Genomic mechanisms of climate adaptation in polyploid bioenergy switchgrass.

Long-term climate change and periodic environmental extremes threaten food and fuel security1 and global crop productivity2-4. Although molecular and adaptive breeding strategies can buffer the effects of climatic stress and improve crop resilience5, these approaches require sufficient knowledge of the genes that underlie productivity and adaptation6-knowledge that has been limited to a small number of well-studied model systems. Here we present the assembly and annotation of the large and complex genome of the polyploid bioenergy crop switchgrass (Panicum virgatum). Analysis of biomass and survival among 732 resequenced genotypes, which were grown across 10 common gardens that span 1,800 km of latitude, jointly revealed extensive genomic evidence of climate adaptation. Climate-gene-biomass associations were abundant but varied considerably among deeply diverged gene pools. Furthermore, we found that gene flow accelerated climate adaptation during the postglacial colonization of northern habitats through introgression of alleles from a pre-adapted northern gene pool. The polyploid nature of switchgrass also enhanced adaptive potential through the fractionation of gene function, as there was an increased level of heritable genetic diversity on the nondominant subgenome. In addition to investigating patterns of climate adaptation, the genome resources and gene-trait associations developed here provide breeders with the necessary tools to increase switchgrass yield for the sustainable production of bioenergy.

Assembly, comparative analysis, and utilization of a single haplotype reference genome for soybean.

Cultivar Williams 82 has served as the reference genome for the soybean research community since 2008, but is known to have areas of genomic heterogeneity among different sub-lines. This work provides an updated assembly (version Wm82.a6) derived from a specific sub-line known as Wm82-ISU-01 (seeds available under USDA accession PI 704477). The genome was assembled using Pacific BioSciences HiFi reads and integrated into chromosomes using HiC. The 20 soybean chromosomes assembled into a genome of 1.01Gb, consisting of 36 contigs. The genome annotation identified 48 387 gene models, named in accordance with previous assembly versions Wm82.a2 and Wm82.a4. Comparisons of Wm82.a6 with other near-gapless assemblies of Williams 82 reveal large regions of genomic heterogeneity, including regions of differential introgression from the cultivar Kingwa within approximately 30 Mb and 25 Mb segments on chromosomes 03 and 07, respectively. Additionally, our analysis revealed a previously unknown large (>20 Mb) heterogeneous region in the pericentromeric region of chromosome 12, where Wm82.a6 matches the 'Williams' haplotype while the other two near-gapless assemblies do not match the haplotype of either parent of Williams 82. In addition to the Wm82.a6 assembly, we also assembled the genome of 'Fiskeby III,' a rich resource for abiotic stress resistance genes. A genome comparison of Wm82.a6 with Fiskeby III revealed the nucleotide and structural polymorphisms between the two genomes within a QTL region for iron deficiency chlorosis resistance. The Wm82.a6 and Fiskeby III genomes described here will enhance comparative and functional genomics capacities and applications in the soybean community.

Host genotype controls ecological change in the leaf fungal microbiome.

Leaf fungal microbiomes can be fundamental drivers of host plant success, as they contain pathogens that devastate crop plants and taxa that enhance nutrient uptake, discourage herbivory, and antagonize pathogens. We measured leaf fungal diversity with amplicon sequencing across an entire growing season in a diversity panel of switchgrass (Panicum virgatum). We also sampled a replicated subset of genotypes across 3 additional sites to compare the importance of time, space, ecology, and genetics. We found a strong successional pattern in the microbiome shaped both by host genetics and environmental factors. Further, we used genome-wide association (GWA) mapping and RNA sequencing to show that 3 cysteine-rich receptor-like kinases (crRLKs) were linked to a genetic locus associated with microbiome structure. We confirmed GWAS results in an independent set of genotypes for both the internal transcribed spacer (ITS) and large subunit (LSU) ribosomal DNA markers. Fungal pathogens were central to microbial covariance networks, and genotypes susceptible to pathogens differed in their expression of the 3 crRLKs, suggesting that host immune genes are a principal means of controlling the entire leaf microbiome.

JGI Plant Gene Atlas: an updateable transcriptome resource to improve functional gene descriptions across the plant kingdom.

Gene functional descriptions offer a crucial line of evidence for candidate genes underlying trait variation. Conversely, plant responses to environmental cues represent important resources to decipher gene function and subsequently provide molecular targets for plant improvement through gene editing. However, biological roles of large proportions of genes across the plant phylogeny are poorly annotated. Here we describe the Joint Genome Institute (JGI) Plant Gene Atlas, an updateable data resource consisting of transcript abundance assays spanning 18 diverse species. To integrate across these diverse genotypes, we analyzed expression profiles, built gene clusters that exhibited tissue/condition specific expression, and tested for transcriptional response to environmental queues. We discovered extensive phylogenetically constrained and condition-specific expression profiles for genes without any previously documented functional annotation. Such conserved expression patterns and tightly co-expressed gene clusters let us assign expression derived additional biological information to 64 495 genes with otherwise unknown functions. The ever-expanding Gene Atlas resource is available at JGI Plant Gene Atlas (https://plantgeneatlas.jgi.doe.gov) and Phytozome (https://phytozome.jgi.doe.gov/), providing bulk access to data and user-specified queries of gene sets. Combined, these web interfaces let users access differentially expressed genes, track orthologs across the Gene Atlas plants, graphically represent co-expressed genes, and visualize gene ontology and pathway enrichments.

Predicted structural proteome of Sphagnum divinum and proteome-scale annotation.

MOTIVATION: Sphagnum-dominated peatlands store a substantial amount of terrestrial carbon. The genus is undersampled and under-studied. No experimental crystal structure from any Sphagnum species exists in the Protein Data Bank and fewer than 200 Sphagnum-related genes have structural models available in the AlphaFold Protein Structure Database. Tools and resources are needed to help bridge these gaps, and to enable the analysis of other structural proteomes now made possible by accurate structure prediction. RESULTS: We present the predicted structural proteome (25 134 primary transcripts) of Sphagnum divinum computed using AlphaFold, structural alignment results of all high-confidence models against an annotated nonredundant crystallographic database of over 90,000 structures, a structure-based classification of putative Enzyme Commission (EC) numbers across this proteome, and the computational method to perform this proteome-scale structure-based annotation. AVAILABILITY AND IMPLEMENTATION: All data and code are available in public repositories, detailed at https://github.com/BSDExabio/SAFA. The structural models of the S. divinum proteome have been deposited in the ModelArchive repository at https://modelarchive.org/doi/10.5452/ma-ornl-sphdiv.

Transcriptome and DNA methylome divergence of inflorescence development between 2 ecotypes in Panicum hallii.

The morphological diversity of the inflorescence determines flower and seed production, which is critical for plant adaptation. Hall's panicgrass (Panicum hallii, P. hallii) is a wild perennial grass that has been developed as a model to study perennial grass biology and adaptive evolution. Highly divergent inflorescences have evolved between the 2 major ecotypes in P. hallii, the upland ecotype (P. hallii var hallii, HAL2 genotype) with compact inflorescence and large seed and the lowland ecotype (P. hallii var filipes, FIL2 genotype) with an open inflorescence and small seed. Here we conducted a comparative analysis of the transcriptome and DNA methylome, an epigenetic mark that influences gene expression regulation, across different stages of inflorescence development using genomic references for each ecotype. Global transcriptome analysis of differentially expressed genes (DEGs) and co-expression modules underlying the inflorescence divergence revealed the potential role of cytokinin signaling in heterochronic changes. Comparing DNA methylome profiles revealed a remarkable level of differential DNA methylation associated with the evolution of P. hallii inflorescence. We found that a large proportion of differentially methylated regions (DMRs) were located in the flanking regulatory regions of genes. Intriguingly, we observed a substantial bias of CHH hypermethylation in the promoters of FIL2 genes. The integration of DEGs, DMRs, and Ka/Ks ratio results characterized the evolutionary features of DMR-associated DEGs that contribute to the divergence of the P. hallii inflorescence. This study provides insights into the transcriptome and epigenetic landscape of inflorescence divergence in P. hallii and a genomic resource for perennial grass biology.

PEATmoss (Physcomitrella Expression Atlas Tool): a unified gene expression atlas for the model plant Physcomitrella patens.

Physcomitrella patens is a bryophyte model plant that is often used to study plant evolution and development. Its resources are of great importance for comparative genomics and evo-devo approaches. However, expression data from Physcomitrella patens were so far generated using different gene annotation versions and three different platforms: CombiMatrix and NimbleGen expression microarrays and RNA sequencing. The currently available P. patens expression data are distributed across three tools with different visualization methods to access the data. Here, we introduce an interactive expression atlas, Physcomitrella Expression Atlas Tool (PEATmoss), that unifies publicly available expression data for P. patens and provides multiple visualization methods to query the data in a single web-based tool. Moreover, PEATmoss includes 35 expression experiments not previously available in any other expression atlas. To facilitate gene expression queries across different gene annotation versions, and to access P. patens annotations and related resources, a lookup database and web tool linked to PEATmoss was implemented. PEATmoss can be accessed at https://peatmoss.online.uni-marburg.de.

Soybean (Glycine max) Haplotype Map (GmHapMap): a universal resource for soybean translational and functional genomics.

Here, we describe a worldwide haplotype map for soybean (GmHapMap) constructed using whole-genome sequence data for 1007 Glycine max accessions and yielding 14.9 million variants as well as 4.3 M tag single-nucleotide polymorphisms (SNPs). When sampling random subsets of these accessions, the number of variants and tag SNPs plateaued beyond approximately 800 and 600 accessions, respectively. This suggests extensive coverage of diversity within the cultivated soybean. GmHapMap variants were imputed onto 21 618 previously genotyped accessions with up to 96% success for common alleles. A local association analysis was performed with the imputed data using markers located in a 1-Mb region known to contribute to seed oil content and enabled us to identify a candidate causal SNP residing in the NPC1 gene. We determined gene-centric haplotypes (407 867 GCHs) for the 55 589 genes and showed that such haplotypes can help to identify alleles that differ in the resulting phenotype. Finally, we predicted 18 031 putative loss-of-function (LOF) mutations in 10 662 genes and illustrated how such a resource can be used to explore gene function. The GmHapMap provides a unique worldwide resource for applied soybean genomics and breeding.

PlantSEED enables automated annotation and reconstruction of plant primary metabolism with improved compartmentalization and comparative consistency.

Genome-scale metabolic reconstructions help us to understand and engineer metabolism. Next-generation sequencing technologies are delivering genomes and transcriptomes for an ever-widening range of plants. While such omic data can, in principle, be used to compare metabolic reconstructions in different species, organs and environmental conditions, these comparisons require a standardized framework for the reconstruction of metabolic networks from transcript data. We previously introduced PlantSEED as a framework covering primary metabolism for 10 species. We have now expanded PlantSEED to include 39 species and provide tools that enable automated annotation and metabolic reconstruction from transcriptome data. The algorithm for automated annotation in PlantSEED propagates annotations using a set of signature k-mers (short amino acid sequences characteristic of particular proteins) that identify metabolic enzymes with an accuracy of about 97%. PlantSEED reconstructions are built from a curated template that includes consistent compartmentalization for more than 100 primary metabolic subsystems. Together, the annotation and reconstruction algorithms produce reconstructions without gaps and with more accurate compartmentalization than existing resources. These tools are available via the PlantSEED web interface at http://modelseed.org, which enables users to upload, annotate and reconstruct from private transcript data and simulate metabolic activity under various conditions using flux balance analysis. We demonstrate the ability to compare these metabolic reconstructions with a case study involving growth on several nitrogen sources in roots of four species.

The Sorghum bicolor reference genome: improved assembly, gene annotations, a transcriptome atlas, and signatures of genome organization.

Sorghum bicolor is a drought tolerant C4 grass used for the production of grain, forage, sugar, and lignocellulosic biomass and a genetic model for C4 grasses due to its relatively small genome (approximately 800 Mbp), diploid genetics, diverse germplasm, and colinearity with other C4 grass genomes. In this study, deep sequencing, genetic linkage analysis, and transcriptome data were used to produce and annotate a high-quality reference genome sequence. Reference genome sequence order was improved, 29.6 Mbp of additional sequence was incorporated, the number of genes annotated increased 24% to 34 211, average gene length and N50 increased, and error frequency was reduced 10-fold to 1 per 100 kbp. Subtelomeric repeats with characteristics of Tandem Repeats in Miniature (TRIM) elements were identified at the termini of most chromosomes. Nucleosome occupancy predictions identified nucleosomes positioned immediately downstream of transcription start sites and at different densities across chromosomes. Alignment of more than 50 resequenced genomes from diverse sorghum genotypes to the reference genome identified approximately 7.4 M single nucleotide polymorphisms (SNPs) and 1.9 M indels. Large-scale variant features in euchromatin were identified with periodicities of approximately 25 kbp. A transcriptome atlas of gene expression was constructed from 47 RNA-seq profiles of growing and developed tissues of the major plant organs (roots, leaves, stems, panicles, and seed) collected during the juvenile, vegetative and reproductive phases. Analysis of the transcriptome data indicated that tissue type and protein kinase expression had large influences on transcriptional profile clustering. The updated assembly, annotation, and transcriptome data represent a resource for C4 grass research and crop improvement.

The Physcomitrella patens gene atlas project: large-scale RNA-seq based expression data.

High-throughput RNA sequencing (RNA-seq) has recently become the method of choice to define and analyze transcriptomes. For the model moss Physcomitrella patens, although this method has been used to help analyze specific perturbations, no overall reference dataset has yet been established. In the framework of the Gene Atlas project, the Joint Genome Institute selected P. patens as a flagship genome, opening the way to generate the first comprehensive transcriptome dataset for this moss. The first round of sequencing described here is composed of 99 independent libraries spanning 34 different developmental stages and conditions. Upon dataset quality control and processing through read mapping, 28 509 of the 34 361 v3.3 gene models (83%) were detected to be expressed across the samples. Differentially expressed genes (DEGs) were calculated across the dataset to permit perturbation comparisons between conditions. The analysis of the three most distinct and abundant P. patens growth stages - protonema, gametophore and sporophyte - allowed us to define both general transcriptional patterns and stage-specific transcripts. As an example of variation of physico-chemical growth conditions, we detail here the impact of ammonium supplementation under standard growth conditions on the protonemal transcriptome. Finally, the cooperative nature of this project allowed us to analyze inter-laboratory variation, as 13 different laboratories around the world provided samples. We compare differences in the replication of experiments in a single laboratory and between different laboratories.

Drought responsive gene expression regulatory divergence between upland and lowland ecotypes of a perennial C4 grass.

Climatic adaptation is an example of a genotype-by-environment interaction (G×E) of fitness. Selection upon gene expression regulatory variation can contribute to adaptive phenotypic diversity; however, surprisingly few studies have examined how genome-wide patterns of gene expression G×E are manifested in response to environmental stress and other selective agents that cause climatic adaptation. Here, we characterize drought-responsive expression divergence between upland (drought-adapted) and lowland (mesic) ecotypes of the perennial C4 grass,Panicum hallii, in natural field conditions. Overall, we find that cis-regulatory elements contributed to gene expression divergence across 47% of genes, 7.2% of which exhibit drought-responsive G×E. While less well-represented, we observe 1294 genes (7.8%) with transeffects.Trans-by-environment interactions are weaker and much less common than cis G×E, occurring in only 0.7% oft rans-regulated genes. Finally, gene expression heterosis is highly enriched in expression phenotypes with significant G×E. As such, modes of inheritance that drive heterosis, such as dominance or overdominance, may be common among G×E genes. Interestingly, motifs specific to drought-responsive transcription factors are highly enriched in the promoters of genes exhibiting G×E and transregulation, indicating that expression G×E and heterosis may result from the evolution of transcription factors or their binding sites.P. hallii serves as the genomic model for its close relative and emerging biofuel crop, switchgrass (Panicum virgatum). Accordingly, the results here not only aid in the discovery of the genetic mechanisms that underlie local adaptation but also provide a foundation to improve switchgrass yield under water-limited conditions.

Genome-wide analysis of lectin receptor-like kinases in Populus.

BACKGROUND: Receptor-like kinases (RLKs) belong to a large protein family with over 600 members in Arabidopsis and over 1000 in rice. Among RLKs, the lectin receptor-like kinases (LecRLKs) possess a characteristic extracellular carbohydrate-binding lectin domain and play important roles in plant development and innate immunity. There are 75 and 173 LecRLKs in Arabidopsis and rice, respectively. However, little is known about LecRLKs in perennial woody plants. RESULTS: Here we report the genome-wide analysis of classification, domain architecture and expression of LecRLKs in the perennial woody model plant Populus. We found that the LecRLK family has expanded in Populus to a total of 231, including 180 G-type, 50 L-type and 1 C-type LecRLKs. Expansion of the Populus LecRLKs (PtLecRLKs) occurred partially through tandem duplication. Based on domain architecture and orientation features, we classified PtLecRLKs into eight different classes. RNA-seq-based transcriptomics analysis revealed diverse expression patterns of PtLecRLK genes among leaves, stems, roots, buds and reproductive tissues and organs. CONCLUSIONS: This study offers a comprehensive view of LecRLKs in the perennial woody model plant Populus and provides a foundation for functional characterization of this important family of receptor-like kinases.

Relics of interspecific hybridization retained in the genome of a drought-adapted peanut cultivar.

Peanut (Arachis hypogaea L.) is a globally important oil and food crop frequently grown in arid, semi-arid, or dryland environments. Improving drought tolerance is a key goal for peanut crop improvement efforts. Here we present the genome assembly and gene model annotation for 'Line8', a peanut genotype bred from drought tolerant cultivars. Our assembly and annotation are the most contiguous and complete peanut genome resources currently available. The high contiguity of the Line8 assembly allowed us to explore structural variation both between peanut genotypes and subgenomes. We detect several large inversions between Line8 and other peanut genome assemblies, and there is a trend for the inversions between more genetically diverged genotypes to have higher gene content. We also relate patterns of subgenome exchange to structural variation between Line8 homeologous chromosomes. Unexpectedly, we discover that Line8 harbors an introgression from A.cardenasii, a diploid peanut relative and important donor of disease resistance alleles to peanut breeding populations. The fully resolved sequences of both haplotypes in this introgression provide the first in situ characterization of A.cardenasii candidate alleles that can be leveraged for future targeted improvement efforts. The completeness of our genome will support peanut biotechnology and broader research into the evolution of hybridization and polyploidy.

Genome resources for three modern cotton lines guide future breeding efforts.

Cotton (Gossypium hirsutum L.) is the key renewable fibre crop worldwide, yet its yield and fibre quality show high variability due to genotype-specific traits and complex interactions among cultivars, management practices and environmental factors. Modern breeding practices may limit future yield gains due to a narrow founding gene pool. Precision breeding and biotechnological approaches offer potential solutions, contingent on accurate cultivar-specific data. Here we address this need by generating high-quality reference genomes for three modern cotton cultivars ('UGA230', 'UA48' and 'CSX8308') and updating the 'TM-1' cotton genetic standard reference. Despite hypothesized genetic uniformity, considerable sequence and structural variation was observed among the four genomes, which overlap with ancient and ongoing genomic introgressions from 'Pima' cotton, gene regulatory mechanisms and phenotypic trait divergence. Differentially expressed genes across fibre development correlate with fibre production, potentially contributing to the distinctive fibre quality traits observed in modern cotton cultivars. These genomes and comparative analyses provide a valuable foundation for future genetic endeavours to enhance global cotton yield and sustainability.

Phylogenetic and functional analysis of tiller angle control homeologs in allotetraploid cotton.

Introduction: Plants can adapt their growth to optimize light capture in competitive environments, with branch angle being a crucial factor influencing plant phenotype and physiology. Decreased branch angles in cereal crops have been shown to enhance productivity in high-density plantings. The Tiller Angle Control (TAC1) gene, known for regulating tiller inclination in rice and corn, has been found to control branch angle in eudicots. Manipulating TAC1 in field crops like cotton offers the potential for improving crop productivity. Methods: Using a homolog-based methodology, we examined the distribution of TAC1-related genes in cotton compared to other angiosperms. Furthermore, tissue-specific qPCR analysis unveiled distinct expression patterns of TAC1 genes in various cotton tissues. To silence highly expressed specific TAC1 homeologs in the stem, we applied CRISPR-Cas9 gene editing and Agrobacterium-mediated transformation, followed by genotyping and subsequent phenotypic validation of the mutants. Results: Gene duplication events of TAC1 specific to the Gossypium lineage were identified, with 3 copies in diploid progenitors and 6 copies in allotetraploid cottons. Sequence analysis of the TAC1 homeologs in Gossypium hirsutum revealed divergence from other angiosperms with 1-2 copies, suggesting possible neo- or sub-functionalization for the duplicated copies. These TAC1 homeologs exhibited distinct gene expression patterns in various tissues over developmental time, with elevated expression of A11G109300 and D11G112200, specifically in flowers and stems, respectively. CRISPR-mediated loss of these TAC1 homeologous genes resulted in a reduction in branch angle and altered petiole angles, and a 5 to 10-fold reduction in TAC1 expression in the mutants, confirming their role in controlling branch and petiole angles. This research provides a promising strategy for genetically engineering branch and petiole angles in commercial cotton varieties, potentially leading to increased productivity.

Comparative transcriptome analyses reveal insights into catkin bloom patterns in pecan protogynous and protandrous cultivars.

In perennial plants such as pecan, once reproductive maturity is attained, there are genetic switches that are regulated and required for flower development year after year. Pecan trees are heterodichogamous with both pistillate and staminate flowers produced on the same tree. Therefore, defining genes exclusively responsible for pistillate inflorescence and staminate inflorescence (catkin) initiation is challenging at best. To understand these genetic switches and their timing, this study analyzed catkin bloom and gene expression of lateral buds collected from a protogynous (Wichita) and a protandrous (Western) pecan cultivar in summer, autumn and spring. Our data showed that pistillate flowers in the current season on the same shoot negatively impacted catkin production on the protogynous 'Wichita' cultivar. Whereas fruit production the previous year on 'Wichita' had a positive effect on catkin production on the same shoot the following year. However, fruiting the previous year nor current year pistillate flower production had no significant effect on catkin production on 'Western' (protandrous cultivar) cultivar. The RNA-Seq results present more significant differences between the fruiting and non-fruiting shoots of the 'Wichita' cultivar compared to the 'Western' cultivar, revealing the genetic signals likely responsible for catkin production. Our data presented here, indicates the genes showing expression for the initiation of both types of flowers the season before bloom.

The reference genome and abiotic stress responses of the model perennial grass Brachypodium sylvaticum.

Perennial grasses are important forage crops and emerging biomass crops and have the potential to be more sustainable grain crops. However, most perennial grass crops are difficult experimental subjects due to their large size, difficult genetics, and/or their recalcitrance to transformation. Thus, a tractable model perennial grass could be used to rapidly make discoveries that can be translated to perennial grass crops. Brachypodium sylvaticum has the potential to serve as such a model because of its small size, rapid generation time, simple genetics, and transformability. Here, we provide a high-quality genome assembly and annotation for B. sylvaticum, an essential resource for a modern model system. In addition, we conducted transcriptomic studies under 4 abiotic stresses (water, heat, salt, and freezing). Our results indicate that crowns are more responsive to freezing than leaves which may help them overwinter. We observed extensive transcriptional responses with varying temporal dynamics to all abiotic stresses, including classic heat-responsive genes. These results can be used to form testable hypotheses about how perennial grasses respond to these stresses. Taken together, these results will allow B. sylvaticum to serve as a truly tractable perennial model system.

Newly identified sex chromosomes in the Sphagnum (peat moss) genome alter carbon sequestration and ecosystem dynamics.

Peatlands are crucial sinks for atmospheric carbon but are critically threatened due to warming climates. Sphagnum (peat moss) species are keystone members of peatland communities where they actively engineer hyperacidic conditions, which improves their competitive advantage and accelerates ecosystem-level carbon sequestration. To dissect the molecular and physiological sources of this unique biology, we generated chromosome-scale genomes of two Sphagnum species: S. divinum and S. angustifolium. Sphagnum genomes show no gene colinearity with any other reference genome to date, demonstrating that Sphagnum represents an unsampled lineage of land plant evolution. The genomes also revealed an average recombination rate an order of magnitude higher than vascular land plants and short putative U/V sex chromosomes. These newly described sex chromosomes interact with autosomal loci that significantly impact growth across diverse pH conditions. This discovery demonstrates that the ability of Sphagnum to sequester carbon in acidic peat bogs is mediated by interactions between sex, autosomes and environment.

Genome analyses reveal population structure and a purple stigma color gene candidate in finger millet.

Finger millet is a key food security crop widely grown in eastern Africa, India and Nepal. Long considered a 'poor man's crop', finger millet has regained attention over the past decade for its climate resilience and the nutritional qualities of its grain. To bring finger millet breeding into the 21st century, here we present the assembly and annotation of a chromosome-scale reference genome. We show that this ~1.3 million years old allotetraploid has a high level of homoeologous gene retention and lacks subgenome dominance. Population structure is mainly driven by the differential presence of large wild segments in the pericentromeric regions of several chromosomes. Trait mapping, followed by variant analysis of gene candidates, reveals that loss of purple coloration of anthers and stigma is associated with loss-of-function mutations in the finger millet orthologs of the maize R1/B1 and Arabidopsis GL3/EGL3 anthocyanin regulatory genes. Proanthocyanidin production in seed is not affected by these gene knockouts.