Blue and red line indicate peak positions of 381 homozygous _LYS3_ and _lys3.c_ genotypes from training dataset. Green and yellow dashed line 382 correspond to distribution of synthetic F 2 datasets under 3:1 and 1:1 segregation, respectively. ( **f** ) 383 Percentage of maternal (salmon) and paternal (blue) _LYS3_ transcripts based on RT-qPCR from 384 reciprocal F 1 MOR-ETH crosses at given DAPs. ( **g** ) Time specificity of imprinted genes from 385 expression clusters 1 to 3 in 8 and 12 DAP RNA-seq data. 15