Biology subjects

Romay, C.

Publications and source records attributed to Romay, C..

At least 19 recordsLinked to original sources

bioRxiv2026

An open field phenomics resource for multimodal maize yield prediction across divergent environments

Temporal drone phenotyping captures crop development, but irregular flight schedules complicate comparisons across environments. We release curated imagery from 356 flights across 19 Genomes to Fields environments containing 1,180 maize (Zea mays L.) hybrids. To evaluate its utility, we integrated functional principal components of vegetation index and weather trajectories with genomic information. Combinined genomic and phenomic kernels improved yield prediction, reaching correlations up to r = 0.501 for held-out hybrids in environments represented in training and 0.408 when environments were also withheld. Accumulated growing degree days offered no consistent predictive advantage over days after planting, and weather contributed modest, task-dependent gains. A transformer neural process learned directly from irregular observations, serving as a novel application of neural process models in agriculture. Mapping vegetation index functional principal components identified recurrent quantitative trait loci on chromosomes 3 and 7. This resource and its reproducible analyses guide the use of temporal spectral data for crop prediction and genetic discovery.

plant biology↗

bioRxiv2025

The origins and adaptive consequences of polyploidy in a dominant prairie grass

Polyploidy is ubiquitous across North American prairies, which provide essential ecosystem services and rich soil for agriculture. Yet the mechanism driving polyploid abundance is unclear. Multiple hypotheses have been proposed including polyploid abundance is proportional to the opportunity for whole genome duplication (WGD), and WGD alters phenotypes that may increase fitness. We tested these two hypotheses together in the mixed-ploidy species Andropogon gerardi, a dominant grass species in endangered North American tallgrass prairies. Leveraging a novel, phased allopolyploid reference genome, we found the A. gerardi hexaploid arose after the C4 grassland expansion in the early Pleistocene, when glacial cycles likely increased secondary contact between the diploid progenitors. We sequenced A. gerardi from 25 popula-tions and examined cytotype performance and morphology in a controlled environment to investigate the consequences of the contemporary mixed-ploidy populations. We found the 9x A. gerardi cytotype is a neopolyploid and a result of recurrent WGD events. Further, we demonstrate the 9x neopolyploids have greater growth and a decreased stomatal pore index, which is adaptive in xeric climates where the 9x cy-totype is most common. Together, our results support both hypotheses for polyploid abundance in North America: WGD is a product of opportunity and can have immediate fitness consequences. Although the changes to fitness may provide an advantage to 9x A. gerardi, the establishment of 9x may lower overall population fitness due to the lower reproductive viability of 9x individuals. Significance statementPolyploid species are abundant in North American prairies and make up many of the dominant species in the ecosystem. This prominence could be a result of whole genome duplication conferring an advantage that increases the frequency of polyploids or could simply indicate that the opportunity for whole genome duplication is higher in this ecosystem, or both. Through examining three polyploidization events in A. gerardi, the dominant species in endangered tallgrass North American prairies, we found whole genome duplication is both surprisingly common and confers traits that are beneficial in some environments.

plant biology↗

bioRxiv2025

Cross-species optimization of nuclei isolation in plants

Single-cell technologies are transforming plant biology, yet broadly transferable nuclei isolation remains a key bottleneck for snRNA-seq. We developed a reproducible, Percoll-based workflow that is applicable to multiple maize tissues and eight additional plant species. In maize, nuclei from root, stem, leaf, and embryo consistently concentrated at the 80% Percoll interface and exhibited high integrity, with typical recoveries > 50,000 nuclei per sample. For other species, gradient compositions were tuned according to genome size to achieve efficient enrichment and clean suspensions, and yields ranged from 20,000 to 50,000 nuclei per sample. Downstream validation showed that nuclei from special interest maize and Tripsacum generated high-quality 10x Genomics snRNA-seq libraries, as supported by cDNA quality profiles. These results demonstrate the versatility and robustness of the method across species and tissues.

plant biology↗

bioRxiv2025

PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms

Understanding how DNA sequence encodes biological function remains a fundamental challenge in biology. Flowering plants (angiosperms), the dominant terrestrial clade, exhibit maximal biochemical complexity, extraordinary species diversity (over 100,000 species), relatively recent origins ([~]160 million years), [~]200-fold variation in genome size and relative compact coding regions compared with other eukaryotes. These features present both a unique challenge and opportunity for pre-training DNA language models to understand plant-specific evolutionary conservation, regulatory architectures and genomic functions. Here, we introduce PlantCAD2, an extended context, plant-specific DNA language model with single-nucleotide resolution, pre-trained on 65 angiosperm genomes, together with a series of public benchmarks for evaluation. Comprehensive zero-shot testing shows that PlantCAD2 (676 million parameters) efficiently captures evolutionary conservation, surpassing the 7-billion-parameter Evo2 model in 10 of 12 tasks. With parameter-efficient fine-tuning, PlantCAD2 also outperforms the 1-billion-parameter AgroNT across seven cross-species tasks including chromatin accessible region, gene expression and protein translation. Moreover, its 8,192bp context window substantially improves accessible chromatin prediction in large genomes such as maize (AUPRC increasing from 0.587 to 0.711), underscoring the importance of long-range context for modeling distal regulation. Together, these results establish PlantCAD2 as a powerful, efficient, and versatile foundation model for plant genomics, enabling accurate genome annotation across diverse species.

bioinformatics↗

bioRxiv2025

Phylogenetic Analysis and Machine Learning Identify Signatures of Selection and Predict Deleterious Mutations in Common Bean

Mutations are continuous source of new alleles and genetic diversity in populations. Domestication and selection influence the accumulation of alleles occurring across a range of deleteriousness. Evidence suggests that mildly deleterious mutations (DelMut) can be purged out of breeding populations, increasing favorable allele accumulation. We used phylogeny-based analyses among 36 legume genomes to identify selection signatures and predict DelMut in common bean. We also developed a multiparent advanced generation intercrossed (MAGIC) population of black beans to characterize DelMut. Genes involved in nitrogen metabolism showed signs of positive selection in the Middle American genome, whereas genes related to phosphorylation were positively selected in the Andean genome. By combining conservation and protein information with machine learning (ML) for high-dimensional feature analysis, we characterized 82,442 sites in the MAGIC founders (36,558 polymorphic) and 4,753 sites evenly sequenced among RILs that could be potentially deleterious. Variation in the number of highly DelMut (high predicted deleterious scores) among lines was observed and later correlated with agronomic traits. Phenotypic analyses showed that calculated genetic load (and number of highly DelMut) was negatively correlated with flowering time, maturity, and yield. A detailed in-silico analysis of predicted mutations showed presence in highly conserved protein regions, which is likely to affect protein functionality. Our results show that variation in genetic load can be observed in breeding populations and potentially impact plant performance. These results contribute to understanding the genome-wide accumulation patterns of DelMut in breeding populations. Our study supports future development of strategies to reduce genetic load in promising germplasm and accelerate breeding programs. Key messages- Genome-wide highly deleterious mutations were predicted in conserved protein domains potentially affecting protein functionality. - Variation in the genetic load and number of highly deleterious mutations can be observed in artificial breeding populations. - Nitrogen- and phosphorous-related genes are under positive selection in Middle American and Andean beans respectively.

plant biology↗

bioRxiv2025

Extensive modulation of a conserved cis-regulatory code across 589 grass species

The growing availability of genomes from non-model organisms offers new opportunities to identify functional loci underlying trait variation through comparative genomics. While cis-regulatory regions drive much of phenotypic evolution, linking them to specific functions remains challenging. We identified 514 cis-regulatory motifs enriched in regulatory regions of five diverse grass species, with 73% consistently enriched across all, suggesting a deeply conserved regulatory code. We then quantified conservation of specific motif instances across 589 grass species, revealing widespread gain and loss over evolutionary time. Conservation declined rapidly over the first few million years of divergence, yet [~]50% of motif instances were conserved back to the origin of grasses [~]100 million years ago. Conservation patterns varied by gene class, with modestly higher conservation at transcription factor genes. To test for adaptive cis-regulatory changes, we used phylogenetic mixed models to identify motif gains and losses associated with ecological niche transitions. Our models revealed polygenic adaptation across 810 motif-orthogroup combinations, including convergent gains of HSF/GARP motifs at an Alpha-N-acetylglucosaminidase gene associated with adaptation to temperate environments. Our results support a "stable code, variable sites" model in which cis-regulatory evolution involves extensive turnover of individual binding site instances while largely preserving transcription factors binding preferences. Cis-regulatory changes at hundreds to thousands of genes appear to contribute to environmental adaptation. Our results highlight the potential of comparative genomics and phylogenetic mixed models to reveal the genetic basis of complex traits.

genomics↗

bioRxiv2025

Higher order repeat structures reflect diverging evolutionary paths in maize centromeres and knobs

BackgroundHighly repetitive tandem repeat arrays, known as satellite DNAs, are enriched in low-recombination regions such as centromeres. Satellite arrays often contain complex internal structures called higher-order repeats (HORs), which may have functional significance. Maize is unusual in that its satellites occur in two distinct genomic contexts: centromeres, which interact with kinetochore proteins, and knobs, which undergo meiotic drive in the presence of Abnormal chromosome 10. Whether maize centromeres or knobs contain HOR patterns, and how such patterns relate to function, remains unclear. ResultsHere, we generate 13 repeat-sensitive genome assemblies of maize and its recent ancestor, teosinte. We develop a new graph-based pipeline, HiReNET, to classify HORs and demonstrate its utility in both Arabidopsis and maize. We find that HORs are ubiquitous in maize satellites but are typically low-frequency with small patterns, unlike the large, continuous HOR blocks characteristic of human centromeres. Approximately 38% of centromeric CentC monomers occur in HORs; however, no specific HOR class dominates any functional centromere as marked by centromeric histone H3. Arabidopsis centromeres have a similar HOR landscape. In contrast, maize knobs exhibit a more structured HOR distribution. Large knobs contain megabase-scale similarity blocks with repeated HOR patterns. These repeat units likely promote unequal crossing over, enabling rapid knob expansion, and may harbor motifs recognized by trans-acting factors involved in meiotic drive. ConclusionsHORs occur in all major maize satellite arrays. Specific HORs are not associated with centromere function, but knobs contain conserved HOR patterns within similarity blocks that may facilitate meiotic drive.

genomics↗

bioRxiv2024

Cross-species plant genomes modeling at single nucleotide resolution using a pre-trained DNA language model

Interpreting function and fitness effects in diverse plant genomes requires transferable models. Language models (LMs) pre-trained on large-scale biological sequences can learn evolutionary conservation and offer cross-species prediction better than supervised models through fine-tuning limited labeled data. We introduce PlantCaduceus, a plant DNA LM based on the Caduceus and Mamba architectures, pre-trained on a curated dataset of 16 Angiosperm genomes. Fine-tuning PlantCaduceus on limited labeled Arabidopsis data for four tasks, including predicting translation initiation/termination sites and splice donor and acceptor sites, demonstrated high transferability to 160 million year diverged maize, outperforming the best existing DNA LM by 1.45 to 7.23-fold. PlantCaduceus is competitive to state-of-the-art protein LMs in terms of deleterious mutation identification, and is threefold better than PhyloP. Additionally, PlantCaduceus successfully identifies well-known causal variants in both Arabidopsis and maize. Overall, PlantCaduceus is a versatile DNA LM that can accelerate plant genomics and crop breeding applications.

bioinformatics↗

bioRxiv2024

Muller's Ratchet in Action: The Erosion of Sexual Reproduction Genes in Domesticated Cassava (Manihot esculenta)

Centuries of clonal propagation in cassava (Manihot esculenta) have engaged Mullers Ratchet, leading to the accumulation of deleterious mutations due to the absence of sexual recombination. This has resulted in both inbreeding depression affecting yield and a significant decrease in reproductive performance, creating hurdles for contemporary breeding programs. Cassava is a member of the Euphorbiaceae family, including notable species such as rubber tree (Hevea brasiliensis) and poinsettia (Euphorbia pulcherrima). Expanding upon preliminary draft genomes, we annotated 7 long-read genome assemblies and aligned a total of 52 genomes, to analyze selection across the genome and the phylogeny. Through this comparative genomic approach, we identified 48 genes under relaxed selection in cassava. Notably, we discovered an overrepresentation of floral expressed genes, especially focused at six pollen-related genes. Our results indicate that domestication and a transition to clonal propagation has reduced selection pressures on sexually reproductive functions in cassava leading to an accumulation of mutations in pollen-related genes. This relaxed selection and the genome-wide deleterious mutations responsible for inbreeding depression are potential targets for improving cassava breeding, where the generation of new varieties relies on recombining favorable alleles through sexual reproduction.

evolutionary biology↗

bioRxiv2023

Fishing for a reelGene: evaluating gene models with evolution and machine learning

Assembled genomes and their associated annotations have transformed our study of gene function. However, each new assembly generates new gene models. Inconsistencies between annotations likely arise from biological and technical causes, including pseudogene misclassification, transposon activity, and intron retention from sequencing of unspliced transcripts. To evaluate gene model predictions, we developed reelGene, a pipeline of machine learning models focused on (1) transcription boundaries, (2) mRNA integrity, and (3) protein structure. The first two models leverage sequence characteristics and evolutionary conservation across related taxa to learn the grammar of conserved transcription boundaries and mRNA sequences, while the third uses conserved evolutionary grammar of protein sequences to predict whether a gene can produce a protein. Evaluating 1.8 million gene models in maize, reelGene found that 28% were incorrectly annotated or nonfunctional. By leveraging a large cohort of related species and through learning the conserved grammar of proteins, reelGene provides a tool for both evaluating gene model accuracy and genome biology.

genomics↗

bioRxiv2022

Cross-species predictive modeling reveals conserved drought responses between maize and sorghum

Drought tolerance is a highly complex trait controlled by numerous interconnected pathways with substantial variation within and across plant species. This complexity makes it difficult to distill individual genetic loci underlying tolerance, and to identify core or conserved drought responsive pathways. Here, we collected drought physiology and gene expression datasets across diverse genotypes of the C4 cereals sorghum and maize and searched for signatures defining water deficit responses. Differential gene expression identified few overlapping drought associated genes across sorghum genotypes, but using a predictive modeling approach, we found a shared core drought response across development, genotype, and stress severity. Our model had similar robustness when applied to datasets in maize, reflecting a conserved drought response between sorghum and maize. The top predictors are enriched in functions associated with various abiotic stress responsive pathways as well as core cellular functions. These conserved drought response genes were less likely to contain deleterious mutations than other gene sets, suggesting that core drought responsive genes are under evolutionary and functional constraints. Our findings support a broad evolutionary conservation of drought responses in C4 grasses regardless of innate stress tolerance, which could have important implications for developing climate resilient cereals. Significance StatementDrought is a complex and variable stress that is difficult to quantify and link to underlying mechanisms both within and across species. Here, we developed a predictive model to classify drought stress responses in sorghum and identify important features that are responsive to water deficit. Our model has high predictive accuracy across development, genotype, and stress severity, and the top features are enriched in genes related to classical stress responses and have functional and evolutionary conservation. We applied this sorghum trained model to maize, and observed similar predictive accuracy of drought responses, supporting transfer learning across plant species. Our findings suggest there are deeply conserved drought responses across C4 grasses that are unrelated to tolerance.

plant biology↗

bioRxiv2022

Melon pan-genome and multi-parental framework for high-resolution trait dissection

Linking between genotype and phenotype is a fundamental goal in biology and requires robust data for both layers. The prominent increase in plant genome sequencing and comparisons of multiple related individuals, exposed the abundance of structural genomic variation and suggest that a single reference genome cannot represent the complete sequence diversity of a crop species, leading to the expansion of the pan-genome concept. For high-resolution forward genetics, this unprecedented access to genomic variation should be paralleled by availability and phenotypic characterization of genetic diversity, and effective integration between these layers. Here, we describe a multi-parental framework for trait dissection in melon, leveraging a novel pan-genome constructed for this crop. Melon (Cucumis melo L.) is an important crop from the Cucurbitaceae family, which display extensive phenotypic variation available for breeding. A diverse core set of 25 founder lines (MelonCore25) was sequenced using a combination of short and long-read technologies and their genomes were assembled de novo. The construction of a melon pan-genome exposed substantial variation in genome size and structure, including detection of ~300,000 structural variants and ~9 million SNPs. A half-diallel derived set of 300 F2 populations representing all possible MelonCore25 parental combinations was constructed as framework for trait dissection through integration with the pan-genome. We demonstrate the potential of this unified framework for genetic analysis of various melon traits, including rind color and mottling pattern, fruit sugar content and resistance to fungal diseases. We anticipate that utilization of this integrated resource will enhance genetic dissection of important traits and accelerate melon breeding. Significance statementPan-genomes aim to address the abundance of genome structural variation within species for improved genomic analyses. New pan-genome, constructed from de novo genome assemblies of 25 diverse melon (Cucumis melo L.) accessions is integrated with a half-diallel derived set of 300 F2 populations representing all possible parental combinations. The potential of this unified multi-parental trait dissection framework for melon genetics and breeding is presented.

genetics↗

bioRxiv2022

Elucidating the patterns of pleiotropy and its biological relevance in maize

Pleiotropy - when a single gene controls two or more seemingly unrelated traits - has been shown to impact genes with effects on flowering time, leaf architecture, and inflorescence morphology in maize. However, the genome-wide impact of true biological pleiotropy across all maize phenotypes is largely unknown. Here we investigate the extent to which biological pleiotropy impacts phenotypes within maize through GWAS summary statistics reanalyzed from previously published metabolite, field, and expression phenotypes across the Nested Association Mapping population and Goodman Association Panel. Through phenotypic saturation of 120,597 traits, we obtain over 480 million significant quantitative trait nucleotides. We estimate that only 1.56-32.3% of intervals show some degree of pleiotropy. We then assessed the relationship between pleiotropy and various biological features such as gene expression, chromatin accessibility, sequence conservation, and enrichment for gene ontology terms. We find very little relationship between pleiotropy and these variables when compared to permuted pleiotropy. We hypothesize that biological pleiotropy of common alleles is not widespread in maize and is highly impacted by nuisance terms such as population structure and linkage disequilibrium. Natural selection on large standing natural variation in maize populations may target wide- and large-effect variants, leaving the prevalence of detectable pleiotropy relatively low. Author SummaryThe genetic basis of complex traits has been thought to exhibit pleiotropy, which is the notion that a single locus can control two or more unrelated traits. Widespread reports in the human disease literature show genomic signatures of pleiotropic loci across many traits. However, little is known about the prevalence and behavior of pleiotropy in maize across a large number of phenotypes. Using association mapping of common alleles in over one hundred thousand traits, we determine how pleiotropic each region was and use these quantitative scores to functionally characterize each region of the genome. Our results show little evidence that pleiotropy is a common phenomenon in maize. We observed that maize does not exhibit the same pleiotropic characteristics as human diseases in terms of prevalence, gene expression, chromatin accessibility, or sequence conservation. Rather than pervasive pleiotropy, we hypothesize that strong selection on large and wide effect loci and the need for trait independence at the gene level keep the prevalence of pleiotropy low, thus, allowing for the adaptation of maize varieties to novel environments and conditions.

plant biology↗

bioRxiv2022

A happy accident: a novel turfgrass reference genome

Poa pratensis, commonly known as Kentucky bluegrass, is a popular cool-season grass species used as turf in lawns and recreation areas globally. Despite its substantial economic value, a reference genome had not previously been assembled due to the genomes relatively large size and biological complexity that includes apomixis, polyploidy, and interspecific hybridization. We report here a fortuitous de novo assembly and annotation of a P. pratensis genome. Instead of sequencing the genome of a C4 grass, we accidentally sampled and sequenced tissue from a weedy P. pratensis whose stolon was intertwined with that of the C4 grass. The draft assembly consists of 6.09 Gbp with an N50 scaffold length of 65.1 Mbp, and a total of 118 scaffolds, generated using PacBio long reads and Bionano optical map technology. We annotated 256K gene models and found 58% of the genome to be composed of transposable elements. To demonstrate the applicability of the reference genome, we evaluated population structure and estimated genetic diversity in P. pratensis collected from three North American prairies, two in Manitoba, Canada and one in Colorado, USA. Our results support previous studies that found high genetic diversity and population structure within the species. The reference genome and annotation will be an important resource for turfgrass breeding and study of bluegrasses.

plant biology↗

bioRxiv2021

A multiple genome alignment workflow shows the impact of repeat masking and parameter tuning on alignment of functional regions in plants

Alignments of multiple genomes are a cornerstone of comparative genomics, but generating these alignments remains technically challenging and often impractical. We developed the msa_pipeline workflow (https://bitbucket.org/bucklerlab/msa_pipeline) based on the LAST aligner to allow practical and sensitive multiple alignment of diverged plant genomes with minimal user inputs. Our workflow only requires a set of genomes in FASTA format as input. The workflow outputs multiple alignments in MAF format, and includes utilities to help calculate genome-wide conservation scores. As high repeat content and genomic divergence are substantial challenges in plant genome alignment, we also explored the impact of different masking approaches and alignment parameters using genome assemblies of 33 grass species. Compared to conventional masking with RepeatMasker, a k-mer masking approach increased the alignment rate of CDS and non-coding functional regions by 25% and 14% respectively. We further found that default alignment parameters generally perform well, but parameter tuning can increase the alignment rate for non-coding functional regions by over 52% compared to default LAST settings. Finally, by increasing alignment sensitivity from the default baseline, parameter tuning can increase the number of non-coding sites that can be scored for conservation by over 76%.

genomics↗

bioRxiv2021

Genome-wide Imputation Using the Practical Haplotype Graph in the Heterozygous Crop Cassava

Genomic applications such as genomic selection and genome-wide association have become increasingly common since the advent of genome sequencing. Genotype imputation makes it possible to infer whole genome information from limited input data, making large sampling for genomic applications more feasible, especially in non-model species where resources are less abundant. Imputation becomes increasingly difficult in heterozygous species where haplotypes must be phased. The Practical Haplotype Graph is a recently developed tool that can accurately impute genotypes, using a reference panel of haplotypes. The Practical Haplotype Graph is a haplotype database that implements a trellis graph to predict haplotypes using minimal input data. Genotyping information is aligned to the database and missing haplotypes are predicted from the most likely path through the graph. We showcase the ability of the Practical Haplotype Graph to impute genomic information in the highly heterozygous crop cassava (Manihot esculenta). Accurately phased haplotypes were sampled from runs of homozygosity across a diverse panel of individuals to populate the graph, which proved more accurate than relying on computational phasing methods. At 1X input sequence coverage, the Practical Haplotype Graph achieves a high concordance between predicted and true genotypes (R=0.84), as compared to the standard imputation tool Beagle (R=0.69). This improved accuracy was especially visible in the prediction of rare and heterozygous alleles. We validate the Practical Haplotype Graph as an accurate imputation tool in the heterozygous crop cassava, showing its potential for application in heterozygous species.

genomics↗

bioRxiv2017

Genetic Analysis of Lodging in Diverse Maize Hybrids

Damage caused by lodging is a significant problem in corn production that results in estimated annual yield losses of 5-20%. Over the past 100 years, substantial maize breeding efforts have increased lodging resistance by artificial selection. However, less research has focused on understanding the genetic architecture underlying lodging. Lodging is a problematic trait to evaluate since it is greatly influenced by environmental factors such as wind, rain, and insect infestation, which make replication difficult. In this study over 1,723 diverse inbred maize genotypes were crossed to a common tester and evaluated in five environments over multiple years. Natural lodging due to severe weather conditions occurred in all five environments. By testing a large population of genetically diverse maize lines in multiple field environments, we detected significant correlations for this highly environmentally influenced trait across environments and with important agronomic traits such as yield and plant height. This study also permitted the mapping of quantitative trait loci (QTL) for lodging. Several QTL identified in this study overlapped with loci previously mapped for stalk strength in related maize inbred lines. QTL intervals mapped in this study also overlapped candidate genes implicated in the regulation of lignin and cellulose synthesis.

genetics↗

bioRxiv2016

Identifying the diamond in the rough: a study of allelic diversity underlying flowering time adaptation in maize landraces

Landraces (traditional varieties) of crop species are a reservoir of useful genetic diversity, yet remain untapped due to the genetic linkage between the few useful alleles with hundreds of undesirable alleles1. We integrated two approaches to characterize the genetic diversity of over 3000 maize landraces from across the Americas. First, we mapped the genomic regions controlling latitudinal and altitudinal adaptation, identifying 1498 genes. Second, we developed and used F-One Association Mapping (FOAM) to directly map genes controlling flowering time across 22 environments, identifying 1,005 genes. In total 65% of the SNPs associated with altitude were also associated with flowering time. In particular, we observed many of the significant SNPs were contained in large structural variants (inversions, centromeres, and pericentromeric regions): 29.4% for flowering time, 58.4% for altitude and 13.1% for latitude. The combined mapping results indicate that while floral regulatory network genes contribute substantially to field variation, over 90% of contributing genes likely have indirect effects. Our strategy can be used to harness the diversity of maize and other plant and animal species.

genetics↗