henhaobio
Sign in

Genome biologyPlant BiologyPeer-reviewed journal article

Spatial and single-cell expression analyses reveal complex expression domains in early wheat spike development

Xiaosa XuiD, Huiqiong LiniD, Junli ZhangiD, Germán BurgueneriD, Francine ParaisoiD, Kun LiiD, Connor TumeltyiD, Chengxia LiiD, Yuchen Liu, Jorge DubcovskyiD

PDFDOIPreprintCodeData
In-depth full-text digest available · Jump to digest ↓

Abstract

Wheat is important for global food security and understanding the molecular mechanisms driving spike and spikelet development can inform the development of more productive varieties. In this study, we integrated single-molecule fluorescence in situ hybridization (smFISH) and single-cell RNA sequencing (scRNA-seq) to generate an atlas of cell clusters and expression domains during the early stages of wheat spike development. We characterized spatiotemporal expression of 99 genes by smFISH in 48,225 cells at the early transition, late double ridge, and floret primordia stages. These cells were grouped into 21 different expression domains, including four in the basal region of the developing spikelets and three different meristematic regions, which were consistent across spikelets and sections. Using induced mutants, we revealed functional roles associated with the specific expression patterns of LFY in intercalary meristems, SPL14 in inflorescence meristems, and FZP in glume axillae. Complementary scRNA-seq profiling of 26,009 cells from W2.5 and W3.5 stages identified 23 distinct cell clusters. We used the scRNA-seq information to impute the expression of 74,464 genes into the spatially anchored smFISH-labelled cells and generated a public website to visualize them. We used experimental and imputed expression profiles, together with co-expression studies and correlation matrices, to annotate the scRNA-seq clusters. From co-expression analyses, we uncovered genes associated with boundary genes TCP24 and FZP, as well as the meristematic genes AGL6 and ULT1. The smFISH and scRNA-seq studies provided complementary tools for dissecting gene networks that regulate spike development and identifying new co-expressed genes for functional characterization.

In-depth summary

Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 2 (10 Aug 2025) · CC BY 4.0

Combining 99-gene smFISH across 48,225 cells with scRNA-seq of 26,009 cells maps 21 spatial expression domains and imputes 74,464 genes in early tetraploid wheat spikes.

Breeding more productive wheat spikes needs mechanistic insight, yet the spatial diversity of expression domains in the early spike and the gene networks behind them were poorly characterized despite many genetically defined regulators. Pairing spatially anchored smFISH with scRNA-seq from the same genotype and stages lets imputation extend spatial maps genome-wide, sharpen single-cell annotation, and nominate co-expressed genes for functional study.

Key results

  1. ShownA 99-gene smFISH atlas resolved 21 expression domains across three early spike stages (W1.5, W2.5, W3.5), including four at the base of developing spikelets and three meristematic regions. Cell number and cluster diversity rose as the spike developed.

    48,225 QuPath-segmented cells from two sections per stage; per stage, 7,624 at W1.5, 16,400 at W2.5 and 24,210 at W3.5, forming 14, 19 and 21 clusters. Spikelet-proximal clusters recurred across spikelets and sections, and patterns agreed with earlier in situ hybridization and published MERFISH data. · Abstract; 1. Spatial expression and cell segmentation

  2. ShownscRNA-seq resolved 23 spike cell clusters. Nearest-neighbor imputation transferred 74,464 genes onto spatially anchored smFISH cells, recovering expected epidermal patterns and homeolog-specific profiles that helped annotate the clusters.

    26,009 cells (12,304 at W2.5, 13,705 at W3.5; three replicates per stage); nine of 25 initial clusters were cell-cycle and removed, leaving 17,283 cells in 23 clusters. Imputation from the five nearest scRNA-seq cells using 91 shared genes yielded 74,464 genes, correctly placing six epidermal markers absent from the smFISH panel and separating ULT-A1 from ULT-B1 domains. · 2. Single-cell RNA sequencing (scRNA-seq) of the…; 4.3.4. Imputation of scRNA-seq gene expression…; 2.3. Epidermal clusters; 2.10. Meristem clusters

  3. ShownExpression domains guided mutant analyses. LFY marks a putative stem intercalary meristem (c10) and influences stem width and solidity. FZP defines the glume axilla domain (c14) and is needed to suppress glume axillary meristems.

    lfy-null plants had nodes 22.3% ± 1.8% and internodes 11.5% ± 2.0% wider, and a 62% ± 17% smaller hollow-center diameter than WT, with no significant change in stem length. Eight independent CRISPR T0 lines mutated in both FZP homeologs replaced spikelets with nested glume arrays that repeated until the meristems were exhausted. · 1.1. The basal region below the spike; 1.4.5. Cluster c14 “Glume axilla”; 4.3.9. Generation of loss-of-function mutants of…

  4. ShownSPL14 expression falls sharply in the apical meristem region as the inflorescence meristem begins converting to a terminal spikelet. CRISPR loss of both SPL14 homeologs caused a premature transition and fewer spikelets per spike.

    In the terminal meristem plus the two youngest lateral meristems, SPL14 hybridization density fell 6.7-fold from W2.5 to W3.5 (P= 0.0031), one of 11 additional genes that changed (P< 0.05). Plants of one transgene-free T4 line with both homeologs edited had 42% lower SNS (spikelets per spike), were 12 cm taller and had 2.5 more leaves. · 1.6. Changes within the inflorescence meristem…; 4.3.9. Generation of loss-of-function mutants of…

  5. SuggestedCo-expression with the boundary genes FZP and TCP24 and the meristem genes ULT1 and AGL6 recovered orthologs of known rice and maize boundary regulators, outlined a transcription factor network, and nominated uncharacterized candidates for functional study.

    608 genes co-expressed with FZP and 606 with TCP24 (226 shared, 988 total; threshold mean R plus four SD), including CUC3 and MFS1. GENIE3's top-500 links formed a 95-gene network made up entirely of transcription factors. 82 of the 100 top ULT-A1 correlates encode ribosomal proteins; DDM1a, BTBN16 and SHI1 were prioritized. · 2.12.1. Genes co-expressed with boundary genes…; 2.12.2. Genes co-expressed with meristem genes…; 4.3.7. scRNA-seq co-expression analysis; 4.3.8. Network analysis

Methods, data and reproducibility

Tetraploid Kronos spikes at W1.5, W2.5 and W3.5 were profiled by 99-gene smFISH (Molecular Cartography; QuPath segmentation, Seurat clustering) and, at W2.5 and W3.5, by 10x protoplast scRNA-seq; scRNA-seq expression was imputed into smFISH cells from cosine-weighted five nearest neighbors, alongside CRISPR mutants, co-expression and GENIE3 networks.

  • Longitudinal sections of developing Kronos spikes at W1.5, W2.5 and W3.5 profiled by 99-gene smFISH (two sections per stage) · tetraploid wheat (Kronos) · n = 48,225 cells in six sections · 10.5281/zenodo.14872904
  • 10x Genomics scRNA-seq of spike protoplasts at W2.5 and W3.5 (three replicates per stage) · tetraploid wheat (Kronos) · n = 26,009 cells; 17,283 after removing cell-cycle clusters · PRJNA1199502
  • Bulk RNA-seq of basal, middle and top thirds of W3.0 spikes (NovaSeq X Plus PE150) used to localize scRNA-seq clusters · tetraploid wheat (Kronos) · n = three replicates per section; 20 to 25 apices each · PRJNA1217025
  • Published hexaploid wheat spatial transcriptomics normalized data, provided by C. Uauy, used for cluster correlations · hexaploid wheat · n = 200 genes; three early stages used
  • Kronos genome assembly v1.1 (annotation v1.0) used as the read-mapping reference · tetraploid wheat (Kronos) · 10.5281/zenodo.11106422
  • Maize genes previously flagged as responsive to protoplasting, mapped to Kronos homologs · maize · n = 713 genes
  • Kronos CRISPR-Cas9 mutants in both FZP or both SPL14 homeologs, plus an lfy-null mutant from a previous study (genetic background not stated in the text) · wheat (Kronos CRISPR lines; lfy-null background not stated) · n = eight FZP T0 lines; one spl14 T4 line
Methods & tools
Molecular Cartography 100-plex combinatorial smFISH (Resolve BioSciences), QuPath, Resolve BioSciences Recognize (Seurat v4 backend), Seurat v5, STARsolo, Monocle 3, GENIE3, Cytoscape 3.10.3, DAVID (KEGG pathway analysis), 10x Genomics Chromium Next GEM Single Cell 3' Kit v3.1, CRISPR-Cas9 GRF4-GIF1 binary vector (Addgene Plasmid #160393), ImageJ with Polylux plugin
Reagents & environment
Seurat v4 (Recognize backend) and Seurat v5; Monocle 3; Cytoscape 3.10.3; STARsolo mapping to Kronos genome v1.1 (annotation v1.0, gene 3' ends extended by 500 bp); 10x Next GEM Single Cell 3' Kit v3.1.
Reproducibility gaps
Probe design used Resolve BioSciences' proprietary algorithm, and the vendor ran imaging and decoding. The normalized hexaploid wheat spatial data used for the correlations were provided directly by C. Uauy's lab.
Limitations (6)
  • An smFISH panel can hold only a limited number of genes (99 here, with no epidermal markers), so genome-wide spatial coverage depends on imputation from scRNA-seq rather than direct measurement. · stated by the authors
  • Probes were designed from the higher-expressing homeolog but are not genome-specific and can detect both homeologs, so smFISH signals cannot separate A-genome from B-genome expression. · stated by the authors
  • No marker genes cleanly separated the inflorescence meristem from early spikelet meristems (both in cluster c8). Changes in the inflorescence meristem were instead assessed in manually delineated regions using within-gene density comparisons. · stated by the authors
  • Protoplasting stress likely affected scRNA-seq clusters sc3 and sc22, which were left unannotated. About 19% and 28% of their preferentially expressed genes were undetected in RNA-seq of intact spikes. · stated by the authors
  • Imputation quality is supported only by visual agreement for selected markers. No held-out-gene test or other quantitative accuracy metric is reported for the five-nearest-neighbor transfer across the 74,464 imputed genes. · noted in this summary
  • The spl14 phenotypes (SNS, height, leaf number) come from a single transgene-free edited line, and the atlas uses two sections per stage. Line-specific or section-specific effects therefore cannot be fully excluded. · noted in this summary

Read this if you study grass inflorescence or spikelet development, or want to impute scRNA-seq expression into targeted smFISH cells in a polyploid crop.

Citations
15
FWCI
8.06
Field percentile
top 2%

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISLong '25Qu '25VanGessel '22Li '26Liu '25Paraiso '23
75% semantic matchPlant Biology
Katie Long, Ashleigh Lister, Maximillian R. W. Jones+16 more·The Plant Cell·Dec 12, 2025· 32 citations