Targeted development of new plant varieties

Abstract

The present disclosure provides improved breeding methods that allow for the selection of plants having improved agronomic traits using clustering to identify genotypes similar to the genotypes of previously successful commercial plants.

Description

TARGETED DEVELOPMENT OF NEW PLANT VARIETIESBACKGROUND

[0001] In a target population of environments (TPE) there can arise outstanding hybrids that are not amenable to rapid improvement. These products may require intense and extensive breeding and testing to generate improved lines. Traditional breeding processes may take up to seven years or more to develop improved hybrids, and the complete agronomics (e.g., susceptibility to local diseases or to abiotic failures) of such hybrids may remain unknown until testing is conducted over wide and varied environments. Accordingly, there exists a need to efficiently improve and replace highly successful plant lines though a fast-track approach which uses background information (performance of tested genetics in a specific TPE) and available genetics.SUMMARY

[0002] Provided are methods for generating a hybrid plant having one more improved agronomic characteristics as compared to an elite hybrid plant by genotyping a first inbred parent and a second inbred parent of an elite hybrid to produce a first parent genotypic profile and a second parent genotypic profile, screening a plurality of inbred lines to produce a plurality of genotypic profiles for four or more inbred lines of the plurality of inbred lines, selecting from the plurality of genotypic profiles a first subset of inbred lines having genetic similarity to the first parent genomic profile and a second subset of inbred lines having genetic similarity to the second parent genomic profile, generating an in silica hybrid population comprising a plurality of simulated hybrids each simulated hybrid of the population having a parent from the first subset of inbred lines and a parent from the second subset of inbred lines, generating a predicted genotypic profile for one or more plants of the hybrid population, comparing the predicted genotypic profile for the one or more plants of the hybrid population to the genotypic profile of the elite hybrid, identifying hybrid plants of the hybrid population that have genotypic similarity to the elite hybrid and are predicted to have the one or more improved agronomic characteristics, and growing the identified hybrid plants of the hybrid population and selecting hybrid plants that have one or more improved agronomic traits as compared to the elite hybrid plant. In certain embodiments, members of the first subset of inbred plant lines, members of the second subset ofinbred plant lines, or both are selected using cluster analysis, such as, for example, hierarchical clustering.

[0003] Also provide are methods of selecting a hybrid plant having one more improved agronomic characteristics as compared to an elite hybrid plant by genotyping a first inbred parent and a second inbred parent of an elite hybrid to produce a first parent genotypic profile and a second parent genotypic profile, generating a genotypic profile for members of a population, the population comprising a plurality of inbred lines, clustering the members of the population of inbred lines to generate clusters of inbred lines having genetic similarity, selecting a first cluster of inbred lines having genetic similarity to the first parent genomic profile and a second cluster of inbred lines having genetic similarity to the second parent genomic profile, generating an in silico hybrid population comprising a plurality of simulated hybrids each simulated hybrid of the population having a parent from the first cluster of inbred lines and a parent from the second cluster of inbred lines, generating a predicted phenotypic profile for at least one trait for one or more plants of the in silico hybrid population, identifying hybrid plants of the in silico hybrid predicted to have improved phenotypic profile for the at least one trait as compared to the phenotypic profile of the elite hybrid, and growing the identified hybrid plants of the hybrid in silico population and selecting hybrid plants that have one or more improved agronomic traits as compared to the elite hybrid plant.BRIEF DESCRIPTION OF THE DRAWINGS

[0004] The disclosure can be more fully understood from the following detailed description and the accompanying drawings, which form a part of this application.

[0005] Fig. 1 is a schematic illustrating the different steps of the targeted development of new hybrids approach. First, a target hybrid or variety (i.e., a genetic element showing a successful performance across several years and locations) is detected. Then, a hierarchical clustering analysis is performed including all available germplasm and those clusters including the parents of the target hybrid or variety are detected. The next step is the in-silico creation of all possible diallei crosses between the genetic elements more similar to the target hybrid / variety parents. The final step involves the prediction and selection of best combinations that outperform the target hybrid / variety that will be subject to field testing after hybrid / variety creation.

[0006] Fig. 2 is a block diagram illustrating an exemplary computer system including a server and a computing device according to an embodiment as disclosed herein.DETAILED DESCRIPTION

[0007] The present disclosure provides improved breeding methods that allow for the selection and / or production of plants having one or more improved agronomic traits using clustering to identify genotypes similar to the genotypes of previously successful commercial plants, such that, for example, the methods described herein generate new hybrid plants or populations of hybrid plants that have an improved characteristic as compared to the previously successful commercial hybrid plant.

[0008] In certain embodiments, the method comprises genotyping a first inbred parent and a second inbred parent of an elite hybrid to produce a first parent genotypic profile and a second parent genotypic profile, screening a plurality of inbred lines to produce a genotypic profile for two or more (e g., 2, 3, 4, 5, 6, 7, 8, 9, 10, 100, 1000, 10,000 or more) inbred lines of the plurality of inbred lines, selecting from the genotypic profiles a first subset of inbred lines having genetic similarity to the first parent genomic profile and a second subset of inbred lines having genetic similarity to the second parent genomic profile, generating an in silica hybrid population comprising a plurality of simulated hybrids each simulated hybrid of the population having a parent from the first subset of inbred lines and a parent from the second subset of inbred lines, generating a predicted genotypic and / or phenotypic profile for one or more plants of the hybrid population, comparing the predicted genotypic and / or phenotypic profile for the one or more plants of the hybrid population to the genotypic and / or phenotypic profile of the elite hybrid, identifying plants of the hybrid population that have genotypic and / or phenotypic similarity to the elite hybrid, and growing the identified plants of the hybrid population and selecting plants that have one or more improved agronomic traits as compared to the elite hybrid plant to generate hybrid plants with one or more improved agronomic traits as compared to the elite hybrid.

[0009] A “genomic profile”, “genotypic profile”, or the like as used herein generally refers to a set of information about the entire genome of a given plant or group of plants (genome-wide), or it can encompass a specific subset of the genome of a given plant or group of plants, or any combination thereof in a given plant or group of plants. The genotypic profile (also referred toherein as genotype) of the plants and members of the populations described herein (e g., inbred parents, inbred population, hybrid population) may be determined or generated using any method known in the art. In certain embodiments, the genomic profile of the plants and members of the populations described herein (e.g., inbred parents, inbred population, hybrid population) includes information regarding the presence or absence in the genome of a specific set of mutations, single nucleotide polymorphisms (SNPs), insertion of bases, deletion of bases, genotypic markers, other sequence information, or any combination thereof.

[0010] In certain embodiments of the methods described herein, the genotypic profile is determined using molecular biological assays such as, for example, PCR, DNA sequencing, restriction fragment length polymorphism identification, SNP genotyping or whole genome sequencing, such that the genotypic profile comprises an observed genotype (e.g., observed SNP markers). The plants can be genotyped individually or from a pooled DNA sample. In certain embodiments of the methods described herein, the genotypic profile is a predicted genotypic profile such that, for example, the genotypic profile comprises an imputed genotype (e.g., imputed SNP markers). In certain embodiments of the methods described herein, the genotypic profile comprises both observed genotypic information and predicted genotypic information, such that, for example, the genotype comprises both observed SNP markers and imputed SNP markers.

[0011] In certain embodiments, the genotypic profile is predicted using variational autoencoders (VAEs). For example, in certain embodiments, to predict the genotypes of the plants and members of the populations described herein parental SNPs are imputed using VAEs trained for optimal reconstruction of the population, such as, for example, samples from a breeding program. Non-limiting examples of using VAEs trained for optimal reconstruction include, but are not limited to, those found in US Patent No. 11,174,522. In certain embodiments, the VAEs produce intermediate latent representations that could be decoded into imputed SNPs. The genotype of the plants and members of the populations described herein can be simulated from the imputed SNPs. In certain embodiments, the genotypes of the plants and members of the populations described herein can be simulated using a statistical model of recombination. In certain embodiments, the statistical model of recombination is based on a Poisson process. For example, in certain embodiments to simulate SNPs based on a Poisson process recombination break points in the genetic map may be sampled from an exponential distribution with a rateparameter equal to one, and the sampled distance multiplied by 100 for conversion to centimorgans. In certain embodiments, the crossover interference rate can be sampled from a gamma distribution with shape and scale parameters equal to two. As would be understood by a person of ordinary skill in the art, the exact parameters to simulate the population using a statistical model of recombination based on the Poisson process may be adjusted based on the complexity of the genome of the population.

[0012] VAEs are hybrids of deep neural networks and probabilistic graphical models that enable construction of a compressed latent representation that is independent of the underlying data generation (e.g., genotyping platform) and serves as a basis of imputing characteristics of a desired data set (e.g., multiple germplasm characterization). The core of VAEs is rooted in Bayesian inference, which includes modeling of the underlying probability distribution of data, such that new data can be sampled from that distribution, which is independent of the dataset that resulted in the probability distribution. VAEs have a property that separates them from standard autoencoders that is suitable for generative modeling: the latent spaces that VAEs generate are, by nature of the framework, probability distributions, thereby allowing simpler random sampling and interpolation for desirable end-uses. VAEs accomplish this latent space representation by making its encoder not output an encoding vector of size n, rather, outputting two vectors of size n: a vector of means, p, and another vector of standard deviations, G. Some of the basic notions for VAE include for example:X: data that needs to be modeled, for example, genotypic data (such as SNPs, markers, haplotype, sequence information) z: latent variableP(X): probability distribution of the data, for example, genotypic dataP(z): probability distribution of latent variable (e.g., genotypic associations from the underlying genotypic data)P(X|z): distribution of generating data given latent variable, e.g. prediction or imputation of the desired outcome based on the latent variable.

[0013] VAE is based on the principle that if there exists a hidden variable z, which generates an observation or an outcome x, then one of the objectives is to model the data, i.e., to find P(X). However, one can observe x, but the characteristics of z need to be inferred. Thus, p(z|x) needs to be computed.p(z|x) = p(x|z)p(z) / p(x)

[0014] However, computing p(x) is based on probability theory, in relation to z. This function can be expressed as follows: p(x) = J p(x|z)p(z)dz

[0015] While the p(x) function is an intractable distribution, variational inference is used to optimize the joint distribution of x and z. The function p(z|x) is approximated by another distribution q(z|x), which is defined such that it is a tractable distribution. The parameters of q(z|x) are defined such that they are highly similar to p(z|x) and therefore, it can be used to perform approximate inference of the intractable distribution. KL divergence is a measure of difference between two probability distributions. Therefore, if the goal is to minimize the KL divergence between the two distributions, this minimization function is expressed as: min KL (q(z|x) | |p(z|x))

[0016] This expression is minimized by maximizing the following:Eq z\x) logp(x|z) - KL(q(z|x) | |p(z))

[0017] Reconstruction likelihood is represented by the first part, and the second term penalizes departure of probability mass in q from the prior distribution, p. q is used to infer hidden variables (latent representation) and this is built into a neural network architecture where the encoder model learns the mapping relation from x to z and the decoder model learns the mapping from z back to x. Therefore, the neural network for this function includes two terms - one that penalizes reconstruction error or maximizes the reconstruction likelihood and the other that encourages the learned distribution q(z|x) to be highly similar to the true prior distribution p(z), which is assumed to follow a unit Gaussian distribution, for each dimension j of the latent space. This is represented by:

[0018] It should be appreciated that the variational autoencoder is one of several techniques that may be used for producing compressed latent representations of raw samples, for example, genotypic association data. Like other autoencoders, the variational autoencoder places a reduced dimensionality bottleneck layer between an encoder and a decoder neural network. Optimizing the neural network weights relative to the reconstruction error then produces separation of the samples within the latent space. However, unlike generative adversarial networks (GAN), theencoder neural network’s outputs are parameterized univariate Gaussian distributions with standard N(0,l) priors. Thus, unlike other autoencoders, which tend to memorize inputs and place them in arbitrarily small locations within the latent space, the variational autoencoder produces a smooth, continuous latent space in which semantically-similar samples tend to be geometrically close - e.g., haplotypes that co-segregate to provide a certain phenotype.

[0019] For example, in the context of genomic characterization, a smooth spatial organization of the latent space captures varying levels of ancestral relationships that are present within a dataset. Genomic variation within a population such as a plant breeding program may be characterized by a variety of methods. For example, genotypes are characterized with a common platform that interrogates localized variants such as single nucleotide polymorphisms (SNPs) and / or insertions / deletions (indels). Due to the ancestral recombination and demographic history of the population, these variants tend to co-segregate within linked segments (haplotypes). Further, single genotypes may then be further characterized by the set of haplotypes they contain. For example, as described further below, VAEs may be used to compress the information contained within a given set of production markers to a common, marker-invariant, latent space capable of capturing these co-segregation patterns genome- wide.

[0020] As used herein, an inbred plant line refers to a line that has been bred for genetic homogeneity and includes, but is not limited to, plants generated from plant breeding techniques (e g., self-pollination, backcrossing) and double haploid plants, (e.g., plants derived from microspores or haploid embryos), including but not limited to Fl or F2 doubled haploid plants, offspring or progeny thereof. The number of inbred plant lines of the plurality of inbred lines for use in the methods described herein is not particularly limited. In certain embodiments, the plurality of inbred lines comprises at least 1000, 5000, 10,000, 50,000, 100,000, 250,000, 500,000, 1,000,000, or 10,000,000 inbred plant lines. In certain embodiments, the plurality of inbred lines comprises inbred lines developed by traditional breeding, genome edited breeding, or a combination thereof, such that, in certain embodiments, the plurality of inbred lines comprises inbred lines having one or more genome edits.

[0021] Traditional breeding refers to methods to develop new plants with desirable traits by choosing plants with favorable characteristics and crossing them to produce offspring with improved traits. Plants (e.g., inbred lines) developed by traditional breeding for the methods provided herein include non-modified plants and plants comprising transgenes, such as thoseintroduced by trait introgression or transformation of the plant. Genome edited breeding refers to methods to develop new plants (e.g., inbred lines) with desirable traits that uses genome editing technologies, such as, for example CRISPR / Cas, TALENs, or ZFNs, to precisely modify the DNA of an organism to produce plants with improved traits. Inbred lines developed by genome edited breeding for the methods provided herein comprise at least one genome edit. Inbred lines containing at least one genome edit are created through trait integration or through direct editing of the plant material.

[0022] In certain embodiments, the first subset of inbred lines comprises inbred lines having 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 97.5%, 98%, 98.5%, 99%, 99.1%, 99.2%, 99.3%, 99.4%, 99.5%, 99.6%, 99.7%, 99.8%, or 99.9% or more genetic similarity to the first inbred parent. In certain embodiments of the method described herein, the second subset of inbred lines comprise inbred lines having 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 97.5%, 98%, 98.5%, 99%, 99.1%, 99.2%, 99.3%, 99.4%, 99.5%, 99.6%, 99.7%, 99.8%, or 99.9% or more genetic similarity to the second inbred parent. In certain embodiments of the methods described herein, both the first subset of inbred lines and the second subset of inbred lines comprise inbred lines having 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 97.5%, 98%, 98.5%, 99%, 99.1%, 99.2%, 99.3%, 99.4%, 99.5%, 99.6%, 99.7%, 99.8%, or 99.9% to the first parent and second parent, respectively. In certain embodiments, the first subset of inbred lines, the second subset of inbred lines or both the first subset and second subset comprise inbred lines having one or more genome edits.

[0023] As used herein, genetic similarity refers to the genetic relatedness among individuals (e.g., an inbred plant of the first subset and the first parent). The genetic similarity between individuals can be calculated using a percent similarity. As used herein, “percent similarity” “percent genetic similarity” or the like refers to the comparison of the homozygous alleles of first plant with a second plant, and if the homozygous allele of the first plant matches at least one of the alleles from the second plant, then they are scored as similar. Percent genetic similarity can be calculated by comparing a statistically significant number of loci and recording the number of loci with similar alleles as a percentage.

[0024] In certain embodiments, the first subset of inbred plant lines selected from the plurality of inbred lines comprise no more than 70%, 65%, 60%, 55%, 50%, 45%, 40%, 35%, 30%, 25%, 20%, 15%, 10%, or 5% of the plants from the plurality of inbred plant lines. In certainembodiments, the second subset of inbred plant lines selected from the plurality of inbred lines comprise no more than 70%, 65%, 60%, 55%, 50%, 45%, 40%, 35%, 30%, 25%, 20%, 15%, 10%, or 5% of the plants from the plurality of inbred plant lines. In certain embodiments, both the first subset of inbred plant lines and second subset of inbred plant lines selected from the plurality of inbred lines comprise no more than 70%, 65%, 60%, 55%, 50%, 45%, 40%, 35%, 30%, 25%, 20%, 15%, 10%, or 5% of the plants from the plurality of inbred plant lines.

[0025] In certain embodiments, the first subset of inbred lines, the second subset of inbred lines, or both the first and second subset of inbred lines comprise at least 10, 25, 50, 100, 250, or 500 inbred lines and fewer than 10,000, 5,000, 1,000, 500, 250, or 150 inbred lines.

[0026] In certain embodiments, the first subset of inbred lines, the second subset of inbred lines, or both the first and second subset of inbred lines comprise at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 25, 50, 100, 250, 500, or more inbred line or all inbred lines that do not share both parents with the first inbred parent line or second inbred parent line, respectively.

[0027] As used herein, “clustering”, “cluster analysis” or the like refers to a grouping of objects (e.g., inbred plant lines) based on similarity, such that objects in the same cluster are more similar to each other than objects in other clusters. The inbred plant lines of the methods described herein can be grouped (clustered) based on any characteristic of the plant (e.g., genotypic and / or phenotypic characteristic). In certain embodiments, the inbred plant lines of the methods described herein are clustered based on genotype. In certain embodiments of the methods described herein, the inbred plant lines are grouped into at least 2, 5, 10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 250, 500, or 1000 or more clusters.

[0028] The method for clustering the inbred plant lines of the methods described herein, may be any cluster analysis known in the art that can group the inbred plant lines, such as for example, hierarchical clustering, centroid-based clustering, distribution-based clustering, and densitybased clustering. In certain embodiments, the clustering comprises the use of hierarchical clustering. In certain embodiments, the hierarchical clustering uses agglomerative hierarchical clustering to group inbreds into clusters based on their genomic molecular marker similarity.

[0029] As used herein, a "trait" an “agronomic trait” or the like refers to a physiological, morphological, biochemical, or physical characteristic of a plant or particular plant material or cell. In some instances, this characteristic is visible to the human eye, such as seed or plant size, or can be measured by biochemical techniques, such as detecting the protein, starch, or oilcontent of seed or leaves, or by observation of a metabolic or physiological process, e.g. by measuring uptake of carbon dioxide, or by the observation of the expression level of a gene or genes, e.g., by employing Northern blot analysis, RT-PCR, microarray gene expression assays, or reporter gene expression systems, or by agricultural observations such as stress tolerance, yield, or pathogen tolerance. Examples of improved agronomic traits of the methods described herein includes, but is not limited to, increased yield, increased disease resistance (e.g., southern rust resistance, southern leaf blight resistance, anthracnose stalk rot resistance, northern corn leaf blight resistance, com lethal necrosis resistance, common smut resistance, common rust resistance, diplodia mold resistance, diplodia stalk rot resistance, eyespot resistance, fusarium ear rot resistance, gibberella ear rot resistance, gibberella stalk rot resistance, gray leaf spot resistance, goss’ wilt resistance, helminthosporium carbonum leaf blight resistance, head smut resistance, maize dwarf mosaic complex resistance, rust resistance, powdery or downy mildew resistance, leaf spot resistance, fusarium stalk rot), decreased brittle snap, decreased root lodging, relative maturity, growing degree days for silking, drydown, flowering time, pod shatter, improved abiotic traits (e.g., drought tolerance, iron deficiency tolerance, chloride tolerance, temperature, nitrogen use efficiency), improved seed and / or kernel composition (e.g., moisture, test weight, protein percentage, oil percentage, meal type, oleic acid percentage, linoleic acid percentage, linolenic acid percentage), enhanced herbicide tolerance, enhanced insect resistance, fertility (e.g., pollen shed), silage yield, and morphological traits (e.g., ear height, decreased plant height).

[0030] The agronomic trait in which an improvement is desired is not particularly limited and may be selected based on the local environmental conditions. In certain embodiments, the one or more improved agronomic traits comprises an improved yield as compared to the elite hybrid plant. In certain embodiments, the selected hybrid plants comprise at least 0.5%, 1%, 1.5%, 2%, 2.5%, 3%, 3.5%, 4%, 4.5%, or 5%, increase in yield as compared to the elite hybrid plant.

[0031] As used herein, “yield” refers to the amount of agricultural production harvested per unit of land and may include reference to bushels per acre or kilograms per hectare of a crop at harvest, as adjusted for grain moisture. Grain moisture is measured in the grain at harvest. The adjusted test weight of grain is determined to be the weight in pounds per bushel or kilogram, adjusted for grain moisture level at harvest.

[0032] In certain embodiments, the one or more improved agronomic traits comprises an enhanced or increased resistance or tolerance to at least one disease, insect or herbicide, such as those described herein, as compared to the elite hybrid plant. As used herein, “increase in resistance”, “increase in tolerance”, “increased resistance”, “increased tolerance”, “enhanced resistance”, “enhanced tolerance” and the like refers to any detectable reduction of one or more symptoms in a plant caused by a plant pathogen (e.g., disease resistance), insect, or herbicide.

[0033] As used herein, an “elite line” is an agronomically superior line that has resulted from many cycles of breeding and selection for superior agronomic performance. Numerous elite lines are available and known to those of skill in the art of breeding.

[0034] The hybrid plant of the methods described herein is not particularly limited and may be any hybrid plant for which an improved agronomic trait is desired. Examples of plant species of interest include, but are not limited to, maize (Zea mays), wheat (Triticu aestivum), Brassica sp. (e.g., B. napus, B. rapa, B. juncea), particularly those Brassica species useful as sources of seed oil, sorghum (Sorghum bicolor, Sorghum vulgare), sunflower (Helianthus annuus), potato (Solanum tuberosum), pea (Lathyrus spp), and cotton (Gossypium barbadense, Gossypium hirsutum).

[0035] Also provided are methods for selecting a plant and / or producing hybrid plants with improved agronomic traits comprising genotyping a first inbred parent and a second inbred parent of an elite hybrid to produce a first parent genotypic profile and a second parent genotypic profile, generating a genotypic profile for members of a population, the population comprising a plurality of inbred lines, clustering the members of the population of inbred lines to generate clusters of inbred lines having genetic similarity, selecting a first cluster of inbred lines having genetic similarity to the first parent genomic profile and a second cluster of inbred lines having genetic similarity to the second parent genomic profile, generating an in silico hybrid population comprising a plurality of simulated hybrids each simulated hybrid of the population having a parent from the first cluster of inbred lines and a parent from the second cluster of inbred lines, generating a predicted phenotypic profile for at least one trait for one or more plants of the in silico hybrid population, identifying hybrid plants of the in silico hybrid that are predicted to have an improved phenotypic profile for the at least one trait as compared to the phenotypic profile of the elite hybrid. In certain embodiments, the method further comprises growing theidentified hybrid plants of the hybrid in silico population and selecting hybrid plants that have one or more improved agronomic traits as compared to the elite hybrid plant.

[0036] Further provided are methods for selecting a plant and / or producing hybrid plants with improved agronomic traits comprising genotyping a first inbred parent and a second inbred parent of an elite hybrid to produce a first parent genotypic profile and a second parent genotypic profile, generating a genotypic profile for members of a population, the population comprising a plurality of inbred lines, clustering the members of the population of inbred lines to generate clusters of inbred lines having genetic similarity, selecting a first subset of inbred plant lines from a first cluster of inbred lines having genetic similarity to the first parent genomic profile and a second subset of inbred plant lines from the second cluster of inbred lines having genetic similarity to the second parent genomic profile, generating an in silico hybrid population comprising a plurality of simulated hybrids each simulated hybrid of the population having a parent from the first subset of inbred lines and a parent from the second subset of inbred lines, generating a predicted phenotypic profile for at least one trait for one or more plants of the in silico hybrid population, identifying hybrid plants of the in silico hybrid that are predicted to have an improved phenotypic profile for the at least one trait as compared to the phenotypic profile of the elite hybrid, and growing the identified hybrid plants of the hybrid in silico population and selecting hybrid plants that have one or more improved agronomic traits as compared to the elite hybrid plant.

[0037] The number of inbred plant lines of the population for use in the methods described herein is not particularly limited. In certain embodiments, the population of inbred lines comprises at least 1000, 5000, 10,000, 50,000, 100,000, 250,000, 500,000, 1,000,000, or 10,000,000 inbred plant lines.

[0038] The phenotypic profile of the plants of the in silico hybrid population may be predicted using any method known in the art that can predict a plant phenotype from genotypic information. In certain embodiments, the phenotypic profile is predicted using a machine learning model, a learned prediction model, best linear unbiased prediction (BLUP), calculation of genomic estimated breeding values (GEBVs), or a combination thereof.

[0039] Best linear unbiased prediction (BLUP) is a statistical method to predict breeding values, such as, for example, genomic estimated breeding values, using molecular markers. Methods for using BLUP to predict plant phenotypes are known in the art.

[0040] As used herein, “genomic estimated breeding values” (GEBVs) refer to a measurable degree to which one or more, polymorphic markers, haplotypes and / or genotypes heritably affect the expression of a phenotype associated with a trait. Standard methods of genomic selection estimate effects of genome-wide molecular markers to calculate genomic estimated breeding values (GEBVs) for individuals without phenotypes. For example, GEBV can be used as a selection criterion by predicting phenotypic performance with canonical methods computing predictions using genetic marker data that measure allelic states at genome-wide loci.

[0041] In certain embodiments, the machine learning model for use in the methods described herein is an artificial neural network (ANN). ANNs are configured to synthesize or learn from a plurality of inputs to produce an output. One or more variables in the algorithms can have weights that are applied to each equation and optimized as the neural network is trained. Based on the amount of training information, the deep learning models or networks get better at producing more helpful outputs. In certain embodiments, the ANN includes a plurality of input factors that may be used to train predicted phenotypic information. These factors include, but are not limited to, QTLs, SNPs, haplotypes, yield, and other historical agronomic or breeding phenotypic components.

[0042] In certain embodiments, the learned prediction model comprises a linear statistical model. As used herein, a “linear statistical model” is a model that provides a linear relationship between an independent variable (e.g., genotype) and a dependent variable (e.g., phenotype) to predict the outcome of future events. The type of linear statistical model is not particularly limited and may be any linear statistical model known in the art. In certain embodiments, the linear statistical model comprises a logistic generalized linear prediction model.

[0043] In certain embodiments, the machine learning model for use in the methods described herein is a Bayesian genomic prediction model (e g., BayesA, BayesB, or BayesC). In certain embodiments, the training dataset for the Bayesian genomic prediction model comprises phenotypes at multiple locations for the trait or traits of interest and marker genotype scores from a plurality of material representative of the germplasm of interest. In certain embodiments, the training dataset is used to estimate the marker effects for the trait of interest in the population at a given environment using the prediction model BayesA, BayesB, and / or BayesC.

[0044] In certain embodiments of the method described herein, the first subset of inbred lines from the first cluster comprises inbred lines having 90%, 91%, 92%, 93%, 94%, 95%, 96%,97%, 97.5%, 98%, 98.5%, 99%, 99.1%, 99.2%, 99.3%, 99.4%, 99.5%, 99.6%, 99.7%, 99.8%, or 99.9% or more genetic similarity to the first inbred parent. In certain embodiments of the method described herein, the second subset of inbred lines from the second cluster comprises inbred lines having 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 97.5%, 98%, 98.5%, 99%, 99.1%, 99.2%, 99.3%, 99.4%, 99.5%, 99.6%, 99.7%, 99.8%, or 99.9% or more genetic similarity to the second inbred parent. In certain embodiments of the methods described herein, both the first subset of inbred lines from the first cluster and the second subset of inbred lines from the second cluster comprise inbred lines having 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 97.5%, 98%, 98.5%, 99%, 99.1%, 99.2%, 99.3%, 99.4%, 99.5%, 99.6%, 99.7%, 99.8%, or 99.9% to the first parent and second parent, respectively.

[0045] In certain embodiments, the first subset of inbred lines from the first cluster, the second subset of inbred lines from the second cluster, or both the first and second subset of inbred lines comprise at least 10, 25, 50, 100, 250, or 500 inbred lines and fewer than 10,000, 5,000, 1,000, 500, 250, or 150 inbred lines.

[0046] Also provided are computer readable mediums having stored thereon instructions that, when executed by a processor (or computing device), cause the processor to perform the steps of the methods described herein to provide hybrid plants that have one or more improved characteristics as compared to the elite target hybrid plant.

[0047] Further disclosed herein are systems (e.g., computer systems) for use in hybrid plant breeding that include (a) one or more servers, each of the one or more server storing plant data, and (b) a computing device communicatively coupled to the one or more servers, the computing device including: (1) a memory, and (2) one or more processors configured to perform operations to: (a) screed a plurality of inbred lines to produce genotypic profiles, (b) selecting from the plurality of genotypic profiles a first subset of inbred lines having genetic similarity to the first parent genomic profile and a second subset of inbred lines having genetic similarity to the second parent genomic profile, (c) generating an in silica hybrid population comprising a plurality of simulated hybrids each simulated hybrid of the population having a parent from the first subset of inbred lines and a parent from the second subset of inbred lines, (d) generating a predicted genotypic profile for one or more plants of the hybrid population, (e) comparing the predicted genotypic profile for the one or more plants of the hybrid population to the genotypic profile of the elite hybrid, and (g) identifying hybrid plants of the hybrid population that havegenotypic similarity to the elite hybrid. The one or more processors may also be configured to perform operations to genotype a first inbred parent and a second inbred parent of an elite hybrid to produce a first parent genotypic profile and a second parent genotypic profile.

[0048] The type of system (e.g., computer system) is not particularly limited and may be any system comprising a computing device and one or more servers such as the system provided in Fig. 2. Referring to Fig. 2, a block diagram of a computer system 100 to provide hybrid plants that have one or more improved characteristics as compared to the elite target hybrid plant. To do so, the system 100 may include a computing device 110 and a server 130 that is associated with a computer system. The system 100 may further include one or more servers 140 that are associated with other computer systems such that the computing device 110 may communicate with different computer systems running different platforms. However, it should be appreciated that, in some embodiments, a single server (e.g., a server 130) may run multiple platforms. The computing device 110 is communicatively coupled to the one or more servers 130, 140 via a network 150 (e.g., a local area network (LAN), a wide area network (WAN), a personal area network (PAN), the Internet, etc.).

[0049] In general, the computing device 110 may include any existing or future devices capable of cluster analysis such as, for example, training a machine learning model to perform the cluster analysis. For example, the computing device may be, but not limited to, a computer, a notebook, a laptop, a mobile device, a smartphone, a tablet, wearable, smart glasses, or any other suitable computing device that is capable of communicating with the server 130.

[0050] The computing device 110 includes a processor 112, a memory 114, an input / output (I / O) controller 116 (e.g., a network transceiver), a memory unit 118, and a database 120, all of which may be interconnected via one or more address / data bus. It should be appreciated that although only one processor 112 is shown, the computing device 110 may include multiple processors. Although the I / O controller 116 is shown as a single block, it should be appreciated that the I / O controller 116 may include a number of different types of I / O components (e.g., a display, a user interface (e.g., a display screen, a touchscreen, a keyboard), a speaker, and a microphone).

[0051] The processor 112 as disclosed herein may be any electronic device that is capable of processing data, for example a central processing unit (CPU), a graphics processing unit (GPU), a tensor processing unit (TPU), a system on a chip (SoC), or any other suitable type of processor. It should be appreciated that the various operations of example methods described herein (i.e.,performed by the computing device 110) may be performed by one or more processors 112. The memory 114 may be a random-access memory (RAM), read-only memory (ROM), a flash memory, or any other suitable type of memory that enables storage of data such as instruction codes that the processor 112 needs to access in order to implement any method as disclosed herein. It should be appreciated that, in some embodiments, the computing device 110 may be a computing device or a plurality of computing devices with distributed processing.

[0052] As used herein, the term “database” may refer to a single database or other structured data storage, or to a collection of two or more different databases or structured data storage components. In the illustrative embodiment, the database 120 is part of the computing device 110. In some embodiments, the computing device 110 may access the database 120 via a network such as network 150. The database 120 may store data (e.g., input, output, intermediary data) used for selecting subsets of lines having genetic similarity and / or generating an in silica hybrid population comprising a plurality of simulated hybrids. For example, the data may include genotypic data, such as single nucleotide polymorphisms (SNPs), genetic markers, haplotype, sequence information, phenotypic data, environment data, production costs, pedigree information, co-ancestry information, or combinations thereof that are obtained from one or more servers 130, 140.

[0053] The computing device 110 may further include a number of software applications stored in a memory unit 118, which may be called a program memory. The various software applications on the computing device 110 may include specific programs, routines, or scripts for performing processing functions associated with the methods described herein. Additionally, or alternatively, the various software applications on the computing device 110 may include general -purpose software applications for data processing, database management, data analysis, network communication, web server operation, or other functions described herein or typically performed by a server. The various software applications may be executed on the same computer processor or on different computer processors. Additionally, or alternatively, the software applications may interact with various hardware modules that may be installed within or connected to the computing device 110. Such modules may implement part of or all of the various exemplary method functions discussed herein or other related embodiments.

[0054] Although only one computing device 110 is shown in Fig. 2, the server 130, 140 is capable of communicating with multiple computing devices similar to the computing device 110.Although not shown in Fig. 2, similar to the computing device 1 10, the server 130, 140 also includes a processor (e.g., a microprocessor, a microcontroller), a memory, and an input / output (I / O) controller (e.g., a network transceiver). The server 130, 140 may be a single server or a plurality of servers with distributed processing. The server 130, 140 may receive data from and / or transmit data to the computing device 110.

[0055] In certain embodiments, the computing device 110 may generate an in silico hybrid population comprising a plurality of simulated hybrids each simulated hybrid of the population having a parent from the first subset of inbred lines and a parent from the second subset of inbred lines. In certain embodiments, the computing device 110 may further generate predictions of plant genotype performance by using at least one learned prediction model to generate a predicted phenotype for the plants of the hybrid population.

[0056] The following are examples of specific embodiments of some aspects of the invention. The examples are offered for illustrative purposes only and are not intended to limit the scope of the invention in any way.EXAMPLE 1

[0057] This example demonstrates generation of a new maize hybrid variety having improved characteristics compared to a top selling hybrid in Argentina.

[0058] The maize target hybrid for improvement is a top selling hybrid in Northern Argentina and was a top hybrid for several years in the southern United States. Even though this target hybrid is a top-end yield product, some of its main weaknesses include Northern Leaf Blight tolerance (having intermediate to low tolerance scores), plant height, stalk lodging, root strength, and plant integrity. A hierarchical clustering (HC) analysis was initiated using all commercial inbreds and early stage pre-commercial inbreds readily available from Latin American and North American evaluation zones. In total, the analysis included nearly 460,000 genetic elements (GEs) from several different evaluation zones. Observed and imputed marker scores were retrieved from approximately 24,000 current production SNPs and transformed into 64-dimensional latent encodings based on the neural network encoders used in the worldwide com bridge imputation process. Clusters of GEs were then formed using the agglomerative hierarchical clustering algorithm implemented in scikit-learn, with connectivity constraints imposed from a 50-neighbor KNN graph of the latent vectors based on Euclidean distance. The granularity of clusters wasvaried by running the algorithm with 30, 50, and 100 total clusters. For the parents of the reference hybrid, the 100 nearest female neighbors and 100 nearest male neighbors across commercial inbreds and early stage pre-commercial inbreds were identified. Then, the 100 most genetically similar combinations to the target hybrid were generated in silico and processed through WGP, followed by index selection across estimation sets (ES) predictions. Hybrid combinations were selected through improvement of target hybrid weaknesses based on local ES. Eleven hybrids were created in winter nursery experiments and tested in the TPE (late planting in Argentina). A comparison of the predicted phenotypic scores and the field phenotypic scores for the trait targets is shown in Table 1 . The Northern Leaf Blight field scores correspond to a scale of 1-9 where 1 is fully susceptible and 9 is fully tolerant.Table 1 : Predicted Phenotypic Scores Compared to Field Trials Conducted in Argentina in 2022* Estimation set based on 5 years data. Values are provided as adjusted relative values (ARV).EXAMPLE 2

[0059] This example demonstrates generation of a new maize hybrid variety having improved characteristics compared to a top selling hybrid in Mexico.

[0060] The maize target hybrid for improvement two top selling hybrids in Mexico. A hierarchical clustering (HC) analysis was initiated using all commercial inbreds and early stage pre-commercial inbreds readily available from Latin American and North American evaluation zones. In total, the analysis included nearly 460,000 genetic elements (GEs) from several different evaluation zones. Observed and imputed marker scores were retrieved from approximately 24,000 current production SNPs and transformed into 64-dimensional latent encodings based on the neural network encoders used in the worldwide com bridge imputation process. Clusters of GEs were then formed using the agglomerative hierarchical clustering algorithm implemented in scikit-learn, with connectivity constraints imposed from a 50-neighbor KNN graph of the latent vectors based on Euclidean distance. The granularity of clusters was varied by running the algorithm with 30, 50, and 100 total clusters. For the parents of the reference hybrids, the 100 nearest female neighbors and 100 nearest male neighbors across commercial inbreds and early stage pre-commercial inbreds were identified. Then, the 100 most genetically similar combinations to the target hybrids were generated in silico and processed through WGP, followed by index selection across estimation sets (ES) predictions. Hybrid combinations were selected through improvement of the Target hybrids weaknesses based on local ES. Twenty hybrids were created in winter nursery experiments and tested in 2022 in the TPE (late planting in Argentina). A comparison of the predicted phenotypic scores and the field phenotypic scores for the trait targets is shown in Table 2.

[0061] A similar clustering was performed for a third target hybrid. Twenty hybrids were created in winter nursery experiments and tested in 2023 in the TPE (late planting in Argentina). A comparison of the predicted phenotypic scores and the field phenotypic scores for the trait targets is shown in Table 3.Table 2: Predicted Phenotypic Scores Compared to Field Trials Conducted in Mexico in 2022* Estimation set based on 5 years data. Values are provided as adjusted relative values (ARV).Table 3: Predicted Phenotypic Scores Compared to Field Trials Conducted in Mexico in 2023* Estimation set based on 5 years data. Values are provided as adjusted relative values (ARV).EXAMPLE 3

[0062] This example demonstrates generation of a new sunflower hybrid variety having improved characteristics.

[0063] The sunflower target hybrid is a top hybrid for central Europe. In the analysis, 244 inbreds and 4,304 conventional breeding GEs were involved. The 100 most genetically similar to both hybrid parental inbreds were detected and predicted with WGP, followed by index selection to short-list the most suitable genetic combinations. Then, genetically similar combinations to the target hybrids are generated in silico and processed through WGP, followed by index selection across estimation sets (ES) predictions. Hybrid combinations are selected through improvement of the Target hybrid weaknesses based on local ES. Hybrids are created in nursery experiments and are tested in a TPE.

[0064] All publications and patent applications in this specification are indicative of the level of ordinary skill in the art to which this invention pertains. All publications and patent applications are herein incorporated by reference to the same extent as if each individual publication or patent application was specifically and individually indicated by reference.

[0065] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Unless mentioned otherwise, the techniques employed or contemplated herein are standard methodologies well known to one of ordinary skill in the art. The materials, methods and examples are illustrative only and not limiting.

[0066] Many modifications and other embodiments of the inventions set forth herein will come to mind to one skilled in the art to which these inventions pertain having the benefit of the teachings presented in the foregoing descriptions and the associated drawings. Therefore, it is to be understood that the inventions are not to be limited to the specific embodiments disclosed and that modifications and other embodiments are intended to be included within the scope of theappended claims. Although specific terms are employed herein, they are used in a generic and descriptive sense only and not for purposes of limitation.

[0067] Units, prefixes and symbols may be denoted in their SI accepted form. Unless otherwise indicated, nucleic acids are written left to right in 5’ to 3’ orientation; amino acid sequences are written left to right in amino to carboxy orientation, respectively. Numeric ranges are inclusive of the numbers defining the range. Amino acids may be referred to herein by either their commonly known three letter symbols or by the one-letter symbols recommended by the IUPAC-IUB Biochemical Nomenclature Commission. Nucleotides, likewise, may be referred to by their commonly accepted single-letter codes.

Claims

We claim:

1. A method of generating a hybrid plant having one more improved agronomic characteristics as compared to an elite hybrid plant, the method comprising: a. genotyping a first inbred parent and a second inbred parent of an elite hybrid to produce a first parent genotypic profile and a second parent genotypic profile; b. screening a plurality of inbred lines to produce a plurality of genotypic profiles for four or more inbred lines of the plurality of inbred lines; c. selecting from the plurality of genotypic profiles a first subset of inbred lines having genetic similarity to the first parent genomic profile and a second subset of inbred lines having genetic similarity to the second parent genomic profile; d. generating an in silica hybrid population comprising a plurality of simulated hybrids each simulated hybrid of the population having a parent from the first subset of inbred lines and a parent from the second subset of inbred lines; e. generating a predicted genotypic profile for one or more plants of the hybrid population; f. comparing the predicted genotypic profile for the one or more plants of the hybrid population to the genotypic profile of the elite hybrid; g. identifying hybrid plants of the hybrid population that have genotypic similarity to the elite hybrid and are predicted to have the one or more improved agronomic characteristics; and h. growing the identified hybrid plants of the hybrid population and selecting hybrid plants that have the one or more improved agronomic traits as compared to the elite hybrid plant.

2. The method of claim 1, wherein the plurality of inbred plant lines comprises at least 10,000 inbred plant lines.

3. The method of claim 1 or 2, wherein the first subset of inbred plant lines, the second subset of inbred plant lines, or both comprise no more than 50% of the plants from the plurality of inbred plant lines.

4. The method of any one of claims 1-3, wherein members of the first subset of inbred plant lines, members of the second subset of inbred plant lines, or both are selected using cluster analysis.

5. The method of claim 4, wherein the cluster analysis comprises hierarchical clustering.

6. The method of any one of claims 1-5, wherein the phenotypic profile is predicted using a machine learning model.

7. The method of any one of claims 1-6, wherein the phenotypic profile is predicted using best linear unbiased prediction (BLUP).

8. The method of any one of claims 1-7, wherein the one or more improved traits is selected from the group consisting of yield, lodging, disease resistance, maturity, herbicide tolerance, insect resistance, plant height, or any combination thereof.

9. The method of any one of claims 1-8, wherein the selected hybrid plants comprise at least a 5% increase in yield as compared to the elite hybrid variety.

10. A method of selecting a hybrid plant having one more improved agronomic characteristics as compared to an elite hybrid plant, the method comprising: a. genotyping a first inbred parent and a second inbred parent of an elite hybrid to produce a first parent genotypic profile and a second parent genotypic profile; b. generating a genotypic profile for members of a population, the population comprising a plurality of inbred lines; c. clustering the members of the population of inbred lines to generate clusters of inbred lines having genetic similarity; d. selecting a first cluster of inbred lines having genetic similarity to the first parent genomic profile and a second cluster of inbred lines having genetic similarity to the second parent genomic profile;e. generating an in silica hybrid population comprising a plurality of simulated hybrids each simulated hybrid of the population having a parent from the first cluster of inbred lines and a parent from the second cluster of inbred lines; f. generating a predicted phenotypic profile for at least one trait for one or more plants of the in silica hybrid population; g. identifying hybrid plants of the in silica hybrid predicted to have improved phenotypic profile for the at least one trait as compared to the phenotypic profile of the elite hybrid; and h. growing the identified hybrid plants of the hybrid in silica population and selecting hybrid plants that have one or more improved agronomic traits as compared to the elite hybrid plant.

11. The method of claim 10, wherein the population comprises at least 10,000 inbred plant lines.

12. The method of claim 10 or 11, wherein the genotypic profile for members of the population is generated using observed SNP markers, imputed SNP markers, or a combination thereof.

13. The method of any one of claims 10-12, wherein the clustering comprises hierarchical clustering.

14. The method of any one of claims 10-13, wherein the population is separated into at least 20 total clusters.

15. The method of any one of claims 10-14, wherein the first cluster of inbred plant lines, the second cluster of inbred plant lines, or both comprise no more than 50% of the plants from the population.

16. The method of any one of claims 10-15, wherein the phenotypic profile is predicted using a machine learning model.

17. The method of any one of claims 10-16, wherein the phenotypic profile is predicted using best linear unbiased prediction (BLUP).

18. The method of any one of claims 10-17, wherein the one or more improved traits is selected from the group consisting of yield, lodging, disease resistance, maturity, herbicide tolerance, insect resistance, plant height, or any combination thereof.

19. The method of any one of claims 10-18, wherein the selected hybrid plants comprise at least a 5% increase in yield as compared to the elite hybrid variety.

20. A method of selecting a target plant, the method comprising: a. genotyping a first inbred parent and a second inbred parent of an elite hybrid to produce a first parent genotypic profile and a second parent genotypic profile; b. generating a genotypic profile for members of a population, the population comprising a plurality of inbred lines; c. clustering the members of the population of inbred lines to generate clusters of inbred lines having genetic similarity; d. selecting a first subset of inbred plant lines from a first cluster of inbred lines having genetic similarity to the first parent genomic profile and a second subset of inbred plant lines from the second cluster of inbred lines having genetic similarity to the second parent genomic profile; e. generating an in silico hybrid population comprising a plurality of simulated hybrids each simulated hybrid of the population having a parent from the first subset of inbred lines and a parent from the second subset of inbred lines; f. generating a predicted phenotypic profile for at least one trait for one or more plants of the in silico hybrid population; g. identifying hybrid plants of the in silico hybrid predicted to have improved phenotypic profile for the at least one trait as compared to the phenotypic profile of the elite hybrid; and h. growing the identified hybrid plants of the hybrid in silico population and selecting hybrid plants that have one or more improved agronomic traits as compared to the elite hybrid plant.

21. The method of claim 20, wherein the population comprises at least 10,000 inbred plant lines.

22. The method of claim 20 or 21, wherein the genotypic profile for members of the population is generated using observed SNP markers, imputed SNP markers, or a combination thereof.

23. The method of any one of claims 20-22, wherein the clustering comprises hierarchical clustering.

24. The method of any one of claims 20-23, wherein the population is separated into at least 20 total clusters.

25. The method of any one of claims 20-24, wherein the first cluster of inbred plant lines, the second cluster of inbred plant lines, or both comprise no more than 50% of the plants from the population.

26. The method of any one of claims 20-25, wherein the first subset of inbred lines comprises inbred lines having 95% or more genetic similarity to the first inbred parent, the second subset of inbred lines comprises inbred lines having 95% or more genetic similarity to the second inbred parent, or both the first subset of inbred lines and the second subset of inbred lines comprise inbred lines having 95% to the first parent and second parent, respectively.

27. The method of any one of claims 20-26, wherein the first subset of inbred lines, the second subset of inbred lines, or both the first and second subset of inbred lines comprise at least 50 inbred lines and fewer than 500 inbred lines.

28. The method of any one of claims 20-27, wherein the phenotypic profile is predicted using a machine learning model.

29. The method of any one of claims 20-28, wherein the phenotypic profile is predicted using best linear unbiased prediction (BLUP).

30. The method of any one of claims 20-29, wherein the one or more improved traits is selected from the group consisting of yield, lodging, disease resistance, maturity, herbicide tolerance, insect resistance, plant height, or any combination thereof.

31. The method of any one of claims 20-30, wherein the selected hybrid plants comprise at least a 5% increase in yield as compared to the elite hybrid variety.

Images