Identification and characterization of herbicides and plant growth regulators

Abstract

The present invention provides high-throughput methods that enable the screening of compounds for herbicidal activity or plant growth regulating activity and further allow the prediction of the mode of action of herbicide compounds or plant growth regulators. The methods provided herein also facilitate the identification of mutations responsible for herbicide resistance in plants and the identification of herbicide targets. The methods further provide the identification of mutations responsible for plant growth regulation and the identification of plant growth regulator targets.

Description

TECHNICAL FIELD

[0001] The present invention relates generally to the field of herbicides and plant growth regulators. More specifically, the present invention relates to methods of facilitating the discovery of herbicides and plant growth regulators, the prediction of herbicide and / or plant growth regulator mode of action, the identification of herbicide or plant growth regulator targets, and / or the identification of mutations conferring resistance to herbicides or plant growth regulators. BACKGROUND

[0002] The following discussion of the background of the application is merely provided to aid the reader in understanding the application and is not admitted to be prior art.

[0003] Prior to the introduction of herbicides, many farmers relied on manual and mechanical weed control measures.

[0004] Manual weed control methods are extremely labor intensive and have adverse social costs, while mechanical weed control measures, such as tilling, create environmental problems due to damage to soil structure and exposure to erosion, and often stimulate further weed invasion. Burning, if done prior to seed shed, can prevent further spread of weeds and can be done over a wide area with minimal labor input, but like tilling, also leaves the soil surface exposed to erosion.

[0005] Since the introduction of selective herbicides, approximately 70 years ago, crop protection chemicals such as herbicides have been an important tool for farmers. Non-selective herbicides (e.g., paraquat in the 1960s and glyphosate in the 1970s) have enabled the adoption of no-till (direct drilling; no- tillage planting) and other simplified crop production systems that are being established. If weeds are removed by herbicides prior to planting, there is no need to plow the land to bury the weeds. No-till systems can increase yields when the crop is properly established and has many environmental and economic benefits.

[0006] The potential benefits of no-till systems, in the case of herbicides only, include a reduction in the labor required, less soil erosion, water resource conservation, less fuel used, reduced greenhouse gas emissions, and an increase in biodiversity. Combating climate change is of particular current concern. Plowing lets air into the soil, causing oxidation of organic matter. This not only destroys good soil structure (e.g., building after rotational pastures), but also releases large amounts of carbon dioxide. Spraying herbicides to burn weeds prior to planting in no-till systems can reduce CO2 emissions by more than 80%.

[0007] The importance of herbicides globally, in 2018, they accounted for over 40% of the global crop and non-crop pesticide market, valued at $64 billion. However, since the late 1990s, new herbicides have reached the market at a much slower pace. Many of the largest selling and most widely used active ingredients (e.g., glyphosate) have been used by farmers for decades. In the past three decades, major herbicides with novel modes of action have not been commercialized. This poses a very serious problem for global crop protection as new herbicides that become available to farmers provide the same limited modes of action, which lead to resistant weeds.

[0008] Early herbicide discovery was achieved primarily by spraying chemicals on whole plants under greenhouse conditions and observing phenotypic changes after a few days or weeks. A more recent approach is to mimic the binding of a compound to a known or predicted herbicide target; this approach is called in silico screening. Another approach is to empirically test a screening compound for inhibition of activity of a known or predicted herbicide target isolated from its biological context; this approach is called in vitro screening. Another approach is to empirically test a screening compound for inhibition of growth or viability of a relevant living system; this approach is called in vivo screening. In the case of herbicide discovery, the target organism for in vivo screening is typically a whole weed plant. However, the quality of the output data produced by in vivo herbicide screening platforms, in which whole weed plants are exposed to screening compounds and scored for growth / viability, is obtained at a lower throughput than in vitro or in silico methods.

[0009] There is an unmet need for new herbicides and particularly those with novel modes of action. In addition to providing much-needed alternatives to existing herbicides with modes of action to which weeds have developed resistance, knowledge of the mode of action of a herbicide can also help to generally predict safety to humans, wildlife, and the environment. SUMMARY

[0010] The present invention satisfies at least one unmet need in the art by providing a method for screening candidate compounds for herbicidal or plant growth regulating activity.

[0011] Additionally or alternatively, the method can be used to identify the mode of action of a known or newly identified herbicide compound or plant growth regulator. The method can also be used to identify mutations that are the cause of certain phenotypes of plants, including, for example, herbicide resistance or plant growth regulation.

[0012] Output data quality and throughput are two key parameters of herbicide screening platforms, and achieving an effective trade-off between them is a significant challenge. The inventors have developed a highly efficient in vivo herbicide screening method with high throughput capabilities. This method utilizes a screening platform based on spores and / or germinating spores, such as those from non-vascular plants (e.g., bryophytes), to screen candidate compounds for their herbicidal activity. In addition to identifying novel compounds with herbicidal or plant growth-regulating activities, this method can be used to predict or determine the mode of action of newly identified and / or known herbicide compounds or plant growth regulators. Currently, the herbicide industry uses the dicotyledonous plant Arabidopsis thaliana (… Arabidopsis thaliana (This is used as a gold standard screening system.)

[0013] Previously, non-vascular plants such as *Millettia divaricata* (…) Marchantia polymorpha (This is considered a potential tool for gene screening, such as Ishizaki) et al. 2016 (Ishizaki, K., et al. The details are provided in "Molecular genetic tools and techniques for Marchantia polymorpha research", Plant and Cell Physiology, 2016, Vol. 57, No. 2, pp. 262-270; and for genetic models, see, for example, Sugano. et al. 2014 (Sugano, SS, et al., 'CRISPR / Cas9-mediated targeted mutagenesis in the liverwort Marchantia polymorpha L.', Plant and Cell Physiology. 2014, Vol. 55, No. 3, pp. 475-481). However, the use of non-vascular whole plants, spores, germinated spores, explants, protoplasts, or plant protoplasts in the context of screening herbicides or plant growth regulators is not obvious. Although studies on photosynthetic electron transport inhibitors have considered polymorphic Marchantia cells in a limited way (see Sato, F., et al."Photoautotrophic cultured plant cells: a novel system to survey new photosynthetic electron transport inhibitors", Zeitschrift fur Naturforschung C. 1991, Vol. 46, Issues 7-8, pp. 563-568), but there are different challenges and fundamental biological differences that mean non-vascular whole plants, spores, gemmae, explants, protoplasts and / or plant protoplasts would not be considered as a screening platform for herbicide compounds or plant growth regulators. For example, Marchantia polymorpha represents a much smaller and simpler plant system than the unwanted weed plant and indeed than the current gold standard screening system, Arabidopsis thaliana. Furthermore, non-vascular plants are structurally different in many ways to higher plants, for example the cuticle formed in gemmae, propagules, explants and whole non-vascular plants is chemically very different to higher plants. Furthermore, non-vascular plants such as Marchantia polymorpha are not a logical choice for screening herbicides or plant growth regulators as it is known that liverworts are insensitive to major herbicides such as glyphosate.

[0014] However, the present inventors have surprisingly found that gemmae of non-vascular plants, including liverworts, can be used to screen for herbicides or plant growth regulators. Furthermore, the use of non-vascular plants overcomes a number of limitations associated with the use of Arabidopsis thaliana and provides a significantly improved screening method. Given the large size and complex nature of Arabidopsis thaliana gemmae or whole plants, the throughput of the screen is limited by the time duration required to grow the plants and the associated additional space, resources and personnel required to cultivate the plants. For example, non-vascular plants typically only require 4 days of growth before they can be used for screening. This is in contrast to the typical value of 7 days for Arabidopsis thaliana. The present invention therefore provides a method with a throughput that is typically at least about 10 times higher than methods based on Arabidopsis. In addition to throughput limitations, the use of Arabidopsis gemmae or whole plants precludes methods for high content screening of herbicides or plant growth regulators based on fluorescence imaging. High content screening methods benefit from the fit of the imaged objects in the smallest possible spatial level volume, and the size and complexity of Arabidopsis prohibits their use in such methods. Given their small size, simple body development, sensitivity to herbicides with different modes of action and / or susceptibility to genetic manipulation, non-vascular plants are advantageous to use in the present invention. This results in a higher throughput, less expensive and more efficient screening system for candidate compounds such as herbicides and plant growth regulators, compared to the more complex screening systems currently in existence such as Arabidopsis.

[0015] In a first aspect, the present invention provides a method of screening candidate compounds for herbicidal or plant growth regulating activity, the method comprising the steps of:

[0016] (i) contacting a range of different candidate compounds with a plurality of test samples from non-vascular plants; and

[0017] (ii) determining whether the test samples provide a phenotypic response to the range of different candidate compounds by comparison with a control sample from a non-vascular plant that has not been contacted with a candidate compound;

[0018] wherein the test samples and the control sample comprise whole plants, spores, germinating spores, explants, protoplasts or vegetative propagules, and the phenotypic response is indicative of the herbicidal or plant growth regulating activity.

[0019] In one embodiment, the candidate compounds are candidate compounds for herbicidal activity.

[0020] In another embodiment, the non-vascular plants are mosses, hornworts or liverworts.

[0021] In another embodiment, the test samples and the control sample are germinating spores.

[0022] In a further embodiment, the test and control germinating spores are derived from spores of non-vascular plants of the same species.

[0023] In one embodiment, the test germinating spores are moss germinating spores, liverwort germinating spores, hornwort germinating spores or any combination thereof.

[0024] In another embodiment, each member of the range of different candidate compounds is contacted with a different test sample.

[0025] In another embodiment, a plurality of members of the range of different candidate compounds are contacted with a single test sample.

[0026] In yet another embodiment, the test samples and control sample are leafy liverwort germinating spores, simple thallus liverwort germinating spores, complex thallus liverwort germinating spores or any combination thereof.

[0027] In one embodiment, the test sample and control sample are selected from the group consisting of Marchantia alpestris gemmae, Marchantia aquatica gemmae, Marchantia berteroana gemmae, Marchantia carrii gemmae, Marchantia chenopoda gemmae, Marchantia debilis gemmae, Marchantia domingenis gemmae, Marchantia emarginata gemmae, Marchantia foliacia gemmae, Marchantia grossibarba gemmae, Marchantia inflexa gemmae, Marchantia linearis gemmae, Marchantia macropora gemmae, Marchantia novoguineensis gemmae, Marchantia paleacea gemmae, Marchantia palmata gemmae, Marchantia papillate gemmae, Marchantia pappeana gemmae, Marchantia polymorpha gemmae, Marchantia rubribarba gemmae, Marchantia solomonensis gemmae, Marchantia streimannii gemmae, Marchantia subgeminata gemmae, Marchantia vitiensis gemmae, Marchantia wallisii, Marchantia nepalensis, and combinations thereof.

[0028] In one embodiment, the plurality of test gemmae are disposed in a series of different wells, each well comprising: 400-800 gemmae / mL, 300-900 gemmae / mL, or 200-1000 gemmae / mL.

[0029] In another embodiment, the test sample and / or the control sample has been engineered to express a fluorescent molecule.

[0030] In one embodiment, the control sample is a positive control.

[0031] In another embodiment, the positive control is contacted with a known herbicide or plant growth regulator.

[0032] In one embodiment, the control sample is a negative control.

[0033] In another embodiment, the negative control sample is not contacted with a known herbicide or plant growth regulator.

[0034] In a further embodiment, step (ii) of the method comprises comparing the phenotype of the test sample to the phenotype of a positive control sample that was contacted with a known herbicide or plant growth regulator compound, and further comprises comparing the phenotype of the test sample to the phenotype of a negative control sample that was not contacted with a known herbicide or plant growth regulator compound.

[0035] In a further embodiment, the known herbicide compound has a known mode of action, and the comparison of the test sample phenotype to the positive control sample phenotype is used to predict the mode of action of candidate compounds identified as having herbicidal or plant growth regulating activity.

[0036] In another embodiment, the test sample, the negative control sample, and the positive control sample are germinating spores derived from spores of a non-vascular plant of the same species.

[0037] In another embodiment, the test sample, the negative control sample, and the positive control sample have been engineered to express a fluorescent molecule.

[0038] In one embodiment, step (ii) of the method comprises measuring the phenotypic response of the test sample after the contacting by growing the test sample in a suitable medium with the candidate compound under suitable conditions for a period of time between 1 and 3 days, between 1 and 5 days, between 3 and 6 days, between 3 and 5 days, between 2 and 3 days, between 1 and 10 days, less than 5 days, less than 4 days, or less than 3 days, and wherein the phenotypes of the control germinating spores are determined after growing the control germinating spores in the suitable medium under the suitable conditions for an equivalent period of time.

[0039] In one embodiment, step (ii) of the method comprises obtaining a measurement of any one or more of: sample length, sample width, sample shape, sample pigmentation, sample circularity, sample chlorophyll concentration, and / or cell number per sample.

[0040] In another embodiment, the measurements are digitally recorded.

[0041] In another embodiment, comparing the phenotypic response of the test sample to the phenotype of any of the control samples comprises any one or more of: distributed random neighborhood embedding, principal component analysis (generalized weighted least squares), principal component analysis (minimize weighted chi-square), principal component analysis (minimize residual), common factor analysis (principal axis), common factor analysis (maximum likelihood), or common factor analysis (weighted least squares).

[0042] In another embodiment, the candidate compounds are selected as potential herbicides using an artificial intelligence algorithm, such as a random forest algorithm or a neural network algorithm.

[0043] In one embodiment, step (ii) of the method comprises:

[0044] obtaining phenotypic measurements from the test sample and any of the control samples and thereby generating a dataset, and

[0045] using at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 95%, or 99% of the dataset as a training set for the artificial intelligence algorithm.

[0046] In one embodiment, the control samples comprise positive control samples, and the artificial intelligence algorithm is used to predict the mode of action of any of the candidate compounds.

[0047] In another embodiment, the method further comprises step (iii) of:

[0048] (a) contacting candidate compounds identified in steps (i) and (ii) as having herbicidal or plant growth regulating activity with a series of mutagenized samples comprising whole plants, spores, germinating spores, explants, protoplasts, or vegetative propagules, wherein the test and mutagenized samples are from non-vascular plants of the same species;

[0049] (b) extracting DNA from resistant mutagenized samples that survive the contacting in (a) or that do not exhibit abnormal growth following the contacting in (a);

[0050] (c) sequencing the genome or genomic portion of the resistant mutagenized samples, thereby obtaining mutagenized sample DNA sequences;

[0051] (d) aligning the mutagenized DNA sequences obtained in (c) to a reference DNA sequence and identifying a first set of sequence mismatches between the mutagenized sample DNA sequences and the reference DNA sequence;

[0052] (e) aligning the DNA sequence from the first comparison sample to the reference DNA sequence and identifying a second set of mismatches between the first comparison DNA sequence and the reference DNA sequence; and

[0053] (f) filtering the first set of mismatches against the second set of mismatches to identify a first subset of mismatches that is unique to the first set of mismatches, wherein the first subset of mismatches is a candidate mutation that is likely to confer resistance to a herbicide or a plant growth regulator;

[0054] wherein the first comparison sample is from an independent sample that did not survive contact with the candidate compound or exhibited abnormal growth after contact with the candidate compound, and belongs to the same genus as the resistant mutagenized sample, and wherein the reference DNA sequence is a known reference sequence for a plant of the genus.

[0055] In another embodiment, the method further comprises:

[0056] (e-i) aligning the DNA sequence of the second comparison sample to the reference DNA sequence and identifying a third set of mismatches between the second comparison sample and the reference DNA sequence; and

[0057] (f) filtering the first set of mismatches against the third set of mismatches to facilitate identification of a second subset of mismatches that is unique to the first set of mismatches, and generating a third subset of mismatches by filtering the first subset of mismatches against the second subset of mismatches, wherein the first and second subsets of mismatches are candidate mutations that are likely to confer resistance to a herbicide or resistance to a plant growth regulator;

[0058] wherein the second comparison sample is from an independent sample that did not survive contact with the candidate compound or exhibited abnormal growth after contact with the candidate compound, and belongs to the same genus as the mutagenized sample.

[0059] In another embodiment, the mutagenized sample is an Ml sample.

[0060] In another embodiment, the mutagenized sample comprises a non-naturally occurring mutation.

[0061] In another embodiment, the method does not comprise a step of isolating analysis, complex segregation analysis, or bulk segregation analysis.

[0062] In one embodiment, the alignment of (e) comprises aligning the DNA sequence of 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or more comparison samples to the reference DNA sequence and identifying the second set of sequence matches between the two sequences.

[0063] In another embodiment, the method further comprises filtering the candidate mutations with a biological filter.

[0064] In one embodiment, the mutagenized sample is haploid.

[0065] In one embodiment, the candidate mutation is in a gene encoding a protein targeted by the candidate compound identified as having herbicidal or plant growth regulating activity.

[0066] In another embodiment, step (iii) is computer implemented.

[0067] In another embodiment, the method further comprises identifying the plant molecule or biological pathway targeted by the candidate compound identified as having herbicidal activity by the method using any one or more of the following: enzyme assays, chlorophyll fluorescence kinetics assays, photosynthetic oxygen evolution assays, electrolyte leakage assays, radiometric assays, spectrophotometric assays, fluorescence assays, absorbance assays, colorimetric assays, mass spectrometry, mitotic index analysis, quantitative PCR analysis, transcriptomic profiling analysis, proteomic profiling analysis, whole genome analysis and / or quantitative quantitative trait locus analysis, computer modeling docking studies, chemical structure analysis.

[0068] In one embodiment, the plurality of test samples do not contain whole plants.

[0069] Definitions

[0070] Certain terms used herein should have the meanings set forth below.

[0071] As used in this application, the singular forms "a," "an," and "the" include plural referents unless the context clearly dictates otherwise. For example, as used herein, the term "a compound" also encompasses multiple compounds, unless otherwise indicated.

[0072] As used herein, the term "comprising" means "including." Variations of the word "comprising," such as "comprise" and "comprises," have correspondingly varied meanings. For example, a composition "comprising" material A can consist of only material A, or can include material A and any other number of additional components (e.g., material B and / or material C).

[0073] As used herein, the term "plurality" means more than one. In certain specific aspects or embodiments, plurality can mean 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 35, 40, 45, 50, 60, 70, 80, 90, 100, 250, 300, 350, 400, 450, 500, 600, 700, 800, 900, 1000, 1250, 1500, 1750, 2000, 2500, 3000, 3500, 4500, 5000, 6000, 7000, 8000, 9000, 10000, 15000, 20000, 30000, 40000, 50000, 60000, 70000, 80000, 90000, 100000, 200000, 300000, 400000, 500000, 600000, 700000, 800000, 900000, 1000000, or more, and any number derivable therein, and any range derivable therein.

[0074] As used herein, the term "between," when used in reference to a range of values, encompasses the values at each endpoint of the range.

[0075] As used herein, "high throughput screening" (HTS) refers to a method in which a plurality of synthetic compounds or natural products are screened for activity against one or more biological targets. HTS can be used to identify compounds having a range of biological activities, e.g., pharmaceutical, pesticide, herbicide, etc.

[0076] As used herein, the term "non-vascular plant" refers to a plant that lacks a vascular system (i.e., xylem and phloem). "Non-vascular plant" is understood herein to encompass whole non-vascular plants, components of whole non-vascular plants, spores of whole non-vascular plants, and germinating spores of whole non-vascular plants. Non-limiting examples of non-vascular plants include bryophytes such as mosses, liverworts, and hornworts.

[0077] As used herein, "whole plant" refers to a complete plant, particularly a complete non-vascular plant. Whole plant can also be used to refer to a miniaturized plant or a juvenile plant (i.e., a plant that has not reached its adult form and thus can not be fully mature or sexually mature). In the case of P. polymorpha, whole plant can refer to a gametophyte having a rooting system, one or more meristems, photosynthetic tissue, epidermis, cuticle, one or more gemmae cups, and one or more antheridial stalks (i.e., reproductive organs). For P. polymorpha juvenile plants or miniaturized plants, the plant has a subset of the tissues, organs, and characteristics of a whole plant, including at least a rooting structure, photosynthetic tissue, and one or more meristems.

[0078] As used herein, "weed" is understood to mean a plant that grows in adverse competition with a cultivated plant for any one or more of water, light, nutrients, and / or space. Non-limiting characteristics of a weed include low or no economic value compared to a cultivated plant, providing ecological and / or economic damage, a vigorous growth characteristic, and / or producing a large number of seeds.

[0079] As used herein, the terms "herbicide" and "herbicide compound" refer to a synthetic compound or natural product that is capable of killing or inhibiting the growth of a plant, plant cell, plant seed, or plant tissue, including but not limited to a weed and its seeds.

[0080] As used herein, the term "plant growth regulator" refers to a compound (natural or organic) that regulates plant growth, e.g., increases speed, promotes, accelerates, slows, inhibits, delays, or otherwise changes the plant growth, development, or maturation of a plant.

[0081] As used herein, the term "phenotype" shall be deemed to mean a set of observable characteristics of an individual (e.g., an individual plant or a portion of the plant). The term "phenotypic response" shall be deemed to mean an observable response of an individual (e.g., an individual plant or a portion of the plant) in response to a unique environment, e.g., in response to exposure to a candidate compound, such as a candidate herbicide or a candidate plant growth regulator.

[0082] As used herein, "reference DNA sequence" refers to a reference genome sequence of a given plant. The reference DNA sequence is publicly available, e.g., on a database or the like.

[0083] As used herein, "mismatch" refers to a difference in a portion of a read sequence (e.g., a portion of a DNA sequence of a plant being tested to identify a causal mutation) as compared to a portion of a reference DNA sequence to which the read optimally aligns.

[0084] As used herein, the term "causal mutation" shall be deemed to mean a mutation that causes or contributes to a phenotype of interest, e.g., a mutation that causes or contributes to herbicide resistance.

[0085] As used herein, "M0" denotes the population of plants in a mutagenesis experiment prior to exposure to a mutagen (i.e., the parental population). As used herein, "M1" is a notation referring to the same population of plants after exposure to a mutagen (i.e., the M0 population). As used herein, "M2" refers to the M1 progeny after selfing (i.e., the process of a mutant crossing with itself).

[0086] As used herein, "segregation analysis" refers to a statistical technique used to fit formal genetic models to data on traits or diseases (phenotypes) expressed in members of a biological family to determine the most likely genetic mode of the trait or disease under study. Segregation analysis requires multiple generations of family members to determine the genetic pattern of the phenotype under analysis.

[0087] Any description of prior art documents herein, or statements herein that are derived from or based on those documents, is not an admission that the documents or derived statements are part of the common general knowledge of the relevant art.

[0088] For the purposes of description, all documents herein, unless otherwise indicated, are incorporated by reference in their entirety. BRIEF DESCRIPTION OF DRAWINGS

[0089] Preferred embodiments of the present application will now be described, by way of example only, with reference to the accompanying drawings as set out below:

[0090] Figure 1 is a far-red fluorescence micrograph showing 50 P. multiflora gemmae growing in one well of a 96-well assay plate. Gemma density was adjusted to include the maximum number of gemmae and to minimize the number of gemmae growing in contact with each other. The former increases the statistical significance of the assay, while the latter is required to generate accurate data describing the response of the whole plant to herbicide treatment.

[0091] Figure 2 Associated 10X micrographs of 5-day-old P. multiflora gemmae are provided. In the left image, transmitted light is recorded (brightfield image). In the middle image, only far-red photons are recorded by the camera chip (far-red image). In the right image, only cyan photons are recorded (cyan image). The images are cropped to show a single gemma; the same P. multiflora gemma is shown in all three images. The images were generated using third-party software, FIJI.

[0092] Figure 3 Associated 10X micrographs of 5-day-old gemmae are provided. Left, far-red micrograph overlaid with the outline of the object segmented using brightfield image (purple line), far-red image (red line), and cyan image (blue line). Middle, cyan image. Right, brightfield image. The images were generated using third-party software, GE Developer.

[0093] Figure 4 : A 1-day-old Marchantia polymorpha pro Mp EF1a::YFP-NLS(3X) 10X micrographs of germinating spores. Left, transmitted light micrograph showing ungerminated spores (arrow) and 2-cell germinating spores. Middle, fluorescence micrograph showing the two nuclei of a 2-cell germinating spore. Images were generated using third-party software FIJI. Right, graph providing outlines of nuclei segmented using third-party software GE Developer.

[0094] Figure 5 Provided are the distribution of the 10 variables from the low-resolution dataset (2X) before scaling (top panel) and after scaling (bottom panel).

[0095] Figure 6 Is a scree plot of the factors created from a set of 10 variables in the low-resolution dataset (2X). The horizontal line represents the recommended number of factors formed according to the Kaiser criterion, the Elbow method, or the Joliffe criterion. Images were generated using third-party software StratomineR.

[0096] Figure 7 Is a hit selection plot showing the average phenotypic distance of plants in wells containing 0.1% DMSO (negative control, red), a low concentration of the known herbicide isoxaben (positive control, green), or a chemically unknown active (screening compound, blue). The red dashed line represents the threshold above which a compound causes a statistically significant response in treated plants and is selected as a hit. Images were generated using third-party software StratomineR.

[0097] Figure 8 Is a hit selection plot showing the average phenotypic distance of plants in wells containing 0.1% DMSO (negative control, red), a low concentration of the known herbicide isoxaben (positive control, green), or a chemically unknown active (screening compound, blue). The red dashed line represents the threshold above which a compound causes a statistically significant response in treated plants and is selected as a hit. Images were generated using third-party software StratomineR.

[0098] Figure 9 Provided are contour plots showing the predicted coordinates of the phenotypic response of plants treated by 0.1% DMSO (red surface) or by 3 different herbicides with different modes of action. Circles represent a 20% of the dataset used to test the phenotypic models: if a circle lies on a surface of the same color, the corresponding model is accurate. Images were generated using third-party software StratomineR.

[0099] Figure 10A cluster plot showing hits belonging to the same cluster in different colors is provided.

[0100] Figure 11 A photograph of a petri dish containing 14-day-old germinated spores grown on 1 µM, 100 µM, or 1000 µM of the known herbicide, bentazon, is provided. The germinated spores did not all die on 1000 µM bentazon. Three replicates are shown.

[0101] Figure 12 A photograph of a petri dish containing 14-day-old germinated spores grown on 1 µM, 100 µM, or 1000 µM of the known herbicide, chlorimuron, is provided. The germinated spores did not all die on 1000 µM chlorimuron, but showed a clear reduction in growth. Three replicates are shown.

[0102] Figure 13 A dose response curve for the genus Marchantia mutagenized by different amounts of UV-B radiation. Figure 13 Two replicates are provided. 50% kill was achieved by spores exposed to UV-B radiation for 20 s.

[0103] Figure 14 Dark field micrographs of 14-day-old germinated spores grown on 0.1 ppm chlorimuron (left and middle) with no UV-B treatment (left) or with UV-B treatment. The larger plant is a chlorimuron resistant mutant; it has a similar phenotype to the UV-B treated 14-day-old germinated spores shown in the right panel grown on 0.1% DMSO.

[0104] Figure 15 A process flow diagram for a method for identifying a causal mutation that causes a phenotype of interest in a test sample according to an embodiment of the present invention.

[0105] Figure 16 : Rhizoid phenotypes of two-day-old Marchantia polymorpha plants. Wild type rhizoid phenotype (A), wavy rhizoid phenotype (B). Rhizoids are cells that grow straight in the wild type (A) and wavy in some mutants (B).

[0106] Figure 17 : Dorsal epidermis phenotypes of two-month-old Marchantia polymorpha plants. Wild type epidermis phenotype (A), stretched epidermis phenotype (B). The dorsal epidermis shows stomata (A, arrow mark) that are stretched in some mutants (B).

[0107] Figure 18 : Performance of the non-allelic based mutation discovery pipeline in UV4.32. A: Effect of the increase in the number of non-allelic mutation backgrounds on the filtering efficiency. B: Number of UV4.32 mismatches remaining after each filtering step when using 8 non-allelic UV mutant lines.

[0108] Figure 19 : Performance of the non-allelic based mutation discovery pipeline in chlorimuron-resistant mutants. Increasing the number of allelic mutant backgrounds improves filtering efficiency. The leftmost box plot represents the total number of mismatches in chlorimuron-resistant mutant lines before filtering out the mismatches observed in the wild-type genome that were also resequenced.

[0109] Figure 20 Four associated 4X micrographs of 5-day-old germinated spores are provided. The top panel shows plants exposed to 0.1% DMSO, the middle panel shows plants exposed to 10 uM isoxaben and the bottom panel shows plants exposed to 10 uM of the test compound selected as a hit. From left to right: far-red micrograph overlaid with the outline of the object using brightfield image segmentation (outer red line) and the outline of the plant body using far-red image segmentation (nested red line); next, far-red micrograph overlaid with the outline of the object using brightfield image segmentation (outer red line) and the outline of the plant meristem using far-red image segmentation (nested purple line); next, far-red micrograph overlaid with the outline of the object using brightfield image segmentation (outer red line) and the outline of the plant rhizoid using far-red image segmentation (nested green line); next, cyan micrograph overlaid with the outline of the object using brightfield image segmentation (outer red line) and the outline of the cyan fluorescent material using cyan image segmentation (nested yellow line); finally, far-red micrograph overlaid with the outline of the object using brightfield image segmentation (outer red line) and the outline of the chloroplasts using far-red image segmentation (nested blue line). DETAILED DESCRIPTION

[0110] Herbicide resistance is a major problem affecting global crop and pasture production. The level of herbicide-resistant weeds is only expected to increase until herbicides exhibiting different modes of action are identified and commercialized. Methods that enable the identification of novel compounds with herbicidal or plant growth-regulating activity and the identification of the mode of action of such compounds are of great importance to combat the increasing level of herbicide resistance in weeds and other similar plants.

[0111] The present invention provides high throughput methods that enable the screening of compounds for herbicidal or plant growth regulating activity, thereby providing the potential for the discovery of novel herbicides or plant growth regulators. The methods further allow for the prediction of the mode of action of herbicide compounds or plant growth regulators, including compounds identified by the methods described herein and any known herbicide or plant growth regulator that can not have a characterized mode of action. Thus, the methods provided herein can be used to identify compounds, e.g., herbicide compounds or plant growth regulators, that have novel modes of action. The methods provided herein also provide for the identification of mutations responsible for plant, e.g., weed, resistance to herbicides and the identification of herbicide targets. The methods provided herein also provide for the identification of mutations responsible for plant growth regulation and the identification of plant growth regulator targets.

[0112] Currently, screening compounds for herbicidal activity using whole plants is relatively low throughput because the size and complexity of whole plants does not make them amenable to high throughput phenotypic characterization. Described herein is a high throughput screening method using non-vascular plants that enables the simultaneous screening of a large number of compounds for herbicide or plant growth regulator activity, thereby predicting their mode of action and identifying their targets. The methods described herein can also be used to identify causal mutations that confer resistance to herbicide compounds.

[0113] High-throughput screening

[0114] The present invention provides in vivo high throughput methods for screening compounds for herbicidal or plant growth regulating activity. High throughput screening (HTS) is a technique used to rapidly and efficiently sort useful compounds from a large number of candidates for novel drugs, pesticides, herbicides, etc. and can be used to identify compounds with a range of biological activities. Without limitation, HTS as contemplated herein generally includes three elements: a suitable library of compounds for screening, a screening method, and a system for processing and / or analyzing data generated by the assay.

[0115] Compounds for use in the methods of the present invention can be created from combinatorial chemistry or from natural products (e.g., secondary metabolites from plants, animals, and / or microorganisms). In some embodiments of the present invention, natural compounds can be treated prior to inclusion in a library of compounds. A non-limiting example of a suitable treatment technique well known to those skilled in the art is solid phase extraction. In additional embodiments of the present invention, combinatorial libraries to be screened can be synthesized in the compartment in which the assay is performed, thereby providing a reference address for the candidate compound. A range of concentrations of any given compound can be tested. Various solvents can be used to dissolve solid compounds. Any suitable compound can be screened using the methods of the present invention, including but not limited to candidate natural, synthetic, and chemical compounds.

[0116] According to the method of the present invention, non-vascular plants can be contacted with candidate compounds, namely candidate herbicide compounds or candidate plant growth regulator compounds, and their responses to these candidate compounds can be evaluated. As indicated in the definition section, the term "non-vascular plant" includes the whole non-vascular plant, its components, its spores and / or its germinated spores. Plants can be non-vascular plants, such as, for example, mosses, bryophytes, and / or hornworts.

[0117] By way of non-limiting examples, non-vascular plants can be bryophytes. Bryophytes can be thalloid, simple thalloid, or complex thalloid. Non-limiting examples of bryophytes that can be used in the screening method described herein include: *Mallotus altissima*, *Mallotus aquaticus*, *Mallotus peltata*, *Mallotus calceus*, *Mallotus quinquefolius*, *Mallotus licheniformis*, *Mallotus domingosus*, *Mallotus styracifoli ... and *Mallotus styracifolius*. Other non-limiting examples of bryophytes that can be used in the screening methods described herein are Jungermanniopsida (e.g., plants of the subclasses Jungermanniidae or Metzgeriidae), Marchantiopsida (e.g., plants of the subclasses Marchantiidae or Sphaerocarpidae), or Haplomitriopsida.

[0118] By way of non-limiting examples, non-vascular plants can be mosses. Non-limiting examples of mosses that can be used in the screening method described herein are as follows: *Mammillaria pulcherrima* (… Physcomitrella patens ) or Red's small bowl moss ( Physcomitrella readeri Mosses.

[0119] By way of non-limiting examples, non-vascular plants can be hornworts. Non-limiting examples of hornworts that can be used in the screening method described herein are as follows: hornworts of the genera *Anthoceros*, *Dendroceros*, *Folioceros*, *Megaceros*, *Notothylas*, and *Phaeoceros*.

[0120] In some embodiments of the present invention, non-vascular plants may be insensitive to glyphosate and / or glufosinate.

[0121] The methods of the application use plant material from non-vascular plants. The methods can include the steps of (i) contacting a plurality of test samples with a series of different candidate compounds; and (ii) determining whether the test samples provide a phenotypic response to the series of different candidate compounds by comparison to the phenotype of control samples that were not contacted with the candidate compounds. The test samples and control samples can include non-vascular plant material. In one embodiment, the test samples and control samples include whole plants, spores, gemmipores, explants, protoplasts, or vegetative propagules from non-vascular plants. In one embodiment, the test samples and control samples include spores, gemmipores, explants, protoplasts, or vegetative propagules from non-vascular plants. In one embodiment, the test samples and control samples include spores or gemmipores from non-vascular plants. In one embodiment, the test samples and control samples include spores from non-vascular plants. In a preferred embodiment, the test samples and control samples include spores from a bryophyte. Suitable bryophyte spores or gemmipores for use in the methods of the application include, for example, Marchantia spores or gemmipores.

[0122] Gemmipores for use in the methods of the application can have rhizoids, plant photosynthetic cells, and / or nascent meristems. Plants can be used for high-throughput screening at less than 1 day old, less than 2 days old, less than 3 days old, less than 4 days old, less than 5 days old, less than 6 days old, less than 7 days old, less than 8 days old, less than 9 days old, less than 10 days old, less than 11 days old, less than 12 days old, less than 13 days old, or less than 14 days old. Alternatively, older plants can be used.

[0123] Non-vascular plants for use in the methods described herein can autofluoresce (i.e., contain endogenous fluorescent molecules). The nature of the autofluorescence can indicate that the non-vascular plant contains photosynthetic pigments, photoprotective pigments, stress-induced primary metabolites, stress-induced secondary metabolites. For example, chlorophyll is a photosynthetic pigment located in the chloroplasts of non-vascular plants and fluoresces in the "far-red" spectrum. For example, NADH and NADPH accumulate in stressed non-vascular plants and fluoresce in the "cyan" spectrum. The amplitude of the autofluorescence can indicate the extent to which the fluorescent molecules accumulate in the non-vascular plant. By extension, the nature and amplitude of the autofluorescence can indicate various physiological responses of the non-vascular plant to contact with a test compound. For example, chlorophyll content can indicate plant growth, and NAD(P)H content can indicate chemically-induced cellular stress. As another example, the localization of chlorophyll and the localization of NAD(P)H in a cell or plant can indicate impaired light and metabolic processes by contact with a test compound.

[0124] A non-vascular plant used in the methods described herein can be engineered to express a fluorescent or luminescent cell marker. Expression of these fluorescent or luminescent markers can be used to create digital images for use in obtaining a measure of phenotypic response. It has been common in the art for some time to create images of biological structures using expression of fluorescent cell markers. Various techniques exist to express fluorescent proteins in plant cells. Standard plant transformation methods known to those skilled in the art can be used to introduce an exogenous nucleic acid encoding a fluorescent protein into a plant cell. One common method that can be used in some embodiments of the present application is Agrobacterium tumefaciens transfer-DNA (T-DNA) induced insertion mutagenesis. Simple and efficient T-DNA transformation protocols have been available to those skilled in the art for many years, including for example the “dabbing” method (Clough and Bent, The Plant Journal, 1998; 16(6): 735-743). Those skilled in the art will know that T-DNA transformation protocols can be used in non-vascular plants (reviewed in Genetic transformation of moss plant, Jing et al. , 2013, African Journal of Biotechnology; 12(3): 227-232). T-DNA mediated insertion is random, but because the inserted DNA fragment is flanked by 25 bp border sequences (T-DNA), primers designed from the left border of the T-DNA can be used to isolate the genomic / T-DNA sequence junction, which can then be mapped to the genome to precisely identify the chromosomal insertion location. Transposon-mediated mutagenesis (with or without T-DNA) is commonly used in the art and can be used to express fluorescent proteins in plants used in the present application.

[0125] Other common methods for introducing exogenous nucleic acids into plant cells that can be used in the present application include, but are not limited to, cationic or polyethylene glycol treatment of protoplasts (O'Neill et al. The Plant Journal, 1993; 3(5):729-738), calcium phosphate precipitation, electroporation, microinjection, viral infection, protoplast fusion, microprojectile bombardment, agitating a cell suspension in a solution with microbeads or microparticles coated with transforming DNA, direct DNA uptake, and liposome-mediated DNA uptake. Such methods are well described in a wide range of texts commonly used by those skilled in the art, such as Glick, Methods in Plant Molecular Biology and Biotechnology, 2018; CRC Press; Sambrooket al. Molecular Cloning: a laboratory manual, 1998; Cold Spring Harbor Laboratory. CRISPR / Cas9 genome editing techniques can also be used to fluorescently label endogenous plant proteins.

[0126] A wide variety of fluorescent proteins are commercially available (see, e.g., Shaner et al. Nature Methods, 2015; 2(12): 905-909), and one skilled in the art can select a marker for use in the present application based on the desired image. Numerous publications are available to describe the use of fluorescent proteins in imaging of various plant types, plant organs, and use with a range of microscopy techniques (see, e.g., Berg and Beachy, Methods in Cell Biology, 2005; 85: 153-177).

[0127] In some embodiments of the present application, green fluorescent protein (GFP) or a modified version of GFP can be used as a fluorescent marker. The original GFP isolated from the jellyfish Aequorea victoria as well as numerous modifications to wild-type GFP to enable its expression in plants are known in the art. GFP spectral mutants such as cyan and yellow enhanced fluorescent proteins (ECFP and EYFP) are generally classified into seven types based on the type of chromophore (see Zacharias and Tsien, Green Fluorescent Protein: Properties, Applications, and Protocols, 2006, John Wiley and Sons; 83-120). The choice of fluorophore depends on whether more than one fluorescent marker is to be used, which requires selection of pairs that can be spectrally separated. In some embodiments of the present application, red fluorescent protein (RFP) or a modified version of RFP can be used as a fluorescent marker. Plants can be engineered to express fluorescent proteins, e.g., RFP and derivatives, in specific plant structures.

[0128] The non-vascular plants used in the methods described herein can be stained with fluorescent cell dyes or probes. The fluorescent cell dyes or probes can be applied at any time during the assay to label any particular cellular compartment, any particular cell type, or any particular portion of the plant. In some embodiments of the application, such fluorescent dyes include Calcofluor White, S4B, propidium iodide, FM1-43, FM4-64, Mitotracker dyes, or Hoechst dyes. Fluorescent cell dyes or probes can also be used as ion content indicators, including [Ca2+] or pH indicators or redox indicators. In some embodiments of the application, such fluorescent dyes or probes are OxiORANGE, HySOx, HYDROP, hydroxyl phenyl fluorescein.

[0129] The screening methods of the application can use any suitable arrangement of compartments (e.g., wells, tubes, etc.) suitable for HTS assays.

[0130] For example, a 96-well microtiter plate can be used for HTS in both automated and non-automated formats. The assay can be set up manually or by robotic systems such as liquid handling robots.

[0131] Different candidate compounds can be individually screened in separate compartments for herbicidal activity or plant growth regulating activity. A given candidate compound can be screened in a single compartment or in multiple compartments. Alternatively, multiple different candidate compounds (e.g., a natural extract library or a synthetic mixture of molecules) can be screened in a single compartment for herbicidal activity or plant growth regulating activity.

[0132] In some embodiments of the application, in preparation for screening, a solvent can be mixed with the non-vascular plant and the candidate compound. Non-limiting examples of suitable solvents include dimethyl sulfoxide (DMSO), acetone, water, methanol, and ethanol. DMSO is a carrier / general solvent that has the ability to solubilize a large number of small molecules and carry them across membranes. Without wishing to be bound by theory, DMSO or another suitable solvent, a surfactant, and any other suitable additives can also enhance the permeability of the plant cells / tissues by the test compound and aid in the preservation of the plant cells / tissues during the assay.

[0133] Additionally or alternatively, in a preparation for screening, a liquid or jellied nutrient medium can be mixed with non-vascular plants and candidate compounds. Non-limiting examples of suitable nutrient media include Johnson's medium, M51C, Gamborg B5, and MS medium. In some embodiments, Johnson's medium comprising the following is utilized: myo-inositol (100 mg / L), sucrose (10 g / L), KNO3(6000 µM), MgSO4(1000 µM), Ca(NO3)2*4H2O (4000 µM), KCl (25 µM), H3BO4(10 µM), MnSO4*4H2O (1 µM), ZnSO4*7H2O (1 µM), CuSO4*5H2O (0.25 µM), (NH4)6Mo7O24*4H2O (0.25 µM), FeSO4*7H2O (25 µM), Na2EDTA (25 µM), NH4H2PO4(600 µM), and (NH4)2SO4(400 µM).

[0134] One skilled in the art will appreciate that the density of non-vascular plants per well of a given assay plate can be varied to maximize the statistical significance of the assay while avoiding the stacking of material within each well that can affect the accuracy of the measurements. One exemplary embodiment of the present invention uses 50-70 spores or gemmae per well of a standard 96-well microtiter plate. Alternatively, the density can be 40-80, 30-90, or 20-100 plants, spores, or gemmae per well. In some embodiments, the density is 100-225 spores or gemmae per cm 2 , 85-260 spores or gemmae per cm 2 , 55-285 spores or gemmae per cm 2 One skilled in the art will appreciate that the density of non-vascular plants per well of a given assay plate can be increased to saturation when the material can be stacked without affecting the accuracy of other measurements. For example, the saturation density can be 11400-17100 spores or gemmae per cm 2 , 8550-19950 spores or gemmae per cm 2 , 5700-22800 spores or gemmae per cm2or 285-28500 spores or gemmae per cm 2Such other measurements include, but are not limited to, spectrophotometric measurements or fluorescence measurements of autofluorescence of spore or germtube suspensions, or fluorescence of fluorescent proteins or fluorescence of fluorescent dyes or probes in spore or germtube suspensions. During the assay, the non-vascular plant can be exposed to light. In some embodiments, the non-vascular plant can be grown under continuous illumination. Alternatively, exposure to light can be interrupted during the assay. Illumination can be provided at a wavelength between 300 nm and 900 nm, e.g., 400 nm and 700 nm. Illumination can be, for example, ultraviolet (UV) light, visible light, or infrared (IR) light.

[0135] The temperature at which the non-vascular plant is grown during the assay can be, for example, less than 15°C, less than 16°C, less than 17°C, less than 18°C, less than 19°C, less than 20°C, less than 21°C, less than 22°C, less than 23°C, less than 24°C, less than 25°C, less than 26°C, less than 27°C, less than 28°C, less than 29°C, or less than 30°C. In some embodiments, the temperature is between 21°C and 24°C.

[0136] The humidity at which the non-vascular plant is grown during the assay can be, for example, in the range of 40-80%, 45-75%, or 50-60%.

[0137] The duration of time for which the non-vascular plant is grown in the assay can be, for example, less than 1 day, less than 2 days, less than 3 days, less than 4 days, less than 5 days, less than 6 days, less than 7 days, less than 8 days, less than 9 days, less than 10 days, less than 11 days, less than 12 days, less than 13 days, less than 14 days, less than 15 days, less than 16 days, less than 17 days, less than 18 days, less than 19 days, less than 20 days, less than 21 days, less than 22 days, less than 23 days, less than 24 days, less than 25 days, less than 26 days, less than 27 days, or less than 28 days. Alternatively, older plants can be used in the assay.

[0138] In one exemplary embodiment of the application, the plant germtubes (e.g., Marchantia germtubes) are grown under continuous illumination at about 23°C for about 5 days prior to taking the measurements.

[0139] At the end of the assay period and / or upon completion, appropriate comparisons can be made between test samples in which the non-vascular plant was treated with various candidate compounds and control samples in which the non-vascular plant was not treated with various candidate compounds. The control sample can be a negative control sample in which the non-vascular plant was not mixed with a herbicide or plant growth regulator and / or the control sample can be a positive control sample in which the non-vascular plant was mixed with a known herbicide or plant growth regulator (e.g., a herbicide or plant growth regulator with a known mode of action). These comparisons can be used to determine factors including, but not limited to, whether a given candidate compound or mixture of candidate compounds has herbicidal activity, the potency of any herbicidal activity observed, the phenotypic response of the non-vascular plant in response to a given candidate compound found to exhibit herbicidal activity, and / or the predicted mode of action of a given candidate compound found to exhibit herbicidal activity. Similarly, these factors can be determined with respect to plant growth regulators. Any suitable means known in the art for making comparisons between test samples, negative controls, and / or positive control samples can be used. For example, comparisons can be made via visual comparison, via imaging under a microscope (e.g., a fluorescence microscope), etc. In one embodiment, the method screens for candidates with herbicidal activity, and the phenotypic response is death of the non-vascular plant following exposure to the candidate compound (i.e., plant matter dies following exposure to the candidate compound). In one embodiment, death is determined at one week following exposure. In one embodiment, death is determined at two weeks following exposure. In one embodiment, death is determined at three weeks following exposure.

[0140] In one embodiment, the method screens for candidates with plant growth regulator activity, and the phenotypic response is growth of the non-vascular plant following exposure to the candidate compound (i.e., plant matter grows following exposure to the candidate compound). In one embodiment, growth is determined at one week following exposure. In one embodiment, growth is determined at two weeks following exposure. In one embodiment, growth is determined at three weeks following exposure. Those skilled in the art will appreciate that there are many ways to determine plant growth, including, but not limited to, measuring plant size (diameter), density, width, diameter, or height. More complex analysis of the phenotypic response can be made using high content screening as described below.

[0141] In one embodiment, a method of screening candidate compounds for herbicidal activity or plant growth regulating activity is provided, the method comprising the steps of:

[0142] (i) contacting a plurality of test samples with a series of different candidate compounds; and

[0143] (ii) determining whether the test sample provides a phenotypic response to the series of different candidate compounds by comparison with the phenotype of a control sample that has not been contacted with a candidate compound;

[0144] wherein the test sample and the control sample comprise a whole plant, spore, gemmiparous spore, explant, protoplast or vegetative propagule from a non-vascular plant and the phenotypic response is indicative of herbicidal activity or the plant growth regulating activity. In preferred embodiments, the non-vascular plant is a liverwort, most preferably of the genus Marchantia.

[0145] In one embodiment, there is provided a method of screening candidate compounds for herbicidal activity or plant growth regulating activity, the method comprising the steps of:

[0146] (i) contacting a series of different candidate compounds with a plurality of test samples; and

[0147] (ii) determining whether the test sample provides a phenotypic response to the series of different candidate compounds by comparison with the phenotype of a control sample that has not been contacted with a candidate compound;

[0148] wherein the test sample and the control sample comprise a whole plant, spore, gemmiparous spore, explant, protoplast or vegetative propagule from a liverwort plant and the phenotypic response is indicative of herbicidal activity or the plant growth regulating activity.

[0149] In one embodiment, there is provided a method of screening candidate compounds for herbicidal activity or plant growth regulating activity, the method comprising the steps of:

[0150] (i) contacting a series of different candidate compounds with a plurality of test samples; and

[0151] (ii) determining whether the test sample provides a phenotypic response to the series of different candidate compounds by comparison with the phenotype of a control sample that has not been contacted with a candidate compound;

[0152] wherein the test sample and the control sample comprise a whole plant, spore, gemmiparous spore, explant, protoplast or vegetative propagule from a liverwort plant and the phenotypic response is indicative of herbicidal activity or the plant growth regulating activity.

[0153] Non-limiting methods are discussed below and are also presented in the Examples of the present application.

[0154] High-content screening

[0155] In some embodiments, the methods of the present application employ high content screening (HCS). Generally, HCS uses fluorescent or light emission measurements of samples in a high-throughput format and quantitative analysis of various parameters.

[0156] Non-limiting examples of parameters that can be used in HCS according to the methods described herein include non-vascular plant length, width, shape, pigmentation, circularity, chlorophyll content, and cell number per non-vascular plant (note that, as set forth above, as used herein, "non-vascular plant" encompasses whole non-vascular plants, components thereof, spores thereof, and germinated spores thereof). Any one or more of these parameters, optionally including other parameters, include a phenotypic response of the plant to the compound.

[0157] Imaging can be performed by a variety of techniques. Those skilled in the art will appreciate that HCS is typically performed using a fully automated fluorescent imaging system. In some embodiments of the present application, a liquid handling robot is incorporated into the fully automated fluorescent imaging system. In other embodiments, for high-throughput imaging, assays can be manually set up prior to imaging with a fully automated fluorescent imaging system. The fully automated fluorescent imaging system can include a high-throughput fluorescent microscope. Some embodiments of the present application do not require the use of a confocal microscope.

[0158] Several images can be generated for each non-vascular plant sample utilized in a given assay. In some embodiments of the present application, images can be generated by recording transmitted light. Additionally or alternatively, images can be generated by recording transmitted light. Images can be created using only far-red photons and / or only using cyan photons. Far-red images or yellow images can be produced using a 2X objective lens, where the field of view covers one full well of a 96-well microtiter plate. The image can be a far-red fluorescent micrograph. The image can be a yellow fluorescent micrograph. A 4X objective lens can produce far-red images, cyan images, yellow images, or brightfield images, where the field of view fits just within the boundaries of one well of a 96-well microtiter plate. A 10X objective lens can be used to produce a set of 1-9 far-red images, 1-9 cyan images, 1-9 yellow images, and 1-9 brightfield images that cover a small portion of one well, for example 1 / 32-1 / 3 of the well bottom surface. In some embodiments of the present application, images are superimposed with the outline of the object created by another image. Images can be superimposed with the outline created by an image using the same or different photons. By using objective lenses with 2x, 4X, 10X, 20X, or 40X optical magnification, images can have a field of view that is larger, equal, or smaller than the diameter of the well. In some embodiments of the present application, image analysis protocols are used that distinguish plants from subcellular objects such as nuclei from the background.

[0159] In some embodiments of the application, several images are generated at different points through the structure of a non-vascular plant. These images are referred to as slices. The slices can be less than 5 pm, less than 10 pm, less than 20 pm, less than 30 pm, less than 40 pm, less than 50 pm, less than 60 pm, 70 pm, less than 80 pm, less than 90 pm, or 100 pm deep. The skilled person can determine the number of slices required to image the entire structure based on the sample thickness. Manual imaging can also be used to enable the skilled person to develop an automated protocol. For display and / or analysis purposes, spatially and / or temporally related images can be combined to form a "stack". For example, in some embodiments of the application, a far-red image can be created that is a maximum intensity projection of a stack of 5 slices, each 20 pm deep. The stack can be created with 2, 3, 4, 5, 6, 7, 8, 9, 10, or more images. The stack can also be created with fewer than 20, fewer than 30, fewer than 40, fewer than 50, fewer than 60, fewer than 70, fewer than 80, fewer than 90, fewer than 100 images. The images can be stored in a digital format. Any or all of the images can be used in subsequent analysis. Some embodiments of the application use a computer script to add metadata to the images. The metadata can include the barcode of the assay plate, the date, the image acquisition protocol, and / or the image analysis protocol. The computer script can record whether the images are associated with other images.

[0160] Phenotypic "fingerprinting" of the response of a living system to a chemical is widely used in the field of drug discovery (see, for example, Reisen et al. , Assay and Drug Development Technologies, 2015; 13(7): 415-427) and software tools to assist the skilled person are widely available (see, for example, Omta et al. , Assay and Drug Development Technologies, 2016; 14(8): 439-452).

[0161] The phenotypic fingerprint can be consistent with the known effect of the compound. For example, a compound that inhibits pigment synthesis can cause the contacted plant to produce less pigment and grow smaller; thus, the fingerprint can be a quantitative representation of this expected plant response as described by several phenotypic variables. The phenotypic fingerprint can also be unexpected. For example, a compound that inhibits photosynthesis can cause cell elongation and a shift in the cellular localization of chloroplasts; thus, the fingerprint can also be a quantitative representation of a surprising plant response as described by several phenotypic variables.

[0162] In some embodiments of the application, data is normalized by the median of the negative controls at the plate level to minimize noise across the plate. In further embodiments, the normality of the distribution of the variables is checked and transformed if necessary. Transformation of data can be recommended by the software used for the analysis, which can also recommend data transformation methods and / or automatically perform the transformation. Non-limiting examples of data transformation methods that can be adapted for use with the present application include: square root, power of 2, power of 3, log, log2, loglO, reciprocal. Variables can then be scaled to equalize the weight of the variables in different means in downstream analysis steps. In some embodiments of the application, scaling is performed at the plate level using Z-score method. Screening level scaling can also be performed at this step.

[0163] The statistical methods used for analyzing the data are not particularly limited. Many software packages are suitable for the methods of the present application, which provide a data analysis pipeline where each step can be customized by the person skilled in the art by changing the statistical methods and parameters used. In some embodiments of the application, the relevant variables in the negative control and screening compound data are transformed into factors by: Distributed Random Neighbour Embedding, Principal Component Analysis (Generalized Weighted Least Squares), Principal Component Analysis (Minimized Weighted Chi-Squared), Principal Component Analysis (Least Residual), Common Factor Analysis (Principal Axis), Common Factor Analysis (Maximum Likelihood), or Common Factor Analysis (Weighted Least Squares). In further embodiments, an oblimin (non-orthogonal) rotation can be employed to reduce the number of factors. The number of factors retained can be determined automatically according to the Kaiser, Elbow or Joliffe criteria. Additionally or alternatively, the factors to be retained can be selected manually upon visual inspection of the scree plot.

[0164] Alternatively, all or a subset of the original variables can be selected that describe the phenotypic differences between plants treated with the candidate herbicide and the negative or positive controls.

[0165] The retained factors can be used in some embodiments of the application to select compounds as "hits" according to the phenotypic Euclidean distance from the median of the negative controls in the multi-factor space. Alternatively, all or a subset of the original variables can be selected that describe the phenotypic differences between plants treated with the candidate herbicide and the negative or positive controls. In another embodiment of the application, a combination of the original variables and the factors is used for hit selection.

[0166] The level of significance can be chosen by the skilled artisan above which a compound will be considered a "hit". In some embodiments, clustering of "hits" with positive controls can be used to predict the mode of action of a potential herbicide. In some embodiments, clustering of "hits" with positive controls can be used to predict the mode of action of a potential plant growth regulator. The "hits" and positive controls can be clustered using the Ward agglomerative method, in which K-means is chosen automatically and the distance between cluster centroids is calculated as the Euclidean distance in the multi-factor space. Alternative agglomerative methods include McQuitty / Weighted Pair Group using arithmetic mean, single agglomerative, complete agglomerative, centroid agglomerative, or median agglomerative. Alternative distance calculations include maximum distance, Manhattan distance, Canberra distance, Minkowski distance, or cosine distance. In some embodiments, if a "hit" falls outside of a cluster associated with a known mode of action positive control, the hit can be visually inspected for the presence of atypical symptoms, and if atypical symptoms are observed, the hit can be predicted to have a novel mode of action. In some embodiments of the application, it is not possible to predict whether a "hit" is a known or novel mode of action. In this case, the hit can be manually progressed to a dose-response experiment, in which plants are treated with a range of concentrations of the "hit" ranging from 1 nM to 50,000 nM, and all resulting data points are processed and reanalyzed according to the methods described herein. Exemplary concentrations for such a dose-response experiment can range from 1 nM to 50,000 nM.

[0167] Artificial intelligence can be used for "hit" selection. In some embodiments of the application, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, or at least 80% of the data obtained by measuring the phenotypic response of plants to negative control compounds is used as a training set for an artificial intelligence algorithm. The algorithm can be a random forest algorithm. In some embodiments of the application, an artificial intelligence algorithm (e.g., a random forest algorithm) can be used to predict the mode of action. Alternatively, a neural network algorithm can be used to predict the mode of action. At least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, or at least 80% of the data obtained by measuring the phenotypic response of plants to positive control compounds can be used as a training set for an artificial intelligence algorithm for predicting the mode of action. Statistical tests can be used to determine the probability that a "hit" matches the phenotypic model generated by the artificial intelligence. In some embodiments of the application, this decision step can be automated. Additionally or alternatively, this decision step can be performed manually.

[0168] The present invention encompasses measuring morphological and / or physiological characteristics of non-vascular plants (e.g., test samples exposed to a candidate compound, a negative and / or positive control) to create a phenotype. The phenotypic response can be used to predict mode of action. In some embodiments of the present invention, each individual plant can be recorded for measurements of less than 10, less than 20, less than 30, less than 40, less than 50, less than 60, less than 70, less than 80, less than 90, less than 100, less than 250, less than 500, or less than 1000 morphological and / or physiological characteristics. Non-limiting examples of morphological and / or physiological characteristics that can be measured include plant length, plant width, plant shape, plant pigmentation, plant roundness, chlorophyll concentration, and cell number per plant.

[0169] In additional embodiments of the present invention, a control sample can be recorded for measurements of less than 10, less than 20, less than 30, less than 40, less than 50, less than 60, less than 70, less than 80, less than 90, less than 100, less than 250, less than 500, or less than 1000 morphological and / or physiological characteristics. Morphological and / or physiological characteristics that can be measured for a control sample also include non-vascular plant length, width, shape, pigmentation, roundness, chlorophyll concentration, and cell number. The control sample can be a sample of the same non-vascular plant as the test sample subjected to the same assay conditions but without the addition of the test compound. In some embodiments of the present invention, DMSO can be added to the control sample in place of the test compound. In other embodiments, another solvent can be added to the control sample. The solvent added can be the same solvent added to the assay well containing the test compound.

[0170] Compounds selected as potential herbicidal or plant growth regulating activity using the methods of the present invention can be referred to herein as "hits." The difference between the phenotype of the non-vascular plant after the assay and the phenotype of the negative control non-vascular plant is referred to herein as the phenotypic response of the non-vascular plant to the compound. Compounds can be selected as "hits" based on the magnitude of the phenotypic response of the non-vascular plant material or germinating spores to the compound. Additionally or alternatively, compounds can be selected as "hits" based on the nature of the phenotypic response of the non-vascular plant material or germinating spores to the compound.

[0171] In some embodiments of the application, the mode of action of a "hit" can be predicted. This can be achieved by creating a phenotype of a non-vascular plant when assayed with a compound having a known herbicidal activity or plant growth regulator activity. The phenotype of the plant or germinating spore assayed with the test compound can then be compared to the phenotype of the non-vascular plant assayed with a compound having a known herbicidal activity (referred to herein as a "positive control"). Compounds having a known herbicidal activity that can be used as a positive control in the methods of the application include commercial herbicides used at concentrations known to cause symptoms specific to the mode of action. Examples of such compounds include clodinafop-propargyl, cyhalofop-butyl, diclofop-methyl, fenoxaprop-P-ethyl, fluazifop-P-butyl, haloxyfop-R-methyl, propaquizafop, quizalofop-P-ethyl, dalapon, benzofluor, clethodim, cycloxydim, alloxydim, alloxydim, oxadiargyl, oxadiazon, pinoxaden, carfentrazone-ethyl, flumiclorac-pentyl, sulfentrazone, clomazone, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocet, cloquintocetimazamethabenz-methyl, imazamox, imazapyr, imazaquin, imazethapyr, cloransulam-methyl, diclosulam, florasulam, metosulam, penoxsulam, bispyribac-Na, pyriminobac-methyl, flucarbazone-Na, propoxycarbazone-Na, benfluralin, butralin, dinitramine, ethalfluralin, oryzalin, pendimethalin, trifluralin, amiprophos-methyl, butamiphos, dithiopyr, ethofumesate, propyzamide = Ronstar, tebutam, chlorthal-dimethyl, clomeprop, 2,4-D, 2,4-DB, 2,4-DP, MCPA, MCPB, mecoprop, chloramben, dicamba, TBA, clopyralid, fluroxypyr, picloram, triclopyr, quinclorac, quinmerac, benazolin-ethyl, ametrine, atrazine, cyanazine, desmetryne, dimethametryne, propachlor, propazine, sebuthylazine, simazine, simetryne, metamitron, terbumeton, terbutryn, terbutylazine, hexazinone, metamifop, metribuzin, amicarbazone, bromacil, noruron, lenacil, bilanafos, desmedipham, phenmedipham, bromofenoxim, bromoxynil, dichlobenil, bentazone, pyridate, pyridafol, chloridazon, chlorotoluron, dymron, isouron, metoxuron,diuron, ethidimuron, fenuron, fluometuron, isoproturon, isouron, linuron, methabenzthiazuron, methylthiocarb, methoxuron, monuron, neburon, siduron, tebuthiuron, clomazone, butylate, cycloate, dimepiperate, EPTC, esprocarb, molinate, orbencarb, pebulate, prosulfocarb, benthiocarb, tiocarbazil, triallate, vernolate, bensulide, benfuresate, ethofumesate, glyphosate, sulfosate, glufosinate-ammonium, bilanaphos, amitrole, norflurazon, diflufenican, picolinafen, beflubutamid, fluridone, flurochloridone, flurtamone, clomazone, acifluorfen-Na, bifenox, chlomethoxyfen, fluoro- glycofen-ethyl, fomesafen, halosafen, lactofen, oxyfluorfen, fluazolate, pyraflufen-ethyl, cinidon-ethyl, flumioxazin, flumiclorac-pentyl, fluthiacet-methyl, thidiazimin, oxadiargyl, azafenidin, carfentrazone-ethyl, sulfentrazone, pentoxazone, benzfendizone,butafenacil, pyraclonil, profluazol, flufenpyr-ethyl, acetochlor, alachlor, butachlor, dimethachlor, dimethenamid, metazachlor, metolachlor, pethoxamid, pretilachlor, propachlor, propisochlor, thenylchlor, diphenamid, napropamide, naproanilide, flufenacet, mefenacet, fentrazamide, anilofos, cafenstrole, piperophos, DSMA, MSMA, asulam, naptalam, diflufenzopyr-Na, dichlobenil, chlorthiamide, isoxaben, flupoxam, diquat, paraquat, chlorpropham, propham, carbetamide, DNOC, dodine, dinoseb, phenmedipham, flamprop-M-methyl / -isopropyl, quinclorac, TCA, dalapon, flupropanate, difenzoquat, mesotrione, sulcotrione, isoxachlortole, isoxaflutole, benzofenap, pyrazolynate, pyrazoxyfen, benzobicyclon, bromobutide, (chloro)-bromoxynil, cinmethylin, cumyluron, dazomet, daimuron, etobenzanid, fosamine, indanofan,metam, oxaziclomefone, oleic acid, pelargonic acid, pyributicarb. One skilled in the art can select any herbicide with a known mode of action for use as a positive control in the methods described. Similarly, compounds with known plant growth regulatory activity that can be used as positive controls in the methods of the application include commercial plant growth regulators used at concentrations known to cause symptoms specific to the mode of action. One skilled in the art can select any plant growth regulator with a known mode of action for use as a positive control in the methods described.

[0172] Target identification

[0173] The methods of the application can include assays for identifying targets (e.g., protein targets) of herbicides, such as, for example, candidate compounds screened and identified as having herbicidal activity and / or targets of known herbicides for which the mode of action is unknown.

[0174] In some embodiments, target identification can involve contacting a "hit" identified by the methods of the application with a non-vascular plant that has been mutagenized. DNA can then be extracted from the plants that survive the contacting. Mutagenic methods are standard in the art. The mutagen can be, for example, radiation. In some embodiments, the mutagen is selected from the group consisting of ultraviolet (UV) light, x-rays, gamma rays, and neutrons. In further embodiments, the mutagen can be UV light, which can be UV-A, UV-B, or UV-C light. Additionally or alternatively, mutagenesis can be performed using a chemical agent. Non-limiting examples include alkylating agents, such as ethyl methane sulfonate (EMS). In some embodiments, dimethyl sulfate, sodium azide, or methyl nitro nitroso guanidine (MNNG) can be used to introduce mutations into the non-vascular plant. The chemical agent can also be a deaminating agent or an intercalating agent. In further embodiments of the application, the mutagen is a transposable element.

[0175] DNA extraction methods are standard in the art. DNA can be extracted using phenol, chloroform, and isoamyl alcohol. Other well-known DNA extraction techniques include enzymatic, silica (spin) column-based methods, anionic resins, methods using magnetic beads, and CaCl density gradient DNA extraction methods. Cetyltrimethylammonium bromide (CTAB) and 2-β-mercaptoethanol are commonly used for plant DNA extraction where plant tissue contains high levels of polysaccharides, polyphenols, and / or other secondary metabolites (see, e.g., Clark, Plant molecular biology - a laboratory manual, 1997; Springer: 305-328). DNA can be extracted from whole mutant non-vascular plants or a sub-portion of whole mutant non-vascular plants or mutant non-vascular plant mutant spores, gemmae, explants, protoplasts, or vegetative propagules.

[0176] In some embodiments of the application, a genomic DNA library is prepared. In further embodiments, the library is then sequenced. Any high-throughput sequencing technology capable of sequencing the entire genome of a plant can be used, including clonally amplified-based technologies, semiconductor-based technologies, and single molecule real-time (SMRT) sequencing (for a recent review of potentially suitable commercially available platforms, see Reuter et al., Molecular Cell, 2015; 58: 586-597). The raw sequencing reads can then be "trimmed" to remove poor quality sequence and / or artifacts of the sequencing process such as primers and sequencing adapters. Any suitable known software program can be used, such as, for example, Trimmomatic. Trimmomatic trims Illumina sequencing adapters and read portions associated with poor sequencing quality. Other known methods for performing quality trimming can also be used.

[0177] In some embodiments of the application, the read files can be interleaved. Interleaving can be performed using any suitable parsing script. For example, where paired reads are obtained by a sequencing system, a parsing script can be used to recombine both partners of all paired reads into a single file.

[0178] Some embodiments can include a normalization step. For example, the normalization process can be performed by using a script that calls any suitable known software program (e.g., Khmer) to normalize by 31-mers. In this example, the normalization program looks at the distribution of k-mers in all reads using a predetermined value of k, and discards an appropriate amount of reads containing the most common k-mers as they only provide redundant information. This step can be performed to make the alignment process more storage efficient.

[0179] The deinterleaved or decoupled normalized read files can then be resolved using any suitable resolution script that separates the two mates of all paired reads. This step is the reverse of the interleaving step. For each paired read, there are two mates that are identified as belonging to the same paired read. They can be written into the same file (i.e. interleaved) or separate files (deinterleaved). The process from one to the other is resolved by a tag string that identifies the mates as belonging to the same paired read. This tag is derived from the file produced by the sequencing platform and can look like XYZ / 1 for mate 1 and XYZ / 2 for mate 2, for example. The software identifies them by text matching and writes the corresponding DNA sequences into the same file or two separate files.

[0180] The sequenced genome can then be aligned to a reference genome of the plant. The reference DNA sequence can be a known reference sequence of a plant of the genus. Reference DNA sequences are published on publicly available databases. The entire genome sequence is publicly available for many non-vascular plants, including, for example, mosses such as liverworts (reference sequences for the nuclear genome and the organelle genome of liverworts are publicly available).

[0181] In some embodiments of the invention, the reference DNA sequence can be aligned to additional comparison sequences. The comparison sequences can be from independent plants belonging to the same genus, that did not survive contact with the compound. In some embodiments, the method of the invention can involve obtaining a set of mismatches between the DNA sequence of the mutant plant and the reference DNA sequence. A second set of mismatches can then be obtained between the reference DNA sequence and the comparison sequences. Additional embodiments of the invention can then involve filtering the first set of mismatches with respect to the second set of mismatches to identify a subset of mismatches that are unique to the first set of mismatches. The subset of mismatches can be candidate mutations for causal mutations for herbicide resistance or plant growth regulation. This can aid in identifying the target of a novel herbicide or plant growth regulator identified by the method of the invention.

[0182] A method for identifying a causal mutation that causes a phenotype of interest in a test sample according to embodiments of the invention is provided in Figure 15 .

[0183] In one embodiment, target identification for identifying mutations associated with a phenotype of interest in a non-vascular plant is performed according to the following method:

[0184] (a) aligning the DNA sequence of the test sample to a reference DNA sequence and identifying a first set of sequence mismatches between the two sequences;

[0185] (b) aligning the DNA sequence of the at least one comparison sample to the reference DNA sequence and identifying a second set of sequence mismatches between the two sequences;

[0186] (c) filtering the first set of mismatches against the second set of mismatches to identify a subset of mismatches that are unique to the first set of mismatches, wherein the subset of mismatches is a candidate mutation for the causal mutation;

[0187] wherein the test sample and the comparison sample are from independent non-vascular plants that exhibit the phenotype of interest and wherein the independent non-vascular plants are of the same genus; and

[0188] wherein the reference DNA sequence is a known reference sequence of a vascular plant of the genus.

[0189] In one embodiment, the target identification for identifying mutations associated with a phenotype of interest in non-vascular plants is performed according to the following method:

[0190] (a) aligning the DNA sequence of the test sample to the reference DNA sequence and identifying a first set of sequence mismatches between the two sequences;

[0191] (b) aligning the DNA sequence of the at least one comparison sample to the reference DNA sequence and identifying a second set of sequence mismatches between the two sequences;

[0192] (c) filtering the first set of mismatches against the second set of mismatches to identify a subset of mismatches that are unique to the first set of mismatches, wherein the subset of mismatches is a candidate mutation for the causal mutation;

[0193] wherein the test sample is from a non-vascular plant that exhibits the phenotype of interest and wherein the comparison sample is from a non-vascular plant of interest that belongs to the same genus, does not exhibit the phenotype of interest; and

[0194] wherein the reference DNA sequence is a known reference sequence of a vascular plant of the genus.

[0195] In one embodiment, the test sample and / or the at least one comparison sample is biological material from a non-vascular land plant, wherein the non-vascular plant is a bryophyte. In one embodiment, the test sample and / or the at least one comparison sample is biological material from a bryophyte selected from the group consisting of a moss, a liverwort, and a hornwort. In one embodiment, the test sample and / or the at least one comparison sample is biological material from a leafy liverwort, a simple thallose liverwort, or a complex thallose liverwort. In one embodiment, the test sample and / or the at least one comparison sample is from a plant of the genus Marchantia.

[0196] In some embodiments of the application, the DNA sequence of an additional comparative sample can be aligned to the reference DNA sequence to identify a third set of sequence mismatches between the two sequences. The first set of mismatches can then be filtered against the third set of mismatches to identify a subset of mismatches common to both the first and third sets of mismatches. Both the subset of mismatches can then be candidate mutations for herbicide resistance or plant growth regulation. The additional comparative sample can be from an independent plant belonging to the same genus that did not survive contact with the compound. Some embodiments of the application involve aligning the DNA sequence of 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or more comparative samples to the reference DNA sequence to identify sets of mismatches between the sequences, which can then be used to compare the set of mismatches between the DNA sequence of the mutant plant and the reference DNA sequence to identify unique mismatches that are candidate mutations for herbicide resistance or plant growth regulation. Numerous software packages are available to assist the technician in aligning DNA sequences and filtering sets of mismatches.

[0197] In some embodiments of the application, genomic regions that align more reads than expected can be excluded. That is, sequencing depth is defined by the number of sequencing reads from a sample that align to a region of the reference DNA sequence. When sequencing the DNA sequence of a sample, the user can choose how many times to sequence the same portion of the DNA sequence. This choice defines the expected sequencing depth. For example, a sequencing depth of 1 would require the sampling system to sequence the entire DNA sequence of the sample once. For an expected sequencing depth of 20, the sampling system would sequence the DNA of the sample 20 times.

[0198] Thus, as an example, if the observed sequencing depth at a defined position is 10, 10 sequencing reads are aligned to the region of the reference DNA sequence that includes that position. If the expected sequencing depth is 1, this indicates that 9 of the 10 reads are incorrectly aligned to that region of the DNA sequence. For this reason, the software considers any mismatches in the region of the reference DNA sequence where the observed sequencing depth is higher than the expected sequencing depth as a likely result of having incorrectly aligned reads, and thus removes them from the set of mismatch data. In other words, the mismatches are considered to be alignment artifacts rather than candidate mutations, and thus discarded or removed from the data set. In other words, to determine the first or additional sets of mismatch DNA sequence data, the method and software can reject at least one region of the sample DNA sequence that aligns to the reference DNA sequence based on the actual read depth exceeding the expected read depth. Numerous suitable software programs are available to implement this function.

[0199] In addition, the frequency of mismatches occurring in a set of reads aligned at a position in the genome is used to filter out alignment artefacts. For example, if the mutant is a diploid species, the expected frequency of mismatches in the mutant genome is 50%, while in haploid species it is 100%. If the observed frequency of mismatches does not match the expected frequency of mismatches for the defined species, the relevant reads are discarded from the set of data. Again, this applies to the data sets of the sample and the comparative DNA sequence.

[0200] In some embodiments of the application, the non-vascular plant that has been mutagenized is an Ml mutant. "Ml" refers to the first generation of mutants, meaning that the Ml mutant is not a child of the mutagenized non-vascular plant. The mutagenized non-vascular plant can comprise non-naturally occurring mutations. Some embodiments of the application do not include the steps of segregation analysis, complex segregation analysis, or population segregation analysis.

[0201] Additionally or alternatively, target identification can be achieved by any one or more of the following: enzyme assays, chlorophyll fluorescence kinetic assays, photosynthetic oxygen evolution assays, electrolyte leakage assays, radiometric assays, spectrophotometric assays, fluorometric assays, absorbance assays, colorimetric assays, mass spectrometry, mitotic index analysis, quantitative PCR analysis, transcriptomic profiling analysis, proteomic profiling analysis, whole genome analysis, quantitative trait locus analysis, computer modelling docking studies, chemical structure analysis (some reviewed in Dayan 2015 "Biochemical markers and enzyme assays for herbicide mode of action and resistance studies"). The skilled person will be familiar with all of the aforementioned techniques for downstream analysis.

[0202] incorporated by cross-reference

[0203] This application claims priority from Australian provisional patent application number 2019904145 filed on 4 November 2019, the entire contents of which are incorporated herein by cross-reference.

[0204] Examples

[0205] The application will now be described with reference to the following specific examples, which are in no way to be construed as limiting.

[0206] Example One: Geminil spores as a screening system for herbicidal activity

[0207] Preparation of assay plates

[0208] Front end assays were set up manually or by liquid handling robots. Manual assay plate preparation started with the addition of 1 μΐ, of a 10% DMSO solution followed by 99 μΐ, of a cell suspension in liquid nutrient medium to a 96-well microtiter plate. Automated assay plate preparation started with the addition of 7 μΐ, of a 1.4% DMSO solution followed by 93 μΐ, of a cell suspension in liquid nutrient medium to a 96-well microtiter plate. Typically, only one screening compound was tested in one well and only one well was used to test a screening compound.

[0209] To maximize the statistical significance of the assay while avoiding overlap of germinating spores, 50-70 P. polymorpha spores were placed in each well of a 96-well plate Figure 1 ).

[0210] Growth conditions

[0211] The assay plates were placed in a cabinet that controlled light and temperature. The plants were grown for 5 days under continuous illumination. The temperature was set to 23 degrees Celsius. The plants were grown in Johnson's medium.

[0212] Imaging

[0213] Microtiter plates containing 1 -day-old to 5-day-old P. polymorpha germinating spores grown in the presence of negative controls (DMSO), positive controls (commercial herbicides), or screening compounds were placed in a high-throughput fluorescent microscope, InCell Analyzer 2500 (GE Healthcare). Several images were generated Figure 2 , Figure 5): an image recording only far-red photons by the camera chip (far-red image), an image recording only cyan photons (cyan image), an image recording only yellow photons (yellow image), and an image recording transmitted light (brightfield image). A 2X objective lens was used to generate the far-red image, where the field of view covers the entire well. Here, the far-red image is a 2D fluorescence micrograph. A 4X objective lens was used to generate the far-red image, where the field of view fits exactly within the boundaries of the well. Again, the far-red image is a 2D fluorescence micrograph. A 10X objective lens was used to generate a set of 4 far-red images, 4 cyan images, 4 yellow images, and 4 brightfield images, which cover a small fraction (1 / 8) of the well. Here, the far-red image is a maximum intensity projection of a 5-slice stack, where each slice is 20 pm deep. The cyan and yellow images are maximum intensity projections of a 4-slice stack, where each slice is 20 pm deep. The brightfield image is a single 20 pm deep slice. Any or all four channel images can be used in subsequent image analysis. By using objective lenses with 2x, 4X, 10X, 20X, or 40X optical magnification, the images can have a larger, equal, or smaller field of view compared to the diameter of the well.

[0214] Discussion

[0215] To date, Marchantia spores or any spore-forming plant has not been used as a model organism for herbicide discovery. Marchantia polymorpha is not an obvious choice as a model organism for herbicide discovery because, like mosses in general, it is not sensitive to the major herbicides glyphosate and phosphoramidon.

[0216] To grow a large number of plants suitable for high-content phenotypic characterization, the developmental stage of Marchantia gemmiferous spores has to be fine-tuned.

[0217] - Effect of developmental stage

[0218] Marchantia spores were grown for five days until they reached a developmental stage that formed a diverse array of cell types and tissues representative of the whole plant. That is, 5-day-old germinating spores had rhizoids, plant photosynthetic cells, and nascent meristems. In addition, 5-day-old germinating spores presented a size and shape that was more convenient for microscopy than later developmental stages.

[0219] Example Two: Herbicide hit identification and mode of action prediction using MoA Galaxy

[0220] Background

[0221] Plants respond to herbicide treatment in a variety of ways, from barely detectable physiological changes, minor lesions to plant death. The amplitude of the response is related to the potency of the chemical applied to the plant. The nature of the response differs for chemicals acting through different modes of action.

[0222] Quantification of the amplitude of the phenotypic response of plants treated with screening compounds helps to select hits based on herbicide potency. Independently of the intensity of the response, the nature of the symptoms exhibited by plants treated with screening compounds can be used to select hits. In summary, the amplitude and nature of the plant response to screening compounds not only indicates herbicide potency, but can also indicate herbicide mode of action. Phenotypic fingerprinting of life system responses to chemicals is used in the field of drug discovery (linking phenotypes and modes of action through high content screening fingerprints), and software tools to assist the experimenter are widely available. This method of chemical screening is called high content screening, and relies on the simultaneous recording of multiple phenotypic descriptors, such as length, area, color, shape, etc.

[0223] However, translating this method to herbicide discovery is a challenge, as the in vivo model currently used is hampered by its morphological complexity: 3D seedlings are more complex objects for imaging than 2D cell cultures, especially at the throughput necessary to screen large chemical libraries. The inventors have developed a high content screening platform that takes advantage of the aptitude of Marchantia spores to grow close to 2D (and transform with fluorescent cell markers).

[0224] Methods

[0225] - Image analysis

[0226] Images are processed individually using the software Developer. The image analysis protocol is designed to distinguish plants and subcellular objects (such as nuclei) from the background Figure 3 and 20 ).

[0227] When multiple images are produced, segmented objects can be linked so that any object segmented using one image is the ontology of multiple objects segmented using another image. Then 10-50 measurements describing the morphological and physiological phenotype of the plants are extracted. For example, a non-exhaustive list of measurements is e.g. "plant length", "plant width", "plant circularity", "chlorophyll fluorescence intensity" and "number of cells per plant". Measurements are finally recorded in.csv or.txt files for each plant in each imaged well.

[0228] - Data formatting

[0229] Output files are reformatted using custom parsing scripts for downstream analysis compatibility. Scripts automatically add relevant metadata such as the barcode of the assay plate analyzed and / or the date and protocol of image acquisition and image analysis. In addition, scripts for high resolution analysis output files (4X and 10X datasets) introduce a default value + / - error factor for the variable that takes a value only when the ontology brightfield object is not linked to the cyan, far-red or both cyan and far-red objects respectively define the cyan, far-red or both cyan and far-red objects.

[0230] - Data preparation

[0231] Data analysis was performed using HC StratoMineR software developed by CoreLife Analytics. The software provides a data analysis pipeline where each step can be customized by changing the statistical method and parameters used. The steps, methods and parameters used in this example are provided below:

[0232] Data was normalized by the median of the negative (DMSO) controls at plate level to minimize noise across plates. Normality of the distribution of the variables was checked and transformed if needed after automatic recommendations of the software. Data transformation methods can also be manually selected from the following list: square root, power of 2, power of 3, log, log2, log10, inverse. Variables were then scaled to balance the weight of the variables in different means in downstream analysis steps. Scaling was performed at plate level using Z-score method ( Figure 5 ). Filtering level scaling can also be performed at this step when the number of data points per plate is too low for accurate running of downstream analysis steps.

[0233] Relevant variables in negative control and screening compound data were transformed into factors by starting a common factor analysis with 200 iterations of t-SNE, t-SNE perplexity set to 30, orthogonal rotation method and factor score method as ten Berge. The number of factors to retain can be automatically determined according to Kaiser, Elbow or Joliffe criteria or manually selected upon visual inspection of the screening plot ( Figure 6 ).

[0234] - Hit selection

[0235] - Unsupervised hit selection

[0236] Screening compounds are selected as hits based on their phenotypic Euclidean distance from the median of the negative controls in the multi-factor space defined by the previously reserved factors. A p-value of 0.0001 is selected as the significance threshold above which a screening compound is considered significantly different from the median of the negative controls Figure 7 .

[0237] - AI supervised hit selection

[0238] Screening compounds are selected as hits based on their dissimilarity from the phenotypic model of the negative controls Figure 8 . The phenotypic model of the negative controls and positive controls is generated by using a random forest algorithm of 128 trees using 80% of the respective dataset for training and the remaining 20% for testing the generated model Figure 9 .

[0239] - Known or unknown mode of action prediction

[0240] - Unsupervised mode of action prediction

[0241] The mode of action of the hits selected with the unsupervised method is predicted by the cluster of the hits with the positive controls. The positive controls are commercial herbicides used at concentrations known to cause symptoms of a specific mode of action. The concentrations of the commercial herbicides known to cause symptoms of a specific mode of action are determined experimentally by visually inspecting a dose-response experiment where plants are exposed to a range of herbicide concentrations.

[0242] The clustering of the hits and positive control standards is generated using the Wharton clustering method, where K-means is automatically selected and the distance between the cluster centers is calculated as the Euclidean distance in the multi-factor space, applying a significance threshold of p-value = 0.0001 Figure 10 .

[0243] On average, the automatic classification of the hits into clusters is 86%, which is consistent with the manual classification of the hits into phenotypic similarity groups. If a hit falls outside of a cluster associated with a known mode of action positive control, the hit is visually inspected for the presence of unique symptoms, and if unique symptoms are observed, the hit is predicted to have a novel mode of action.

[0244] If a hit cannot be predicted to be a known or novel mode of action, the hit is manually progressed to a dose-response experiment, where plants are treated with a range of concentrations of the hit from 1 nM to 50,000 nM, and all resulting data points are described methodically Image analysis .

[0245] - AI supervised mode of action prediction

[0246] Statistical tests are performed to determine the likelihood that the hit matches the phenotype model of any of the positive controls. If the probability is higher than an arbitrary defined threshold that depends on the number of positive control phenotype models, the hit is expected to have a known mode of action and that mode of action is the mode of action of the herbicide to which the hit is most strongly correlated with its phenotype model. This decision step is currently performed manually, but can be automated by a custom parsing script.

[0247] If the hit falls below an arbitrary defined significance threshold for all positive control phenotype models, the hit is visually inspected for the presence of unique symptoms and if unique symptoms are observed, the hit is predicted to have a novel mode of action.

[0248] If it is not possible to predict whether the hit has a known or novel mode of action, the hit is progressed to a dose-response experiment, in which the plants are treated with a range of concentrations of the hit from 1 to 50000 nM and the resulting data points are described by a method from the prior art. Image analysis Discussion The resulting data points are described by a method.

[0249] Background

[0250] To date, high content screening has not been applied to herbicide discovery screening. Recent technical advances on the application of high content analysis to herbicide discovery and simultaneous mode of action prediction have been limited to a few modes of action (one automated quantitative image analysis tool for identifying microtubule patterns in plants). Furthermore, these methods require higher resolution (confocal microscopy) and are therefore dependent on explants or sub-portions of plants rather than micro- or whole-plant screening systems.

[0251] In contrast, the present method is applicable to more modes of action and uses micro- plants as it does not rely on confocal microscopy. Therefore, the method of the present invention is higher throughput and has a broader range.

[0252] The application of high content screening methods to herbicide discovery activities is non-obvious due to the lack of availability of screening systems using micro- or whole-plants.

[0253] Example Three: Identification of mutants resistant to herbicide hits

[0254] Methods

[0255] Knowledge of the target informs the mode of action, toxicity, resistance break and further screening or lead optimization efforts. Mutations in genes encoding protein targets can confer resistance to herbicides. Therefore, the inventors of the present invention experimented with the reverse identification of target genes.

[0256] Mutagenesis of spores is a convenient system for screening for mutations conferring herbicide resistance due to its small size and amenability to simple methods of radiation mutagenesis. Herbicide resistance screening following spore mutagenesis has not been previously reported in Marchantia.

[0257] Figure 11

[0258] The lethal concentration of the herbicide hit is determined under conditions of mutagenic concentration. 20,000 Marchantia spores are plated in 90 mm Petri dishes containing 25 mL of Johnson's medium containing 1.4% agar and supplemented with a range of herbicide hit concentrations from 1 to 50,000 nM. The lethal concentration is defined as the minimum concentration of herbicide hit sufficient to kill 100% of wild type Marchantia plants Figure 12 ).

[0259] If no lethal concentration is observed, the highest concentration can be used instead of the lethal concentration, provided that it causes the plants to display a phenotype different from that of the untreated plants. For example, this alternative phenotype can be a strong reduction of growth Figure 13 ) or a clear change in plant shape, or a change in plant pigmentation.

[0260] A mutagenized population of Marchantia spores is generated using a dose of physical or chemical mutagen that kills a defined proportion of the spores of the mutagenized population Figure 14 ). The defined proportion of the mutagenized population of spores defining the experimental mutagen dose can be 50% as is standard in the art, or higher or lower, depending on the phenotype of the wild type plants treated with the herbicide hit at the mutagenic screening concentration.

[0261] A population of 400,000 or more mutagenized spores is spread between 20 90 mm Petri dishes containing 25 mL of Johnson's medium containing 1.4% agar and supplemented with the herbicide hit at a concentration 10 times higher than the lethal concentration. Alternatively, a population of 400,000 or more wild type spores is spread between 20 90 mm Petri dishes containing 25 mL of Johnson's medium containing 1.4% agar and supplemented with the herbicide hit at a concentration 10 times higher than the lethal concentration and the wild type spores are then mutagenized.

[0262] One example of a mutagenesis method is UV-B mutagenesis. 400,000 spores are spread between 20 90 mm petri dishes containing 25 mL of Johnson's media with 1.4% agar and supplemented with the herbicide at a concentration 10 times higher than the lethal concentration. The petri dishes are then inverted and inserted into a UV-B transilluminator with the spores directly facing the UV-B light source. The UV-B light is shined onto the spores for the duration of time necessary to achieve the desired mutagenesis dose. The petri dishes are then closed and wrapped in aluminum foil to block light, and the petri dishes are placed in a 23 degree Celsius incubator overnight.

[0263] The mutagenized spores in the petri dishes containing the desired concentration of herbicide hits are placed in an incubator with constant lighting, Lux, and 23 °C, and the plants are grown for 14 days. Survivors Figure 16 ) or otherwise untreated sample plants are transferred to new petri dishes in the absence of the herbicide hits to grow for another 14 days. Herbicide resistance is verified by transferring segments of the growing plants to fresh Johnson's media with 1.4% agar supplemented with a concentration 10 times higher than the lethal concentration.

[0264] If the resistant phenotype is verified, genomic DNA is extracted from any part of the full mutant plant using any DNA extraction method such as but not limited to phenol-chloroform-IAA extraction. A genomic DNA library is prepared and sequenced using any Illumina next generation sequencing newer than the HiSeq2000.

[0265] Example Four: Identifying causal mutations in RHO GTPASES genes of plant-enhancing proteins that impair fertility (B case)

[0266] In this example, the methods of the present invention are used to identify causal mutations associated with the rhizoid / epidermis phenotype in Marchantia polymorpha as detailed below. These methods are equally applicable to determining phenotypes associated with herbicide resistance or plant growth regulation.

[0267] Several independent mutant lines are generated by irradiating Marchantia polymorpha with ultraviolet B. The mutant lines are separated into two phenotypic groups: some have straight rhizoids Figure 17 A) and intact epidermis Figure 16 A), some have wavy rhizoids Figure 17 B) and stretched epidermis Figure 18 B).

[0268] We aimed to identify causal mutations in the UV4.32 mutant line, which has wavy rhizoids and stretched epidermis. DNA was extracted from the UV4.32 mutant with wavy rhizoids and stretched epidermis using whole plants as samples and standard DNA phenol chloroform-IAA extraction. The genomes of UV4.32 and 7 independent mutant lines with straight rhizoids and intact epidermis were sequenced using Illumina's HiSeq-2000 platform technology.

[0269] Raw reads were quality trimmed using Trimmomatic-0.32 and normalized using Khmer 0.7.1 with a kmer size of 31. The resulting reads were aligned to the reference genome using bowtie2-2.1.0 set to very sensitive local mode. The reference genome used was the publicly available draft assembly of the P. multiflora genome on the NCBI Whole Genome Shotgun (WGS) database.

[0270] Alignments were position sorted and mismatches within reads with a q quality higher than 35 were extracted from bio-samtools-2.0.5 using the functions sort and mpileup. Because they can be caused by misalignments, mismatches in regions with coverage over 100X were excluded from the samtools-0.1.9 package's bcftools using the varFilter function. Mismatches were then only retained if they were supported by more than 7 reads and looked sufficiently homozygous based on a negative FQ value or an AF1 value higher than 0.5001.

[0271] A total of 143292 mismatches were identified in UV4.32 before any filtering. The number of mismatches specific to UV4.32 decreased with the number of UV mutant lines with straight rhizoids and intact epidermis used for filtering (Figure 1). Figure 18 A).

[0272] Finally, using all filtered lines sequenced, the number of candidate mismatches was reduced to 12000 mismatches, or more than 90% reduction (Figure 2B). This shows that the filtering step of subtracting the set of mismatches in the comparison samples predicted to not have causal mutations from the set of mismatches in the test sample before the standard filtering steps increases the stringency of the identification of candidate mismatches. Figure 18 B). This shows that the filtering step of subtracting the set of mismatches in the comparison samples predicted to not have causal mutations from the set of mismatches in the test sample before the standard filtering steps increases the stringency of the identification of candidate mismatches.

[0273] Subsequent filtering steps were performed to filter mismatches that are not consistent with the UV signature, filter mismatches outside of the coding sequence of genes and filter non-synonymous mismatches. These three filtering steps further reduced the number of candidate mismatches to 10 mutations consistent with the expected UV mutation signature (Figure 3A), predicted to be located in the coding sequence of genes (Figure 3B). Figure 18 Table 1: Candidate mutations in Marchantia genes and corresponding Arabidopsis homologs of UV4.32. Arabidopsis is the most mature model in plant genetics and the function of Marchantia genes can be inferred by analogy with the function of Arabidopsis genes. ​) and changed the amino acid sequence of the corresponding protein (Table 1).

[0274]

[0275] REN Ren

[0276] Among these 10 mutations, the strongest mutation is a 2 base pair deletion that causes an early stop codon in Mp Figure 19 Figure 19 Mutants exhibit the same phenotype as UV4.32 (Honkanen et al., 2016). This indicates that the subsequent filtering steps are conservative enough.

[0277] In summary, this shows that we are able to identify a small number of mutations (including causal mutations) based on a pipeline version that subtracts the set of mismatches in a comparison sample that is predicted to not have a causal mutation from the set of mismatches in a test sample without the need to cross the mutant line with a wild type.

[0278] Example Five: Discovery of mutations in acetolactate synthase gene that cause resistance to chlorsulfuron (A example)

[0279] Spores of Physcomitrella patens were irradiated with ultraviolet B light and seven independent mutant lines that were resistant to the herbicide chlorsulfuron were identified. Chlorsulfuron resistance was determined by survival of P. patens plants after two weeks of exposure to a lethal dose of chlorsulfuron (0.1 ppm dose, i.e., a dose sufficient to kill 100% of wild type plants).

[0280] Since all mutant plants share the same phenotype - chlorsulfuron resistance - we hypothesized that they each have the same causal mutation. Comparing the chlorsulfuron resistant mutants to the reference genome identified over 100,000 mismatches alone, and we first filtered out mismatches that are also present in the M0 wild type genome (the 2 leftmost scatter boxes). Candidate mutations in chlorosulfuron mutants (example A)

[0281] To test the efficiency of the allele-based pipeline version, we applied it to the combination of 4, 5, 6, and all 7 chlorsulfuron mutants. The more alleles we subtracted the lines with, the more efficient the pipeline became. In fact, using all 7 chlorsulfuron resistant lines, we reduced the number of mismatches from close to 100,000 to 11 candidate mutations that are consistent with the expected mutation signature and are in the coding sequence of the gene ( ​ ).

[0282] ​​Of the 11 candidate mutations that were shared by all 7 chlorimuron- resistant mutants but not in the wild type, 5 caused a change in the amino acid sequence of the encoded protein (Table 2). Of these 5 candidate mutations, only one was in a gene with a predicted function. In fact, this exact mutation in the acetolactate synthase gene is known to cause chlorimuron resistance in other plant models.

[0283]

[0284] Table 2: ​

[0285] Example Six: Mutations causing chlorimuron resistance are found in the acetolactate synthase gene (AB example)

[0286] To improve the capacity of the pipeline exemplified in Example 1 and Example 2, we combined the two approaches: in this implementation of the pipeline, we look for causal mutations in sets of mismatches that are shared by the allelic mutants and not present in the wild type and non-allelic mutants.

[0287] Using 3 chlorimuron-sensitive mutagenized lines, we filtered 4 out of the 11 chlorimuron-resistance-specific mismatches identified as consistent with the expected mutation signature and in the coding sequence of the gene, leaving us with only 4 candidate mutations that are predicted to cause a change in the amino acid sequence of the protein (Table 3).

[0288] This represents an increase in the capacity of the pipeline of 20% to 30% compared to the pipeline exemplified in Example 2 alone. Because the capacity of the pipeline in Example 1 and 2 increased with the number of allelic and non-allelic subtractive lines, respectively, we predict that the capacity of the pipeline exemplified in this example will further increase if we use more allelic and non-allelic subtractive lines.

[0289]

[0290] Table 3. Candidate mutations predicted to cause a change in the amino acid sequence of the protein.

Claims

1. A method of screening candidate compounds for herbicidal or plant growth regulating activity, the method comprising the steps of: (i) contacting a plurality of different candidate compounds with a plurality of test samples from non-vascular plants; (ii) determining whether the test samples provide a phenotypic response to the plurality of different candidate compounds by comparison to the phenotype of control samples from non-vascular plants that have not been contacted with candidate compounds; and based on light emission or fluorescence measurements, determining the nature and / or magnitude of the phenotypic response, wherein the nature and / or magnitude of the phenotypic response is indicative of mode of action and / or potency and / or herbicidal or plant growth regulating activity; wherein: the test samples and the control samples comprise whole plants, spores, gemmiferous spores, or vegetative propagules.

2. The method of claim 1, wherein the test samples and the control samples comprise whole plants, gemmiferous spores, or vegetative propagules.

3. The method of claim 1 or 2, wherein the test samples and the control samples comprise explants or protoplasts.

4. The method of claim 1 or claim 2, wherein the test samples comprise test gemmiferous spores and the control samples comprise control gemmiferous spores.

5. The method of claim 1, wherein the candidate compounds are candidate compounds for herbicidal activity.

6. The method of claim 1, wherein the non-vascular plants are mosses, hornworts, or liverworts.

7. The method of claim 4, wherein the test gemmiferous spores and the control gemmiferous spores are derived from spores of non-vascular plants of the same species.

8. The method of claim 4, wherein the test gemmiferous spores comprise moss gemmiferous spores, liverwort gemmiferous spores, hornwort gemmiferous spores, or any combination thereof.

9. The method of claim 1, wherein each member of the plurality of different candidate compounds is contacted with a different test sample.

10. The method of claim 1, wherein a plurality of members of the plurality of different candidate compounds is contacted with a single test sample.

11. The method of claim 1, wherein the test samples and control samples are leafy liverwort gemmiferous spores, simple thallus liverwort gemmiferous spores, complex thallus liverwort gemmiferous spores, or any combination thereof. ​ 12. The method of claim 1, wherein the test sample and control sample are selected from the group consisting of: P. hornum gemmae, P. aquaticum gemmae, P. bertieri gemmae, P. carrii gemmae, P. chenopodioides gemmae, P. delicatulum gemmae, P. dominii gemmae, P. eurya gemmae, P. foliosum gemmae, P. incanum gemmae, P. incurvum gemmae, P. lineare gemmae, P. major gemmae, P. neoguineense gemmae, P. pallescens gemmae, P. palmiforme gemmae, P. pellucidum gemmae, P. papaveraceum gemmae, P. polymorphum gemmae, P. rubrum gemmae, P. solomonense gemmae, P. stearnianum gemmae, P. subpinnatum gemmae, P. vitreum gemmae, P. wallisii, P. nepalense, and combinations thereof.

13. The method of claim 4, wherein the test gemmae are disposed in a series of different wells, each well containing 400-800 gemmae / mL.

14. The method of claim 4, wherein the test gemmae are disposed in a series of different wells, each well containing 300-900 gemmae / mL.

15. The method of claim 4, wherein the test gemmae are disposed in a series of different wells, each well containing 200-1000 gemmae / mL.

16. The method of claim 1, wherein the test sample and / or the control sample has been engineered to express a fluorescent molecule.

17. The method of claim 1, wherein the control sample is a positive control.

18. The method of claim 17, wherein the positive control has been contacted with a known herbicide or plant growth regulator.

19. The method of claim 1, wherein the control sample is a negative control.

20. The method of claim 19, wherein the negative control has not been contacted with a known herbicide or plant growth regulator.

21. The method of claim 1, wherein step (ii) comprises comparing the phenotype of the test sample to the phenotype of a positive control sample that has been contacted with a known herbicide compound or plant growth regulator compound, and further comprising comparing the phenotype of the test sample to the phenotype of a negative control sample that has not been contacted with a known herbicide or plant growth regulator compound.

22. The method of claim 21, wherein the known herbicide compound has a known mode of action, and the comparison of test sample phenotype to positive control sample phenotype is used to predict the mode of action of candidate compounds identified as having herbicidal or plant growth regulating activity.

23. The method of claim 21, wherein the test sample, the negative control sample, and the positive control sample are gemmae derived from spores of non-vascular plants of the same species.

24. The method of claim 21, wherein the test sample, the negative control sample, and the positive control sample have been engineered to express a fluorescent molecule.

25. The method of claim 1, wherein step (ii) comprises measuring the phenotypic response of the test sample after the contacting, after growing the test sample in a suitable medium with the candidate compound under suitable conditions for a period of time between 1 to 3 days or between 3 to 6 days, and wherein the phenotype of the control sample is determined after growing the control sample in the suitable medium under the suitable conditions for an equivalent period of time.

26. The method of claim 1, wherein step (ii) comprises measuring the phenotypic response of the test sample after the contacting, after growing the test sample in a suitable medium with the candidate compound under suitable conditions for a period of time between 1 to 5 days, and wherein the phenotype of the control sample is determined after growing the control sample in the suitable medium under the suitable conditions for an equivalent period of time.

27. The method of claim 1, wherein step (ii) comprises measuring the phenotypic response of the test sample after the contacting, after growing the test sample in a suitable medium with the candidate compound under suitable conditions for a period of time between 3 to 5 days or between 2 to 3 days, and wherein the phenotype of the control sample is determined after growing the control sample in the suitable medium under the suitable conditions for an equivalent period of time.

28. The method of claim 1, wherein step (ii) comprises measuring the phenotypic response of the test sample after the contacting, after growing the test sample in a suitable medium with the candidate compound under suitable conditions for a period of time between 1 to 10 days, and wherein the phenotype of the control sample is determined after growing the control sample in the suitable medium under the suitable conditions for an equivalent period of time.

29. The method of claim 1, wherein step (ii) comprises measuring the phenotypic response of the test sample after the contacting, after growing the test sample in a suitable medium with the candidate compound under suitable conditions for a period of time of less than 5 days, and wherein the phenotype of the control sample is determined after growing the control sample in the suitable medium under the suitable conditions for an equivalent period of time.

30. The method of claim 1, wherein step (ii) comprises measuring the phenotypic response of the test sample after the contacting, after growing the test sample in a suitable medium with the candidate compound under suitable conditions for a period of time of less than 4 days, and wherein the phenotype of the control sample is determined after growing the control sample in the suitable medium under the suitable conditions for an equivalent period of time.

31. The method of claim 1, wherein step (ii) comprises measuring the phenotypic response of the test sample after the contacting, after growing the test sample in a suitable medium with the candidate compound under suitable conditions for a period of time of less than 3 days, and wherein the phenotype of the control sample is determined after growing the control sample in the suitable medium under the suitable conditions for an equivalent period of time.

32. The method of claim 1, wherein step (ii) comprises obtaining measurements of any one or more of: sample length, sample width, sample shape, sample pigmentation, sample circularity, sample chlorophyll concentration, and / or cell number per sample.

33. The method of claim 32, wherein the measurements are digitally recorded.

34. The method of claim 1, wherein comparing the phenotypic response of the test sample to any of the control sample phenotypes comprises any one or more of: distributed random neighborhood embedding, generalized weighted least squares principal component analysis, minimum weighted chi-squared principal component analysis, minimum residual principal component analysis, principal axis common factor analysis, maximum likelihood common factor analysis, or weighted least squares common factor analysis.

35. The method of claim 1, wherein the candidate compound is selected as a potential herbicide using an artificial intelligence algorithm.

36. The method of claim 35, wherein step (ii) comprises: obtaining phenotypic measurements from the test sample and any of the control samples and thereby generating a dataset, and using at least 50% of the dataset as a training set for the artificial intelligence algorithm.

37. The method of claim 35, wherein step (ii) comprises: obtaining phenotypic measurements from the test sample and any of the control samples and thereby generating a dataset, and using at least 60% of the dataset as a training set for the artificial intelligence algorithm.

38. The method of claim 35, wherein step (ii) comprises: obtaining phenotypic measurements from the test sample and any of the control samples and thereby generating a dataset, and using at least 70% of the dataset as a training set for the artificial intelligence algorithm.

39. The method of claim 35, wherein step (ii) comprises: obtaining phenotypic measurements from the test sample and any of the control samples and thereby generating a dataset, and using at least 80% of the dataset as a training set for the artificial intelligence algorithm.

40. The method of claim 35, wherein step (ii) comprises: obtaining phenotypic measurements from the test sample and any of the control samples and thereby generating a dataset, and using at least 90% of the dataset as a training set for the artificial intelligence algorithm.

41. The method of claim 35, wherein step (ii) comprises: obtaining phenotypic measurements from the test sample and any of the control samples and thereby generating a dataset, and using at least 95% of the dataset as a training set for the artificial intelligence algorithm.

42. The method of claim 35, wherein step (ii) comprises: obtaining phenotypic measurements from the test sample and any of the control samples and thereby generating a dataset, and using 99% of the dataset as a training set for the artificial intelligence algorithm.

43. The method of any one of claims 36-42, wherein the control sample comprises a positive control sample and the artificial intelligence algorithm is used to predict a mode of action of any of the candidate compounds.

44. The method of claim 1, wherein the method further comprises the following step (iii): (a) contacting a candidate compound identified in steps (i) and (ii) as having herbicidal or plant growth regulating activity with a series of mutagenized samples comprising whole plants, spores, germinating spores, or vegetative propagules, wherein the test sample and mutagenized samples are from non-vascular plants of the same species; (b) extracting DNA from resistant mutagenized samples that survive the contacting in (a) or that do not exhibit growth abnormalities following the contacting in (a); (c) sequencing the genome or genomic portion of the resistant mutagenized samples, thereby obtaining mutagenized sample DNA sequences; (d) aligning the mutagenized sample DNA sequences obtained in (c) to a reference DNA sequence and identifying a first set of sequence mismatches between the mutagenized sample DNA sequences and the reference DNA sequence; (e) aligning DNA sequences from the first comparison DNA sequences to the reference DNA sequence and identifying a second set of sequence mismatches between the first comparison DNA sequences and the reference DNA sequence; and (f) filtering the first set of sequence mismatches against the second set of sequence mismatches to identify a first subset of sequence mismatches that are unique to the first set of sequence mismatches, wherein the first subset of sequence mismatches are candidate mutations that can confer resistance to herbicides or plant growth regulators; wherein the first comparison sample is from an independent sample that did not survive the contacting with the candidate compound or that exhibited growth abnormalities following the contacting with the candidate compound and belongs to the same genus as the resistant mutagenized sample, and wherein the reference DNA sequence is a known reference sequence for plants of the genus.

45. The method of claim 44, wherein the mutagenized samples comprise explants or protoplasts.

46. The method of claim 44 or 45, wherein the method further comprises: (e-i) aligning DNA sequences of a second comparison sample to the reference DNA sequence and identifying a third set of sequence mismatches between the second comparison sample and the reference DNA sequence; and (f) filtering the first set of sequence mismatches against the third set of sequence mismatches to facilitate identification of a second subset of sequence mismatches that are unique to the first set of sequence mismatches, and generating a third subset of sequence mismatches by filtering the first subset of sequence mismatches against the second subset of sequence mismatches, wherein the first and second subsets of sequence mismatches are candidate mutations that can confer resistance to herbicides or plant growth regulators; wherein the second comparison sample is from an independent sample that did not survive the contacting with the candidate compound or that exhibited growth abnormalities following the contacting with the candidate compound and belongs to the same genus as the mutagenized sample.

47. The method of claim 44 or 45, wherein the mutagenized sample is a Ml sample, wherein Ml is a symbol referring to the same plant population after exposure of the parental population to a mutagen.

48. The method of claim 44 or 45, wherein the mutagenized sample comprises a non- naturally occurring mutation.

49. The method of claim 44 or 45, wherein the method does not comprise the step of isolating.

50. The method of claim 44 or 45, wherein the alignment of (e) comprises aligning the DNA sequence of 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or more comparison samples to the reference DNA sequence and identifying a second set of sequence matches between the two sequences.

51. The method of claim 44 or 45, wherein the method further comprises filtering the candidate mutations with a biological filter.

52. The method of claim 44 or 45, wherein the mutagenized sample is haploid.

53. The method of claim 44 or 45, wherein the candidate mutation is in a gene encoding a protein targeted by the candidate compound identified as having herbicidal or plant growth regulating activity.

54. The method of claim 44 or 45, wherein step (iii) is implemented using a computer.

Images