Analysis of genome word frameworks on genomic methylation data

Abstract

Genome Word Framework (GWF) analysis of DNA methylation involves the analysis of methylation motifs and digital signal processing. GWFs are stretches of DNA sequence covering differentially methylated positions (DMPs). This analysis allows the identification of DNA sequence methylation motifs found in genes with potential epigenetic regulatory functions, such as those induced by environmental changes or disease. Analytical heuristics can be implemented and used to identify DNA sequences of methylation motifs with high-order DNA base interdependence on methylated cytosines and base distribution that is statistically non-random. These findings set the basis for further model predictions. For example, such model predictions can be used to identify and treat patients with autism, cancer, and other diseases that would benefit from early diagnosis.

Description

[Technical field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims priority under 35 U.S.C. § 119 of U.S. Provisional Patent Application No. 63 / 323,690, filed March 25, 2022, which is incorporated by reference in its entirety, including but not limited to the specification, claims, and abstract, and any drawings, tables, appendices, or figures thereof.

[0002] STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT This invention was made with Government support under Grant No. GM134056 awarded by the National Institutes of Health. The Government has certain rights in this invention.

[0003] The present disclosure relates generally to improvements in computer science that have application in any industry that can benefit from genetic, phenotypic, and / or DNA / RNA research. More specifically, but not exclusively, the present disclosure relates to genome word framework analysis of genomic methylation data. [Background technology]

[0004] The background discussion provided herein provides a context for the present disclosure. The work of the presently named inventors, and aspects of the description that may not otherwise be regarded as prior art at the time of filing, are not admitted, expressly or impliedly, to be prior art.

[0005] DNA sequences carry not only information about how to build proteins, but also regulatory information for organisms to survive and reproduce, including but not limited to epigenetic information that controls chromatin behavior.

[0006] The existence of many diseases and many phenotypes is poorly understood because the penetrance and expressivity of a gene's phenotype varies due to different combinations of modifier alleles present in one genetic background versus another.

[0007] Thus, there is a need in the art for an apparatus that addresses the possibility of overlapping two or more structured languages ​​across the nucleotide sequence of a genome. Summary of the Invention

[0008] The following objects, features, advantages, aspects, and / or embodiments are not exhaustive and do not limit the overall disclosure. No single embodiment is required to provide every object, feature, or advantage. Any of the objects, features, advantages, aspects, and / or embodiments disclosed herein may be integrated with one another, either fully or partially.

[0009] It is a primary object, feature, and / or advantage of the present disclosure to improve or overcome the deficiencies in the art.

[0010] It is a further object, feature, and / or advantage of the present disclosure to provide an extension for a general-purpose programming language or a statistical programming language. In one embodiment, the extension comprises an algorithm(s) for analyzing methylation signals on a stretch of DNA sequence. In one embodiment, the DNA sequence is characterized by (i) methylation information, and (ii) physicochemical information around each methylated cytosine. In one embodiment, the algorithm(s) comprises one or more functions capable of estimating a distance matrix on a set of selected regions of the DNA sequence, capable of analyzing hierarchical clusters on the set of selected regions, capable of classifying the set of selected regions into a specified number of clusters, and capable of aligning multiple DNA sequences from the clusters to methylation motifs.

[0011] In some embodiments, the extension is written in the R statistical language.

[0012] It is a further object, feature, and / or advantage of the present disclosure to reduce the computational load.For example, differentially methylated genes (DMGs) identified using methylation analysis can be incorporated into gene networks to identify network hubs via protein-protein interaction network analysis and weighted correlation network analysis.Weighted correlation network analysis based on prior knowledge of DMPs and DMGs is thus far unprecedented.

[0013] It is a further object, feature, and / or advantage of the present disclosure to provide a computerized heuristic, in some embodiments, that includes high order DNA base interdependencies for methylated cytosines and a statistically non-random distribution of bases.

[0014] According to some aspects of the present disclosure, the heuristics may include: (1) an estimated information divergence (ID) statistic (such as sum, average, or density) for each gene that has at least one DMP on the gene (on the gene body or on the promoter region); (2) a principal component analysis (PCA) in which the first k components with 1% or more of the total sample variance are considered in downstream analysis; (3) calculation of a correlation matrix with pairwise gene correlations expressed as a vector of PCs; (4) analysis of the correlation matrix for the network; and (5) the contribution of each gene to the discrimination of the phenotype evaluated in terms of the percentage of cumulative variance from the total sample variance that the gene carries.

[0015] According to some further aspects of the present disclosure, the ID is selected from the group consisting of Hellinger divergence / distance, J-divergence, total variation distance, and the like.

[0016] According to some further aspects of the present disclosure, PCA can be applied using the pcaLDA function described in the '986 patent. As a result, genes are represented as k-dimensional vectors of PCs, with the square of each coordinate having the contribution (in terms of variance) of the vector to the treatment discrimination from the control group. The correlation matrix can be mathematically equivalent (in terms of information) to a weighted correlation network (WCN).

[0017] According to some further aspects of the present disclosure, the WCN is analyzed as done for the network, which may be a PPI network. New knowledge is obtained from the WCN derived from raw methylation data, which does not depend on the inventors' ideas or biological knowledge about the genes represented in the network. The results from the WCN and PPI network are compared with the identified consistent relationships and contributions of epigenetic genes to phenotypes.

[0018] According to some further aspects of the present disclosure, the heuristic further comprises a measure calculated as the Euclidean norm of the gene represented as a vector of k PCs.

[0019] It is yet another object, feature, and / or advantage of the present disclosure to selectively construct motif libraries.For example, methylation motifs can be identified in all DMGs, providing the raw material for constructing motif libraries.These libraries can then serve as the basic data sets required to build predictive models with applications in plant science and biomedical research.

[0020] The genome word framework and genome methylation data disclosed herein can be used in a wide variety of applications. For example, such GWF-based model predictions can be used to identify and treat patients with autism, cancer, and other diseases that would benefit from early diagnosis. The models can also help provide further understanding in discovering the causes of (1) currently poorly understood phenotypes and (2) multifactorial diseases that appear to be caused by both genetic and environmental factors, such as diabetes and alcoholism.

[0021] The visual representation of the genome word framework analysis can be automatically and intuitively constructed to quickly convey meaning to people interpreting the same. Thus, at least one embodiment disclosed herein can include different aesthetic appearances. The decorative aspects included in such embodiments can help facilitate people's understanding of the potential relationship that genome methylation data has to applications in the physical world (e.g., phenotypes).

[0022] Methods can be implemented that facilitate the use, production, assembly, maintenance, and repair of libraries of DNA methylation motifs that achieve some or all of the foregoing objectives.

[0023] It is a further object, feature, and / or advantage of the present disclosure to provide a method for analyzing methylation signals on a stretch of DNA sequence. In some embodiments, the method comprises analyzing hierarchical clusters on regions of the DNA sequence, hierarchically classifying a set of selected regions into a specified number of clusters, aligning potential DNA sequence motifs from the clusters, and applying digital signal processing to the encoded methylation and physicochemical signals.

[0024] The creation and maintenance of the libraries can further be incorporated into an automated heuristic analysis process that continually refines and improves DNA base interdependencies for methylated cytosines until they achieve a base distribution that is statistically non-random.

[0025] These and / or other objects, features, advantages, aspects, and / or embodiments will become apparent to those skilled in the art after reviewing the following brief and detailed description of the drawings. Moreover, the present disclosure encompasses aspects and / or embodiments not expressly disclosed, but which can be understood from a reading of the present disclosure, including at least (a) combinations of the disclosed aspects and / or embodiments, and / or (b) reasonable modifications not shown or described.

[0026] Certain embodiments in which the present disclosure may be practiced are illustrated and described in detail, wherein like reference characters represent like components throughout the several views. The drawings are presented for illustrative purposes and may not be drawn to scale unless otherwise indicated. [Brief description of the drawings]

[0027] [Figure 1] Figure 1 shows the msh1 system, which is composed of four msh1-derived epigenetic states. In Arabidopsis, four distinct plant states result from MSH1 knockdown or knockout. States 1 and 2 result directly from disruption of msh1, resulting in a highly stress-responsive phenotype. State 1, a short day length, is variable, including a low frequency of a "perennial-like" phenotype. States 3 and 4 involve the interaction of msh1 modification and naive (wild-type) genome via grafting or crossing, resulting in a growth vigor phenotype. [Figure 2A]Figure 2 shows the characteristics of epi lines derived by crossing msh1 T-DNA mutants with isogenic wild type. 2A shows the phenotypes of different epi line F3 populations at 34 days after planting (DAP). Lines are derived from a WT x msh1 cross, Epi24 and Epi8 are derived from one parental cross, and Epi10 and Epi19 are derived from a second parental cross. 2B shows total leaf area (34 DAP), 2C shows days to bolting, and 2D shows seed weight (mg) for the four populations along with the WT control. 2B-2D show bars representing the mean ± SE. A Mann-Whitney U test with two-tailed alternative hypotheses was used to test the significance of the difference in the means between each Epi F3 population and the WT control. 2E shows the root phenotypes of the four different Epi F3 populations grown in sand (33 DAP). 2F shows total leaf area (33 DAP), dry leaf weight (mg), and dry root weight (mg) for the four populations grown on sand along with the WT control. 2F-2H show bars representing the mean ± SE. A Mann-Whitney U test with two-tailed alternative hypotheses was used to test the significance of the differences in means between each Epi F3 population and the WT control. Significance codes: *p<0.05, **p<0.01, ***p<0.001, ns-not significant. [Figure 2B]Figure 2 shows the characteristics of epi lines derived by crossing msh1 T-DNA mutants with isogenic wild type. 2A shows the phenotypes of different epi line F3 populations at 34 days after planting (DAP). Lines are derived from a WT x msh1 cross, Epi24 and Epi8 are derived from one parental cross, and Epi10 and Epi19 are derived from a second parental cross. 2B shows total leaf area (34 DAP), 2C shows days to bolting, and 2D shows seed weight (mg) for the four populations along with the WT control. 2B-2D show bars representing the mean ± SE. A Mann-Whitney U test with two-tailed alternative hypotheses was used to test the significance of the difference in the means between each Epi F3 population and the WT control. 2E shows the root phenotypes of the four different Epi F3 populations grown in sand (33 DAP). 2F shows total leaf area (33 DAP), dry leaf weight (mg), and dry root weight (mg) for the four populations grown on sand along with the WT control. 2F-2H show bars representing the mean ± SE. A Mann-Whitney U test with two-tailed alternative hypotheses was used to test the significance of the differences in means between each Epi F3 population and the WT control. Significance codes: *p<0.05, **p<0.01, ***p<0.001, ns-not significant. [Figure 2C]Figure 2 shows the characteristics of epi lines derived by crossing msh1 T-DNA mutants with isogenic wild type. 2A shows the phenotypes of different epi line F3 populations at 34 days after planting (DAP). Lines are derived from a WT x msh1 cross, Epi24 and Epi8 are derived from one parental cross, and Epi10 and Epi19 are derived from a second parental cross. 2B shows total leaf area (34 DAP), 2C shows days to bolting, and 2D shows seed weight (mg) for the four populations along with the WT control. 2B-2D show bars representing the mean ± SE. A Mann-Whitney U test with two-tailed alternative hypotheses was used to test the significance of the difference in the means between each Epi F3 population and the WT control. 2E shows the root phenotypes of the four different Epi F3 populations grown in sand (33 DAP). 2F shows total leaf area (33 DAP), dry leaf weight (mg), and dry root weight (mg) for the four populations grown on sand along with the WT control. 2F-2H show bars representing the mean ± SE. A Mann-Whitney U test with two-tailed alternative hypotheses was used to test the significance of the differences in means between each Epi F3 population and the WT control. Significance codes: *p<0.05, **p<0.01, ***p<0.001, ns-not significant. [Figure 2D]Figure 2 shows the characteristics of epi lines derived by crossing msh1 T-DNA mutants with isogenic wild type. 2A shows the phenotypes of different epi line F3 populations at 34 days after planting (DAP). Lines are derived from a WT x msh1 cross, Epi24 and Epi8 are derived from one parental cross, and Epi10 and Epi19 are derived from a second parental cross. 2B shows total leaf area (34 DAP), 2C shows days to bolting, and 2D shows seed weight (mg) for the four populations along with the WT control. 2B-2D show bars representing the mean ± SE. A Mann-Whitney U test with two-tailed alternative hypotheses was used to test the significance of the difference in the means between each Epi F3 population and the WT control. 2E shows the root phenotypes of the four different Epi F3 populations grown in sand (33 DAP). 2F shows total leaf area (33 DAP), dry leaf weight (mg), and dry root weight (mg) for the four populations grown on sand along with the WT control. 2F-2H show bars representing the mean ± SE. A Mann-Whitney U test with two-tailed alternative hypotheses was used to test the significance of the differences in means between each Epi F3 population and the WT control. Significance codes: *p<0.05, **p<0.01, ***p<0.001, ns-not significant. [Figure 2E]Figure 2 shows the characteristics of epi lines derived by crossing msh1 T-DNA mutants with isogenic wild type. 2A shows the phenotypes of different epi line F3 populations at 34 days after planting (DAP). Lines are derived from a WT x msh1 cross, Epi24 and Epi8 are derived from one parental cross, and Epi10 and Epi19 are derived from a second parental cross. 2B shows total leaf area (34 DAP), 2C shows days to bolting, and 2D shows seed weight (mg) for the four populations along with the WT control. 2B-2D show bars representing the mean ± SE. A Mann-Whitney U test with two-tailed alternative hypotheses was used to test the significance of the difference in the means between each Epi F3 population and the WT control. 2E shows the root phenotypes of the four different Epi F3 populations grown in sand (33 DAP). 2F shows total leaf area (33 DAP), dry leaf weight (mg), and dry root weight (mg) for the four populations grown on sand along with the WT control. 2F-2H show bars representing the mean ± SE. A Mann-Whitney U test with two-tailed alternative hypotheses was used to test the significance of the differences in means between each Epi F3 population and the WT control. Significance codes: *p<0.05, **p<0.01, ***p<0.001, ns-not significant. [Figure 2F]Figure 2 shows the characteristics of epi lines derived by crossing msh1 T-DNA mutants with isogenic wild type. 2A shows the phenotypes of different epi line F3 populations at 34 days after planting (DAP). Lines are derived from a WT x msh1 cross, Epi24 and Epi8 are derived from one parental cross, and Epi10 and Epi19 are derived from a second parental cross. 2B shows total leaf area (34 DAP), 2C shows days to bolting, and 2D shows seed weight (mg) for the four populations along with the WT control. 2B-2D show bars representing the mean ± SE. A Mann-Whitney U test with two-tailed alternative hypotheses was used to test the significance of the difference in the means between each Epi F3 population and the WT control. 2E shows the root phenotypes of the four different Epi F3 populations grown in sand (33 DAP). 2F shows total leaf area (33 DAP), dry leaf weight (mg), and dry root weight (mg) for the four populations grown on sand along with the WT control. 2F-2H show bars representing the mean ± SE. A Mann-Whitney U test with two-tailed alternative hypotheses was used to test the significance of the differences in means between each Epi F3 population and the WT control. Significance codes: *p<0.05, **p<0.01, ***p<0.001, ns-not significant. [Figure 2G]Figure 2 shows the characteristics of epi lines derived by crossing msh1 T-DNA mutants with isogenic wild type. 2A shows the phenotypes of different epi line F3 populations at 34 days after planting (DAP). Lines are derived from a WT x msh1 cross, Epi24 and Epi8 are derived from one parental cross, and Epi10 and Epi19 are derived from a second parental cross. 2B shows total leaf area (34 DAP), 2C shows days to bolting, and 2D shows seed weight (mg) for the four populations along with the WT control. 2B-2D show bars representing the mean ± SE. A Mann-Whitney U test with two-tailed alternative hypotheses was used to test the significance of the difference in the means between each Epi F3 population and the WT control. 2E shows the root phenotypes of the four different Epi F3 populations grown in sand (33 DAP). 2F shows total leaf area (33 DAP), dry leaf weight (mg), and dry root weight (mg) for the four populations grown on sand along with the WT control. 2F-2H show bars representing the mean ± SE. A Mann-Whitney U test with two-tailed alternative hypotheses was used to test the significance of the differences in means between each Epi F3 population and the WT control. Significance codes: *p<0.05, **p<0.01, ***p<0.001, ns-not significant. [Figure 2H]Figure 2 shows the characteristics of epi lines derived by crossing msh1 T-DNA mutants with isogenic wild type. 2A shows the phenotypes of different epi line F3 populations at 34 days after planting (DAP). Lines are derived from a WT x msh1 cross, Epi24 and Epi8 are derived from one parental cross, and Epi10 and Epi19 are derived from a second parental cross. 2B shows total leaf area (34 DAP), 2C shows days to bolting, and 2D shows seed weight (mg) for the four populations along with the WT control. 2B-2D show bars representing the mean ± SE. A Mann-Whitney U test with two-tailed alternative hypotheses was used to test the significance of the difference in the means between each Epi F3 population and the WT control. 2E shows the root phenotypes of the four different Epi F3 populations grown in sand (33 DAP). 2F shows total leaf area (33 DAP), dry leaf weight (mg), and dry root weight (mg) for the four populations grown on sand along with the WT control. 2F-2H show bars representing the mean ± SE. A Mann-Whitney U test with two-tailed alternative hypotheses was used to test the significance of the differences in means between each Epi F3 population and the WT control. Significance codes: *p<0.05, **p<0.01, ***p<0.001, ns-not significant. [Figure 3A] Tomato plants used for methylome sequencing are shown. 3A shows Rutgers wild type plants at 4 weeks. 3B-3D show Rutgers msh1 memory plants (state 2, right side of each panel) compared to WT at 4 weeks. 3E-3G show Rutgers msh1 RNAi transgenic plants (state 1, right side of each panel) compared to WT. 3H shows Epi lines with higher vegetative growth (three plants on the left), denoted as Epi line-high (Epi-H) in subsequent analyses, and epi lines with lower vegetative growth (three plants on the right), denoted as Epi line-low (Epi-L). Young leaves of 4-week-old plants were collected for whole genome bisulfite sequencing. [Figure 3B]Tomato plants used for methylome sequencing are shown. 3A shows Rutgers wild type plants at 4 weeks. 3B-3D show Rutgers msh1 memory plants (state 2, right side of each panel) compared to WT at 4 weeks. 3E-3G show Rutgers msh1 RNAi transgenic plants (state 1, right side of each panel) compared to WT. 3H shows Epi lines with higher vegetative growth (three plants on the left), denoted as Epi line-high (Epi-H) in subsequent analyses, and epi lines with lower vegetative growth (three plants on the right), denoted as Epi line-low (Epi-L). Young leaves of 4-week-old plants were collected for whole genome bisulfite sequencing. [Figure 3C] Tomato plants used for methylome sequencing are shown. 3A shows Rutgers wild type plants at 4 weeks. 3B-3D show Rutgers msh1 memory plants (state 2, right side of each panel) compared to WT at 4 weeks. 3E-3G show Rutgers msh1 RNAi transgenic plants (state 1, right side of each panel) compared to WT. 3H shows Epi lines with higher vegetative growth (three plants on the left), denoted as Epi line-high (Epi-H) in subsequent analyses, and epi lines with lower vegetative growth (three plants on the right), denoted as Epi line-low (Epi-L). Young leaves of 4-week-old plants were collected for whole genome bisulfite sequencing. [Figure 3D] Tomato plants used for methylome sequencing are shown. 3A shows Rutgers wild type plants at 4 weeks. 3B-3D show Rutgers msh1 memory plants (state 2, right side of each panel) compared to WT at 4 weeks. 3E-3G show Rutgers msh1 RNAi transgenic plants (state 1, right side of each panel) compared to WT. 3H shows Epi lines with higher vegetative growth (three plants on the left), denoted as Epi line-high (Epi-H) in subsequent analyses, and epi lines with lower vegetative growth (three plants on the right), denoted as Epi line-low (Epi-L). Young leaves of 4-week-old plants were collected for whole genome bisulfite sequencing. [Figure 3E] Tomato plants used for methylome sequencing are shown. 3A shows Rutgers wild type plants at 4 weeks. 3B-3D show Rutgers msh1 memory plants (state 2, right side of each panel) compared to WT at 4 weeks. 3E-3G show Rutgers msh1 RNAi transgenic plants (state 1, right side of each panel) compared to WT. 3H shows Epi lines with higher vegetative growth (three plants on the left), denoted as Epi line-high (Epi-H) in subsequent analyses, and epi lines with lower vegetative growth (three plants on the right), denoted as Epi line-low (Epi-L). Young leaves of 4-week-old plants were collected for whole genome bisulfite sequencing. [Figure 3F] Tomato plants used for methylome sequencing are shown. 3A shows Rutgers wild type plants at 4 weeks. 3B-3D show Rutgers msh1 memory plants (state 2, right side of each panel) compared to WT at 4 weeks. 3E-3G show Rutgers msh1 RNAi transgenic plants (state 1, right side of each panel) compared to WT. 3H shows Epi lines with higher vegetative growth (three plants on the left), denoted as Epi line-high (Epi-H) in subsequent analyses, and epi lines with lower vegetative growth (three plants on the right), denoted as Epi line-low (Epi-L). Young leaves of 4-week-old plants were collected for whole genome bisulfite sequencing. [Figure 3G] Tomato plants used for methylome sequencing are shown. 3A shows Rutgers wild type plants at 4 weeks. 3B-3D show Rutgers msh1 memory plants (state 2, right side of each panel) compared to WT at 4 weeks. 3E-3G show Rutgers msh1 RNAi transgenic plants (state 1, right side of each panel) compared to WT. 3H shows Epi lines with higher vegetative growth (three plants on the left), denoted as Epi line-high (Epi-H) in subsequent analyses, and epi lines with lower vegetative growth (three plants on the right), denoted as Epi line-low (Epi-L). Young leaves of 4-week-old plants were collected for whole genome bisulfite sequencing. [Figure 3H] Tomato plants used for methylome sequencing are shown. 3A shows Rutgers wild type plants at 4 weeks. 3B-3D show Rutgers msh1 memory plants (state 2, right side of each panel) compared to WT at 4 weeks. 3E-3G show Rutgers msh1 RNAi transgenic plants (state 1, right side of each panel) compared to WT. 3H shows Epi lines with higher vegetative growth (three plants on the left), denoted as Epi line-high (Epi-H) in subsequent analyses, and epi lines with lower vegetative growth (three plants on the right), denoted as Epi line-low (Epi-L). Young leaves of 4-week-old plants were collected for whole genome bisulfite sequencing. [Figure 4A] Revertant phenotypes in Arabidopsis and soybean are shown. 4A shows plant growth phenotypes of three F3 epi line populations, Epi10 induced by crossing to an msh1 T-DNA mutant, and Epi6 and Epi9 induced by crossing to an msh1 memory line. Dashed circles indicate putative revertants. Col-0 wild type and msh1 memory are shown as controls. 4B shows field-grown soybean (cv. Thorne) F3 epi lines showing evidence of spontaneous reversion (arrows). October harvest was followed by severe cold, and revertants showed cold tolerance similar to msh1 memory. 4C shows principal component analysis-linear discriminant analysis of methylome data from three epi line non-revertants and three revertant full-sib F3 progeny compared to three independently derived msh1 memory plants. [Figure 4B]Revertant phenotypes in Arabidopsis and soybean are shown. 4A shows plant growth phenotypes of three F3 epi line populations, Epi10 induced by crossing to an msh1 T-DNA mutant, and Epi6 and Epi9 induced by crossing to an msh1 memory line. Dashed circles indicate putative revertants. Col-0 wild type and msh1 memory are shown as controls. 4B shows field-grown soybean (cv. Thorne) F3 epi lines showing evidence of spontaneous reversion (arrows). October harvest was followed by severe cold, and revertants showed cold tolerance similar to msh1 memory. 4C shows principal component analysis-linear discriminant analysis of methylome data from three epi line non-revertants and three revertant full-sib F3 progeny compared to three independently derived msh1 memory plants. [Figure 4C] Revertant phenotypes in Arabidopsis and soybean are shown. 4A shows plant growth phenotypes of three F3 epi line populations, Epi10 induced by crossing to an msh1 T-DNA mutant, and Epi6 and Epi9 induced by crossing to an msh1 memory line. Dashed circles indicate putative revertants. Col-0 wild type and msh1 memory are shown as controls. 4B shows field-grown soybean (cv. Thorne) F3 epi lines showing evidence of spontaneous reversion (arrows). October harvest was followed by severe cold, and revertants showed cold tolerance similar to msh1 memory. 4C shows principal component analysis-linear discriminant analysis of methylome data from three epi line non-revertants and three revertant full-sib F3 progeny compared to three independently derived msh1 memory plants. [Figure 5-1] Figure 1 shows the relative DMP frequency between different Arabidopsis epilines. DMPs were assigned to genes, TE-associated, and other genomic regions. The centroid of the wild-type sample was used as the reference. The relative DMP frequency in each genomic feature was estimated as the number of DMPs divided by the number of total genomic cytosine positions in each genomic feature. DMPs were defined as "hyper" if the cytosine methylation level difference was greater than 0, and "hypo" if the methylation level difference was less than 0. [Figure 5-2] Figure 1 shows the relative DMP frequency between different Arabidopsis epilines. DMPs were assigned to genes, TE-associated, and other genomic regions. The centroid of the wild-type sample was used as the reference. The relative DMP frequency in each genomic feature was estimated as the number of DMPs divided by the number of total genomic cytosine positions in each genomic feature. DMPs were defined as "hyper" if the cytosine methylation level difference was greater than 0, and "hypo" if the methylation level difference was less than 0. [Figure 6A] The sum of hyper-DMP and hypo-DMP counts in epi lines are shown. 6A shows Arabidopsis epi lines vs. WT. 6B shows tomato epi lines low (BE) vs. high (GE). 6C shows soybean epi-F4 vs. epi-F6. Each bar represents a single plant in a given population. Cytosine context (CG, CHG, CHH) is shown separately for each plant. DMPs were defined as "hyper" if the cytosine methylation level difference was greater than 0 and "hypo" if the methylation level difference was less than 0. The centroid of the control was used as reference. [Figure 6B] The sum of hyper-DMP and hypo-DMP counts in epi lines are shown. 6A shows Arabidopsis epi lines vs. WT. 6B shows tomato epi lines low (BE) vs. high (GE). 6C shows soybean epi-F4 vs. epi-F6. Each bar represents a single plant in a given population. Cytosine context (CG, CHG, CHH) is shown separately for each plant. DMPs were defined as "hyper" if the cytosine methylation level difference was greater than 0 and "hypo" if the methylation level difference was less than 0. The centroid of the control was used as reference. [Figure 6C]The sum of hyper-DMP and hypo-DMP counts in epi lines are shown. 6A shows Arabidopsis epi lines vs. WT. 6B shows tomato epi lines low (BE) vs. high (GE). 6C shows soybean epi-F4 vs. epi-F6. Each bar represents a single plant in a given population. Cytosine context (CG, CHG, CHH) is shown separately for each plant. DMPs were defined as "hyper" if the cytosine methylation level difference was greater than 0 and "hypo" if the methylation level difference was less than 0. The centroid of the control was used as reference. [Figure 7A] Discrimination of methylation repatterning between different msh1 states is shown. 7A shows hierarchical clustering results using gene methylome data from three different msh1 states in Arabidopsis: msh1 null mutant (state 1), Col-0 / Col-0 msh1 grafted progeny (state 3), Epi F3 population (state 4), and associated Col-0 control. 7B shows hierarchical clustering results using individual plant (P) datasets from three different msh1 states in tomato: Rutgers MSH1-RNAi+ (state 1), Rutgers / Rutgers msh1-RNAi grafted (HEG) progeny (state 3), and Epi F3 populations (Epi-L and Epi-H) (state 4) along with associated Rutgers control. 7C shows hierarchical clustering results using TE DMP data from the same Arabidopsis dataset presented in 7A. Individuals were represented as vectors of the sum of Hellinger divergence (HD) at DMP positions within gene (7A, 7B) or TE (7C) regions. Hierarchical clustering was constructed using the Ward agglomeration method, and Hellinger divergence (HD) was calculated by using the centroid of the corresponding wild-type samples. The HD formula has been reported elsewhere. [Figure 7B]Discrimination of methylation repatterning between different msh1 states is shown. 7A shows hierarchical clustering results using gene methylome data from three different msh1 states in Arabidopsis: msh1 null mutant (state 1), Col-0 / Col-0 msh1 grafted progeny (state 3), Epi F3 population (state 4), and associated Col-0 control. 7B shows hierarchical clustering results using individual plant (P) datasets from three different msh1 states in tomato: Rutgers MSH1-RNAi+ (state 1), Rutgers / Rutgers msh1-RNAi grafted (HEG) progeny (state 3), and Epi F3 populations (Epi-L and Epi-H) (state 4) along with associated Rutgers control. 7C shows hierarchical clustering results using TE DMP data from the same Arabidopsis dataset presented in 7A. Individuals were represented as vectors of the sum of Hellinger divergence (HD) at DMP positions within gene (7A, 7B) or TE (7C) regions. Hierarchical clustering was constructed using the Ward agglomeration method, and Hellinger divergence (HD) was calculated by using the centroid of the corresponding wild-type samples. The HD formula has been reported elsewhere. [Figure 7C]Discrimination of methylation repatterning between different msh1 states is shown. 7A shows hierarchical clustering results using gene methylome data from three different msh1 states in Arabidopsis: msh1 null mutant (state 1), Col-0 / Col-0 msh1 grafted progeny (state 3), Epi F3 population (state 4), and associated Col-0 control. 7B shows hierarchical clustering results using individual plant (P) datasets from three different msh1 states in tomato: Rutgers MSH1-RNAi+ (state 1), Rutgers / Rutgers msh1-RNAi grafted (HEG) progeny (state 3), and Epi F3 populations (Epi-L and Epi-H) (state 4) along with associated Rutgers control. 7C shows hierarchical clustering results using TE DMP data from the same Arabidopsis dataset presented in 7A. Individuals were represented as vectors of the sum of Hellinger divergence (HD) at DMP positions within gene (7A, 7B) or TE (7C) regions. Hierarchical clustering was constructed using the Ward agglomeration method, and Hellinger divergence (HD) was calculated by using the centroid of the corresponding wild-type samples. The HD formula has been reported elsewhere. [Figure 8A]8A shows the significant enriched GO pathways by DMG analysis in epi lines. 8A shows the enriched GO pathways shared by Epi8 and Epi24 in Arabidopsis. 8B shows the enriched GO pathways shared by Epi10 and Epi19 in Arabidopsis. 8C shows the enriched GO pathways identified by F1 hybrid DMGs from C24×Ler crosses by our analysis. The original methylation data for F1 hybrids was reported previously. 8D shows the enriched GO pathways derived from DMG analysis in the comparison of Epi line low vs. Epi line high in tomato. 8E shows the enriched GO pathways based on DMG analysis in the comparison of Epi-F4 vs. Epi-F6 in soybean. For 8A-8C, the bar graphs represent the number of DMGs and the dotted lines represent the -log10(p-value) for enrichment test. For 8D-8E, the bar graphs represent the number of DMGs and the dotted lines represent the enrichment fold of each enriched pathway. Network enrichment analysis test (NEAT) was used to identify enriched GO pathways. p-values ​​were calculated using a one-sided hypergeometric distribution test implemented in NEAT. Thus, a complete list of enriched pathways can be provided. [Figure 8B]8A shows the significant enriched GO pathways by DMG analysis in epi lines. 8A shows the enriched GO pathways shared by Epi8 and Epi24 in Arabidopsis. 8B shows the enriched GO pathways shared by Epi10 and Epi19 in Arabidopsis. 8C shows the enriched GO pathways identified by F1 hybrid DMGs from C24×Ler crosses by our analysis. The original methylation data for F1 hybrids was reported previously. 8D shows the enriched GO pathways derived from DMG analysis in the comparison of Epi line low vs. Epi line high in tomato. 8E shows the enriched GO pathways based on DMG analysis in the comparison of Epi-F4 vs. Epi-F6 in soybean. For 8A-8C, the bar graphs represent the number of DMGs and the dotted lines represent the -log10(p-value) for enrichment test. For 8D-8E, the bar graphs represent the number of DMGs and the dotted lines represent the enrichment fold of each enriched pathway. Network enrichment analysis test (NEAT) was used to identify enriched GO pathways. p-values ​​were calculated using a one-sided hypergeometric distribution test implemented in NEAT. Thus, a complete list of enriched pathways can be provided. [Figure 8C]8A shows the significant enriched GO pathways by DMG analysis in epi lines. 8A shows the enriched GO pathways shared by Epi8 and Epi24 in Arabidopsis. 8B shows the enriched GO pathways shared by Epi10 and Epi19 in Arabidopsis. 8C shows the enriched GO pathways identified by F1 hybrid DMGs from C24×Ler crosses by our analysis. The original methylation data for F1 hybrids was reported previously. 8D shows the enriched GO pathways derived from DMG analysis in the comparison of Epi line low vs. Epi line high in tomato. 8E shows the enriched GO pathways based on DMG analysis in the comparison of Epi-F4 vs. Epi-F6 in soybean. For 8A-8C, the bar graphs represent the number of DMGs and the dotted lines represent the -log10(p-value) for enrichment test. For 8D-8E, the bar graphs represent the number of DMGs and the dotted lines represent the enrichment fold of each enriched pathway. Network enrichment analysis test (NEAT) was used to identify enriched GO pathways. p-values ​​were calculated using a one-sided hypergeometric distribution test implemented in NEAT. Thus, a complete list of enriched pathways can be provided. [Figure 8D]8A shows the significant enriched GO pathways by DMG analysis in epi lines. 8A shows the enriched GO pathways shared by Epi8 and Epi24 in Arabidopsis. 8B shows the enriched GO pathways shared by Epi10 and Epi19 in Arabidopsis. 8C shows the enriched GO pathways identified by F1 hybrid DMGs from C24×Ler crosses by our analysis. The original methylation data for F1 hybrids was reported previously. 8D shows the enriched GO pathways derived from DMG analysis in the comparison of Epi line low vs. Epi line high in tomato. 8E shows the enriched GO pathways based on DMG analysis in the comparison of Epi-F4 vs. Epi-F6 in soybean. For 8A-8C, the bar graphs represent the number of DMGs and the dotted lines represent the -log10(p-value) for enrichment test. For 8D-8E, the bar graphs represent the number of DMGs and the dotted lines represent the enrichment fold of each enriched pathway. Network enrichment analysis test (NEAT) was used to identify enriched GO pathways. p-values ​​were calculated using a one-sided hypergeometric distribution test implemented in NEAT. Thus, a complete list of enriched pathways can be provided. [Figure 8E]8A shows the significant enriched GO pathways by DMG analysis in epi lines. 8A shows the enriched GO pathways shared by Epi8 and Epi24 in Arabidopsis. 8B shows the enriched GO pathways shared by Epi10 and Epi19 in Arabidopsis. 8C shows the enriched GO pathways identified by F1 hybrid DMGs from C24×Ler crosses by our analysis. The original methylation data for F1 hybrids was reported previously. 8D shows the enriched GO pathways derived from DMG analysis in the comparison of Epi line low vs. Epi line high in tomato. 8E shows the enriched GO pathways based on DMG analysis in the comparison of Epi-F4 vs. Epi-F6 in soybean. For 8A-8C, the bar graphs represent the number of DMGs and the dotted lines represent the -log10(p-value) for enrichment test. For 8D-8E, the bar graphs represent the number of DMGs and the dotted lines represent the enrichment fold of each enriched pathway. Network enrichment analysis test (NEAT) was used to identify enriched GO pathways. p-values ​​were calculated using a one-sided hypergeometric distribution test implemented in NEAT. Thus, a complete list of enriched pathways can be provided. [Figure 9A] Figure 9 shows significant enriched GO pathways based on DEG analysis of epi lineages in Arabidopsis. 9A shows enriched GO pathways from DEGs shared by Epi8 and Epi10 leaf tissues. 9B shows enriched GO pathways from DEGs shared by Epi8 and Epi24 inflorescence stem tissues, and 9C shows root tissues. Bars represent DEG numbers, and dotted lines represent enrichment folds of each enriched pathway. One-tailed Fisher's exact test was used to calculate FDR, which is implemented in DAVID GO. Thus, a complete list of enriched pathways can be provided. [Figure 9B]Figure 9 shows significant enriched GO pathways based on DEG analysis of epi lineages in Arabidopsis. 9A shows enriched GO pathways from DEGs shared by Epi8 and Epi10 leaf tissues. 9B shows enriched GO pathways from DEGs shared by Epi8 and Epi24 inflorescence stem tissues, and 9C shows root tissues. Bars represent DEG numbers, and dotted lines represent enrichment folds of each enriched pathway. One-tailed Fisher's exact test was used to calculate FDR, which is implemented in DAVID GO. Thus, a complete list of enriched pathways can be provided. [Figure 9C] Figure 9 shows significant enriched GO pathways based on DEG analysis of epi lineages in Arabidopsis. 9A shows enriched GO pathways from DEGs shared by Epi8 and Epi10 leaf tissues. 9B shows enriched GO pathways from DEGs shared by Epi8 and Epi24 inflorescence stem tissues, and 9C shows root tissues. Bars represent DEG numbers, and dotted lines represent enrichment folds of each enriched pathway. One-tailed Fisher's exact test was used to calculate FDR, which is implemented in DAVID GO. Thus, a complete list of enriched pathways can be provided. [Figure 10A] Figure 1 shows a protein-protein interaction (PPI) network of hubs from a subset of network-associated DMGs and DEGs in full-sib lines of Arabidopsis Epi8 and Epi24. Epi8 and Epi24 represent progeny lines derived from the same cross, with Epi8 exhibiting enhanced vegetative growth rate and Epi24 being significantly enhanced in seed yield. The main sub-network of hubs was obtained by applying machine learning K-means clustering on a set of 3647 (Epi8, 10A) and 3523 (Epi24, 10B) network-associated DMGs and DEGs identified in the Arabidopsis epi line vs. WT comparison. The analysis yielded 153 (10A) and 346 (10B) hub genes, forming closely related sub-networks. GO network enrichment analysis from the String application in Cytoscape was used to identify enriched gene functional pathways in the network. The size of each node is proportional to its node degree value, and the label font size is proportional to its betweenness centrality. [Figure 10B] Figure 1 shows a protein-protein interaction (PPI) network of hubs from a subset of network-associated DMGs and DEGs in full-sib lines of Arabidopsis Epi8 and Epi24. Epi8 and Epi24 represent progeny lines derived from the same cross, with Epi8 exhibiting enhanced vegetative growth rate and Epi24 being significantly enhanced in seed yield. The main sub-network of hubs was obtained by applying machine learning K-means clustering on a set of 3647 (Epi8, 10A) and 3523 (Epi24, 10B) network-associated DMGs and DEGs identified in the Arabidopsis epi line vs. WT comparison. The analysis yielded 153 (10A) and 346 (10B) hub genes, forming closely related sub-networks. GO network enrichment analysis from the String application in Cytoscape was used to identify enriched gene functional pathways in the network. The size of each node is proportional to its node degree value, and the label font size is proportional to its betweenness centrality. [Figure 11A]11A shows the protein-protein interaction (PPI) network of hubs derived from a subset of network-associated DMGs and DEGs in soybean epi lines and tomato grafted progeny (HEG) lines. 11A shows the primary sub-network of hubs obtained by application of the machine learning K-means clustering algorithm on the set of 1264 network-associated DMGs and DEGs identified in the comparison of soybean Epi-F6 vs. Epi-F4 (state 4). This analysis yielded 421 hub genes that form closely related sub-networks. GO network enrichment analysis from the string application in Cytoscape was used to identify enriched gene functional pathways in the network. 109 genes related to gene expression regulation were selected and presented. 11B shows the main sub-network of hubs obtained by application of the machine learning K-means clustering algorithm on the set of 3173 network-associated DMGs and DEGs identified in the comparison of tomato Rutgers / Rutgersmsh1-RNAi vs. Rutgers / Rutgers grafted progeny (HEG) lines (state 3). The analysis yielded 305 hub genes that formed a subnetwork. GO network enrichment analysis from the String application in Cytoscape was used to identify enriched gene functional pathways in the network. 126 genes related to gene expression regulation were selected and presented. For both 11A and 11B, the size of each node is proportional to its value of node degree and the label font size is proportional to its betweenness centrality. The datasets of soybean DEGs and tomato HEGs (DMGs and DEGs) were reported previously. [Figure 11B]11A shows the protein-protein interaction (PPI) network of hubs derived from a subset of network-associated DMGs and DEGs in soybean epi lines and tomato grafted progeny (HEG) lines. 11A shows the primary sub-network of hubs obtained by application of the machine learning K-means clustering algorithm on the set of 1264 network-associated DMGs and DEGs identified in the comparison of soybean Epi-F6 vs. Epi-F4 (state 4). This analysis yielded 421 hub genes that form closely related sub-networks. GO network enrichment analysis from the string application in Cytoscape was used to identify enriched gene functional pathways in the network. 109 genes related to gene expression regulation were selected and presented. 11B shows the main sub-network of hubs obtained by application of the machine learning K-means clustering algorithm on the set of 3173 network-associated DMGs and DEGs identified in the comparison of tomato Rutgers / Rutgersmsh1-RNAi vs. Rutgers / Rutgers grafted progeny (HEG) lines (state 3). The analysis yielded 305 hub genes that formed a subnetwork. GO network enrichment analysis from the String application in Cytoscape was used to identify enriched gene functional pathways in the network. 126 genes related to gene expression regulation were selected and presented. For both 11A and 11B, the size of each node is proportional to its value of node degree and the label font size is proportional to its betweenness centrality. The datasets of soybean DEGs and tomato HEGs (DMGs and DEGs) were reported previously. [Figure 12A]12A shows the relationship of the 871 core hub genes to four different msh1 derived states and biologically meaningful core networks. 12A shows a Venn diagram of Arabidopsis DMGs identified from four different msh1 derived states (Col-0 genetic background): msh1 mutant (state 1), msh1 memory (state 2), grafted progeny (HEG, state 3), and epi line (Epi24, state 4). 12B shows an overview of the PPI network and the individual 871 core hub genes. 12C shows the hierarchical clustering of the individual plant datasets from the four different msh1 derived states based on the sum of the Bayesian methylation level differences of DMPs across the 871 core genes from 12A. The hierarchical clustering was constructed using the Ward agglomeration method. The Bayesian methylation level differences were calculated as previously described. 12D shows the main sub-network of hubs obtained by application of the machine learning K-means clustering algorithm on the set of 871 core genes from 12A. GO network enrichment analysis from the string application in Cytoscape was used to identify enriched gene functional pathways in the network. 67 genes involved in the enriched network were identified. This set of 67 genes provided RdDM candidate target genes for further study. For both 12B and 12D, the size of each node is proportional to its value of node degree and the label font size is proportional to its betweenness centrality. [Figure 12B]12A shows the relationship of the 871 core hub genes to four different msh1 derived states and biologically meaningful core networks. 12A shows a Venn diagram of Arabidopsis DMGs identified from four different msh1 derived states (Col-0 genetic background): msh1 mutant (state 1), msh1 memory (state 2), grafted progeny (HEG, state 3), and epi line (Epi24, state 4). 12B shows an overview of the PPI network and the individual 871 core hub genes. 12C shows the hierarchical clustering of the individual plant datasets from the four different msh1 derived states based on the sum of the Bayesian methylation level differences of DMPs across the 871 core genes from 12A. The hierarchical clustering was constructed using the Ward agglomeration method. The Bayesian methylation level differences were calculated as previously described. 12D shows the main sub-network of hubs obtained by application of the machine learning K-means clustering algorithm on the set of 871 core genes from 12A. GO network enrichment analysis from the string application in Cytoscape was used to identify enriched gene functional pathways in the network. 67 genes involved in the enriched network were identified. This set of 67 genes provided RdDM candidate target genes for further study. For both 12B and 12D, the size of each node is proportional to its value of node degree and the label font size is proportional to its betweenness centrality. [Figure 12C]12A shows the relationship of the 871 core hub genes to four different msh1 derived states and biologically meaningful core networks. 12A shows a Venn diagram of Arabidopsis DMGs identified from four different msh1 derived states (Col-0 genetic background): msh1 mutant (state 1), msh1 memory (state 2), grafted progeny (HEG, state 3), and epi line (Epi24, state 4). 12B shows an overview of the PPI network and the individual 871 core hub genes. 12C shows the hierarchical clustering of the individual plant datasets from the four different msh1 derived states based on the sum of the Bayesian methylation level differences of DMPs across the 871 core genes from 12A. The hierarchical clustering was constructed using the Ward agglomeration method. The Bayesian methylation level differences were calculated as previously described. 12D shows the main sub-network of hubs obtained by application of the machine learning K-means clustering algorithm on the set of 871 core genes from 12A. GO network enrichment analysis from the string application in Cytoscape was used to identify enriched gene functional pathways in the network. 67 genes involved in the enriched network were identified. This set of 67 genes provided RdDM candidate target genes for further study. For both 12B and 12D, the size of each node is proportional to its value of node degree and the label font size is proportional to its betweenness centrality. [Figure 12D]12A shows the relationship of the 871 core hub genes to four different msh1 derived states and biologically meaningful core networks. 12A shows a Venn diagram of Arabidopsis DMGs identified from four different msh1 derived states (Col-0 genetic background): msh1 mutant (state 1), msh1 memory (state 2), grafted progeny (HEG, state 3), and epi line (Epi24, state 4). 12B shows an overview of the PPI network and the individual 871 core hub genes. 12C shows the hierarchical clustering of the individual plant datasets from the four different msh1 derived states based on the sum of the Bayesian methylation level differences of DMPs across the 871 core genes from 12A. The hierarchical clustering was constructed using the Ward agglomeration method. The Bayesian methylation level differences were calculated as previously described. 12D shows the main sub-network of hubs obtained by application of the machine learning K-means clustering algorithm on the set of 871 core genes from 12A. GO network enrichment analysis from the string application in Cytoscape was used to identify enriched gene functional pathways in the network. 67 genes involved in the enriched network were identified. This set of 67 genes provided RdDM candidate target genes for further study. For both 12B and 12D, the size of each node is proportional to its value of node degree and the label font size is proportional to its betweenness centrality. [Figure 13A]Figure 13 shows an investigation of gene methylation repatterning within candidate RdDM target genes identifying four msh1-derived states in Arabidopsis. 13A shows two of the putative cluster motifs identified based on differential gene methylation across the four msh1-derived states. Hierarchical clustering on a set of 14 bp regions encompassing DMPs from 7 of the 67 core hub loci (TOR, SYD, NRPB1, NRPD1, UBP26, SUVR5, UPF1), followed by DNA multiple sequence alignment of each cluster, allowed the identification of methylation motifs. 13B shows differential methylation levels on gene body DMPs within motif cluster 11 of the putative RdDM target gene UBP26. Variation in motif methylation repatterning in DMPs is shown with chromosome and location. Individual detected methylation changes are shown as cross-hatched dots for each plant assayed in each msh1 state, with positives indicating DMPs and negatives indicating undetected methylation changes. Each row represents a single plant data set. 13C shows sample DMPs within motif cluster 11 in UBP26 and UPF1 showing dcl2 / dcl3 / dcl4 sensitivity in grafted rootstocks (state 1) and grafted progeny (state 3) from contrasting mutant rootstock experiments. Note that dcl2 / dcl3 / dcl4 effects can only be assayed on msh1 and grafted progeny data. 13D shows sample putative cluster motifs identified based on differential gene methylation across four msh1-derived states in an analysis of 67 core hub loci, followed by DNA multiple sequence alignments of each cluster for methylation motifs. [Figure 13B]Figure 13 shows an investigation of gene methylation repatterning within candidate RdDM target genes identifying four msh1-derived states in Arabidopsis. 13A shows two of the putative cluster motifs identified based on differential gene methylation across the four msh1-derived states. Hierarchical clustering on a set of 14 bp regions encompassing DMPs from 7 of the 67 core hub loci (TOR, SYD, NRPB1, NRPD1, UBP26, SUVR5, UPF1), followed by DNA multiple sequence alignment of each cluster, allowed the identification of methylation motifs. 13B shows differential methylation levels on gene body DMPs within motif cluster 11 of the putative RdDM target gene UBP26. Variation in motif methylation repatterning in DMPs is shown with chromosome and location. Individual detected methylation changes are shown as cross-hatched dots for each plant assayed in each msh1 state, with positives indicating DMPs and negatives indicating undetected methylation changes. Each row represents a single plant data set. 13C shows sample DMPs within motif cluster 11 in UBP26 and UPF1 showing dcl2 / dcl3 / dcl4 sensitivity in grafted rootstocks (state 1) and grafted progeny (state 3) from contrasting mutant rootstock experiments. Note that dcl2 / dcl3 / dcl4 effects can only be assayed on msh1 and grafted progeny data. 13D shows sample putative cluster motifs identified based on differential gene methylation across four msh1-derived states in an analysis of 67 core hub loci, followed by DNA multiple sequence alignments of each cluster for methylation motifs. [Figure 13C]Figure 13 shows an investigation of gene methylation repatterning within candidate RdDM target genes identifying four msh1-derived states in Arabidopsis. 13A shows two of the putative cluster motifs identified based on differential gene methylation across the four msh1-derived states. Hierarchical clustering on a set of 14 bp regions encompassing DMPs from 7 of the 67 core hub loci (TOR, SYD, NRPB1, NRPD1, UBP26, SUVR5, UPF1), followed by DNA multiple sequence alignment of each cluster, allowed the identification of methylation motifs. 13B shows differential methylation levels on gene body DMPs within motif cluster 11 of the putative RdDM target gene UBP26. Variation in motif methylation repatterning in DMPs is shown with chromosome and location. Individual detected methylation changes are shown as cross-hatched dots for each plant assayed in each msh1 state, with positives indicating DMPs and negatives indicating undetected methylation changes. Each row represents a single plant data set. 13C shows sample DMPs within motif cluster 11 in UBP26 and UPF1 showing dcl2 / dcl3 / dcl4 sensitivity in grafted rootstocks (state 1) and grafted progeny (state 3) from contrasting mutant rootstock experiments. Note that dcl2 / dcl3 / dcl4 effects can only be assayed on msh1 and grafted progeny data. 13D shows sample putative cluster motifs identified based on differential gene methylation across four msh1-derived states in an analysis of 67 core hub loci, followed by DNA multiple sequence alignments of each cluster for methylation motifs. [Figure 13D-1]Figure 13 shows an investigation of gene methylation repatterning within candidate RdDM target genes identifying four msh1-derived states in Arabidopsis. 13A shows two of the putative cluster motifs identified based on differential gene methylation across the four msh1-derived states. Hierarchical clustering on a set of 14 bp regions encompassing DMPs from 7 of the 67 core hub loci (TOR, SYD, NRPB1, NRPD1, UBP26, SUVR5, UPF1), followed by DNA multiple sequence alignment of each cluster, allowed the identification of methylation motifs. 13B shows differential methylation levels on gene body DMPs within motif cluster 11 of the putative RdDM target gene UBP26. Variation in motif methylation repatterning in DMPs is shown with chromosome and location. Individual detected methylation changes are shown as cross-hatched dots for each plant assayed in each msh1 state, with positives indicating DMPs and negatives indicating undetected methylation changes. Each row represents a single plant data set. 13C shows sample DMPs within motif cluster 11 in UBP26 and UPF1 showing dcl2 / dcl3 / dcl4 sensitivity in grafted rootstocks (state 1) and grafted progeny (state 3) from contrasting mutant rootstock experiments. Note that dcl2 / dcl3 / dcl4 effects can only be assayed on msh1 and grafted progeny data. 13D shows sample putative cluster motifs identified based on differential gene methylation across four msh1-derived states in an analysis of 67 core hub loci, followed by DNA multiple sequence alignments of each cluster for methylation motifs. [Figure 13D-2]Figure 13 shows an investigation of gene methylation repatterning within candidate RdDM target genes identifying four msh1-derived states in Arabidopsis. 13A shows two of the putative cluster motifs identified based on differential gene methylation across the four msh1-derived states. Hierarchical clustering on a set of 14 bp regions encompassing DMPs from 7 of the 67 core hub loci (TOR, SYD, NRPB1, NRPD1, UBP26, SUVR5, UPF1), followed by DNA multiple sequence alignment of each cluster, allowed the identification of methylation motifs. 13B shows differential methylation levels on gene body DMPs within motif cluster 11 of the putative RdDM target gene UBP26. Variation in motif methylation repatterning in DMPs is shown with chromosome and location. Individual detected methylation changes are shown as cross-hatched dots for each plant assayed in each msh1 state, with positives indicating DMPs and negatives indicating undetected methylation changes. Each row represents a single plant data set. 13C shows sample DMPs within motif cluster 11 in UBP26 and UPF1 showing dcl2 / dcl3 / dcl4 sensitivity in grafted rootstocks (state 1) and grafted progeny (state 3) from contrasting mutant rootstock experiments. Note that dcl2 / dcl3 / dcl4 effects can only be assayed on msh1 and grafted progeny data. 13D shows sample putative cluster motifs identified based on differential gene methylation across four msh1-derived states in an analysis of 67 core hub loci, followed by DNA multiple sequence alignments of each cluster for methylation motifs. [Figure 14A]Exemplary cluster motifs encompassing DMPs on gene AT1G50030 (TOR) in Arabidopsis thaliana are shown in Figure 14A. Figure 14B shows the cluster motif for cluster 9. Figure 14C shows the cluster motif for cluster 12. Figure 14D shows the cluster motif for cluster 17. Individual consensus nucleotides were evident at variable distances from the target cytosine, nucleotide 7 (on the "+" or "-" strand) within each motif. [Figure 14B] Exemplary cluster motifs encompassing DMPs on gene AT1G50030 (TOR) in Arabidopsis thaliana are shown in Figure 14A. Figure 14B shows the cluster motif for cluster 9. Figure 14C shows the cluster motif for cluster 12. Figure 14D shows the cluster motif for cluster 17. Individual consensus nucleotides were evident at variable distances from the target cytosine, nucleotide 7 (on the "+" or "-" strand) within each motif. [Figure 14C] Exemplary cluster motifs encompassing DMPs on gene AT1G50030 (TOR) in Arabidopsis thaliana are shown in Figure 14A. Figure 14B shows the cluster motif for cluster 9. Figure 14C shows the cluster motif for cluster 12. Figure 14D shows the cluster motif for cluster 17. Individual consensus nucleotides were evident at variable distances from the target cytosine, nucleotide 7 (on the "+" or "-" strand) within each motif. [Figure 14D]Exemplary cluster motifs encompassing DMPs on gene AT1G50030 (TOR) in Arabidopsis thaliana are shown in Figure 14A. Figure 14B shows the cluster motif for cluster 9. Figure 14C shows the cluster motif for cluster 12. Figure 14D shows the cluster motif for cluster 17. Individual consensus nucleotides were evident at variable distances from the target cytosine, nucleotide 7 (on the "+" or "-" strand) within each motif. [Figure 15A] Power spectral densities of control (15A) and treatment (msh1 mutant, dwarf phenotype) (15B) in A. thaliana. Dots help visually identify peaks where spectral differences between control and treatment are found. [Figure 15B] Power spectral densities of control (15A) and treatment (msh1 mutant, dwarf phenotype) (15B) in A. thaliana. Dots help visually identify peaks where spectral differences between control and treatment are found. [Figure 16A] Spectrograms of AT1G50030 (TOR) in A. thaliana. In both 16A and 16B, the upper panel is control and the lower panel is treatment (dwarf phenotype). 16A is the spectrogram of the first region and 16B is the spectrogram of the third region. The methylation signal resolves the power spectral energy around a periodicity at approximately one-third frequency (white dashed line). [Figure 16B] Spectrograms of AT1G50030 (TOR) in A. thaliana. In both 16A and 16B, the upper panel is control and the lower panel is treatment (dwarf phenotype). 16A is the spectrogram of the first region and 16B is the spectrogram of the third region. The methylation signal resolves the power spectral energy around a periodicity at approximately one-third frequency (white dashed line). [Figure 17A]Wavelet power spectra (WPS) of regions AT1G50030.1 (17A), AT1G50030.2 (17B), and AT1G50030.3 (17C) are shown in the control (left panel, respectively) and treated (dwarf phenotype) (right panel, respectively). Differences observed in the WPS between control and treated groups corresponded to which base was changed. [Figure 17B] Wavelet power spectra (WPS) of regions AT1G50030.1 (17A), AT1G50030.2 (17B), and AT1G50030.3 (17C) are shown in the control (left panel, respectively) and treated (dwarf phenotype) (right panel, respectively). Differences observed in the WPS between control and treated groups corresponded to which base was changed. [Figure 17C] Wavelet power spectra (WPS) of regions AT1G50030.1 (17A), AT1G50030.2 (17B), and AT1G50030.3 (17C) are shown in the control (left panel, respectively) and treated (dwarf phenotype) (right panel, respectively). Differences observed in the WPS between control and treated groups corresponded to which base was changed. [Figure 18] Correlograms based on the WPS of 17A are shown. Control vs. control is shown in the upper panel and control vs. treated (dwarf phenotype) is shown in the lower panel. In addition to the expected correlation decomposition in the region 55-60 bits, correlations between energy spectra in the regions 30-35 bits and 55-60 bits are lost in control vs. treated. [Figure 19] Sub-networks of 81 DMG hubs are shown. A PPI network was constructed on the set of 751 DMGs identified by principal component analysis and analyzed with the STRING Cytoscape App. Circles without cross-hatching highlight DMGs involved in biological processes including nervous system, nervous system development, synapse, neuronal projection, ion transport, CNS disease, axon guidance, neurogenesis, ion / cation binding, ion / cation transmembrane transport, voltage-gated channels and TRP channels. [Figure 20]FIG. 19 shows network enrichment analysis identified in the main network of DMG-hubs from FIG.

[0028] Those skilled in the art need not view the nearly infinite number of different permutations of the features described in the following detailed description in an isolated drawing(s) to facilitate understanding of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0029] The present disclosure is not limited to what is described herein. Mechanical, electrical, chemical, procedural, and / or other changes may be made without departing from the spirit or scope of the present disclosure. Unless otherwise indicated, no feature shown or described is essential to enable the basic operation of the present disclosure.

[0030] Genomic word framework (GWF) analysis of DNA methylation involves the analysis of methylation motifs and digital signal processing. GWFs are stretches of DNA sequence that cover differentially methylated positions (DMPs). However, GWFs should not be confused with the concept of DNA sequence motifs. A word framework (WF) may contain one or more motifs. That is, a "sentence of WF" is also a GWF. DMPs can be identified by methylation analysis using extensions in statistical programming languages, such as the R package described by the inventors of the present application in U.S. Pat. No. 10,913,986. This analysis allows the identification of DNA sequence methylation motifs found in genes with potential epigenetic regulatory functions, including those induced by environmental changes or disease.

[0031] One possible embodiment of the analytical heuristics described herein is implemented in an R package named GenomicWordFramework. More specifically, GenomicWordFramework is a utility package for identifying potential genomic word framework (GWF) regions of a hypothetical language. GWFs facilitate further analysis of DNA sequences with genomic signals by applying digital signal processing (DSP) and machine learning (ML) tools from other R packages. GenomicWordFramework includes several functions to achieve read and data transformation into a suitable format for applying different statistical approaches to data analysis, such as clustering algorithms and statistical tests.

[0032] The GenomicWordFramework can take advantage of pre-identification of DMPs with R packages specific for methylation analysis, such as those described in U.S. Patent No. 10,913,986. Thus, the use of methylation and GWF analysis described herein can form a pipeline that implements signal detection and machine learning approaches that allow for rapid filtering of signal from noise.

[0033] The discriminatory power of GenomicWordFramework was first tested on a small dataset of msh1 mutant lines in Arabidopsis thaliana (a biological model) and later applied to a published study of DNA methylation analysis of placental tissue from typically developing autistic children. GenomicWordFramework analysis is a proven concept compatible with a nearly infinite number of differentially methylated network hubs. Differentially methylated network hubs previously identified in methylation and network analyses in autism studies are related to biological processes including nervous system, nervous system development, synapses, neuronal projections, central nervous system diseases, axon guidance, neurogenesis, ion / cation binding, ion / cation transmembrane transport, and voltage-gated channels.

[0034] The results show that the GWF-based heuristic can identify DNA sequences of methylation motifs with high-order DNA base interdependence with respect to methylated cytosines and with base distribution that is statistically nonrandom. These findings set the basis for further model predictions in patients not only for autism but also for other diseases such as cancer that would benefit from early diagnosis.

[0035] In other words, GWF analysis can identify sets / clusters of synonymous methylation word frameworks within genes involved in gene networks that undergo targeted methylation changes and are involved in biological processes related to the system under study. GWF further lays the foundation for the creation of libraries of DNA methylation motifs for patient diagnostic and prognostic purposes. Thus, GWF analysis substantially increases the utility and value of identifying DMPs and differentially methylated genes (DMGs) by methylation analysis.

[0036] GenomicWordFramework R package and GWF specific GenomicWordFramework is an extension written in a static programming language. More specifically, GenomicWordFramework is an R package designed for the analysis of methylation signals on stretches of DNA sequences characterized not only by methylation information but also by physicochemical information around each methylated cytosine (plus adenine in the case of animals and bacteria). The functionality of the package converts the methylation data into a suitable format, making it accessible to further DSP analysis beyond the R package.

[0037] GWFs are identified in the methylome in two possible ways: (1) application of the algorithm described in Sanchez et al., "Information Thermodynamics of Cytosine DNA Methylation," published March 10, 2016, which is incorporated herein by reference in its entirety, and (2) direct extraction of DNA sequence stretches covering a certain number of DNA bases upstream and downstream of the DMP. The GWFs obtained in the first approach can be further analyzed with digital signal processing (DSP) tools. The GWFs obtained in the second approach can be clustered into groups of aligned motifs and further tested to evaluate the deviation of each of the multiple sequence alignments (MSA) from the random Monte Carlo simulated MSA. Statistically significant motif regions are expanded for further application of DSP analysis.

[0038] Binary coding of methylation signals All the information required to search for DNA motifs, DNA sequences, and modified nucleotide bases can be stored in a singular binary string using the triplet representation of each base, taking into account the number of hydrogen bonds in Watson-Crick base pairs, the chemical type of the DNA base (pyrimidine and purine), and the base modification state. By convention, the DNA sequence is referred to as the plus strand. For example, the following coding of DNA bases can be used to study methylation signals: Methylation status: 1 (methylated) and 0 (unmethylated) Strong-weak base pair interactions: 1 (3 hydrogen bonds, S class) and 0 (2 hydrogen bonds, W class) • Chemical types: 1 (pyrimidine, Y class) and 0 (purine, R class). Radical content: 1 (bases containing amino (NH_3, M class)) and 0 (bases containing keto (C=O) groups (K class))

[0039] Each base contains a binary string. • Signal mC, methylated pyrimidine with three hydrogen bonds on the plus strand: 111. • Signal mC, methylated pyrimidine with three hydrogen bonds on the minus strand (‘+’ base G on the plus strand):110. ●C, pyrimidine with three hydrogen bonds, M class: 011 ●G, purine with three hydrogen bonds, K class: 010 ●T, pyrimidine with two hydrogen bonds, K class: 001 • Signal mA, methylated purine with two hydrogen bonds on the plus strand: 100. ● Signal mA, methylated purine with two hydrogen bonds on the minus strand ("+" base T on the plus strand): 101 ●A, purine with two hydrogen bonds, M class: 000 ●U, same as "T": 001.

[0040] Each base contains a complex number (a+ib), and the cyclic group is the eighth root of unity:

number

number

number

number

number

number

[0041] A number of complex codings are possible. For DSP analysis, coding based on group structure (here Abelian group) may be preferred. This is a group structure defined on a set of DNA bases including methylated cytosine and adenine members of the alphabet, as described further below. Analysis of complex signals in the GWF R package can be difficult, and the complex coding(s) and signal(s) can be later exported to other languages ​​such as C++, Python®, or MatLab®.

[0042] Specific details for the implementation of this coding are provided in the documentation of a function that codes previously detected binary signals of 0 and 1 from a DNA sequence into a user-defined numerical code. Given two objects, one with a signal and the other with a DNA sequence, such a function is able to perform the coding set by the user. The binary code of the methylation signals allows the incorporation of physicochemical information in the DSP analysis of DNA sequence motifs.

[0043] Motif clustering Small regions, usually 7-30 bp, spanning DMPs on at least three samples were identified and considered as DNA methylation motif candidates in 67 genes. A distance matrix can be estimated on the set of selected regions using a function from the R package that calculates the matrix distance between aligned sequences from each multiple sequence alignment (MSA). A hierarchical cluster analysis on the set of selected regions (using the previously estimated distance matrix) can then be achieved with a function that utilizes a matrix of selected information divergences to classify the selected regions into 100 clusters. It is understood that the Hellinger divergence is only one of the possible information divergences that can be estimated and applied here. For example, the J divergence is more appropriate for applications intended to extract new knowledge in terms of information thermodynamics of epigenomic phenomena. The J divergence is a symmetric version of the relative entropy.

[0044] The unweighted pair group method with arithmetic mean ("UPGMA") approach was applied as the agglomeration algorithm. In one embodiment, clusters with less than 10 regions on top are discarded. DNA multiple sequence alignment on each cluster of sequences can be achieved using the MUltiple Sequence Comparison by Log-Expectation (MUSCLE) algorithm implemented on an R package for the analysis and understanding of genomic data generated by wet-lab experiments in molecular biology.

[0045] Further partitioning of the set of motifs can be applied for downstream analysis. There is a wide range of clustering algorithms that can be applied. For example, fast k-medoids clustering can be applied and implemented using distance-based k-medoids clustering algorithms such as simple and fast k-medoids, ranked k-medoids, and increasing the number of clusters in k-medoids. In one embodiment, the algorithm is the algorithm contained in the kmed R package.

[0046] Cluster results can be plotted in marked bar plots or pca biplots. The final division into clusters depends on the applied clustering algorithm and their corresponding parameter settings, including the type of metric applied to calculate the distance matrix required for the clustering algorithm. Methylation motifs are objective DNA sequence features, and the applied clustering algorithm is merely a supporting tool leading to motif identification.

[0047] Motif Score The motif score s of aligned sequences j and k jk can be defined in an intuitive way as the base 2 logarithm of the number of matched bases found in the alignment. Formally,

number

number

number

[0048] For an MSA with M sequences each of length N, the number of pairwise comparisons is:

number

[0049] The same results are obtained if, instead of matching, the letter frequencies of MSA and Shannon entropy are estimated before (full alignment) and after. The alignment information is then calculated as follows: JPEG2025510748000013.jpg6111

[0050] Monte Carlo Test of MSA Randomness A Monte Carlo test (MCT) of how a given DNA multiple sequence alignment differs from a randomly generated MSA was implemented in an R function included in the GenomicWordFramework R package. To achieve MCT, the following is assumed: A count vector n summarizing the DNA base frequencies observed in N columns (k=A,C,G,T) k is the parameter p k The vectors are distributed according to a multinomial distribution with In a Bayesian framework, the Dirichlet distribution is a conjugate prior to the multinomial distribution, so one of two approaches to generating the parameter vector is assumed: When the number of genes or genomic regions for which motifs are identified is small, the parameter vector P = (pA, pC, pG, pT) is drawn from a Dirichlet distribution D_αP with parameter vector α = (α1, α2, α3, α4). When the number of genes or genomic regions for which motifs are identified is large, a parameter vector P = (pA, pC, pG, pT) is drawn from a Dirichlet mixture distribution, with each Dirichlet component D_(α_i Q_i) having parameters α_i = (α_i1, α_i2, α_i3, α_i4) and Q = (qA, qC, qG, qT). Formally,

number

[0051] Dirichlet distribution estimation The parameter vector α = (α1,…,α k ) is estimated from the entire set of identified DNA motif candidates. Given a matrix of DNA methylation motifs with N columns corresponding to the entire set of identified motif candidates, the frequency of each DNA base in each column is tallied to obtain a four-dimensional vector of counts for each column of the dataset, where each vector coordinate has the absolute frequency of one of the four DNA bases in a given alignment column. The resulting N × 4 matrix of counts is the raw count data used for parameter estimation of the Dirichlet distribution, which applies a function from an R package that estimates a family of continuous multivariate probability distributions, i.e., a multivariate generalization of the beta distribution. Random DNA MSA sequences are then generated from the R package according to the estimated Dirichlet distribution with a probability density function (PDF) or cumulative density function (CDF). That is, random DNA MSAs were sampled from the estimated Dirichlet distribution.

[0052] Monte Carlo p-value For MSA of fixed length N, log2N is a constant, so for the purposes of MCT, it is sufficient to consider the score statistic given in Equation 2 and evaluate how statistically different the observed aligned motifs are from the Monte Carlo simulated aligned sequences. The Monte Carlo p-value is estimated as follows:

number

number

[0053] It is important to note that the raw observed frequencies from a small matrix of motifs are often a poor approximation to the distribution of DNA bases among all the motifs that the model is predicted to represent. However, for a typical analysis of 50 or more genes, the matrix of motifs required for Dirichlet distribution model estimation will generally have several thousand motifs.

[0054] Digital signal processing of binary-coded methylation signals The binary coded methylation signals are the raw data for digital signal processing (DSP) tools, for which there are a vast number of possible applications. The GenomicWordFramework R package can include several applications of wavelet spectrograms via wavelet transforms, as well as traditional Fourier power spectra and spectrograms.

[0055] Multiplicative groups of DNA extension alphabets with methylated bases

number

number

number

number

[0056] The group defined above

number

number

number

number

number

number

[0057] Complex coding of methylation signals using the GWF R package Library (GenomicWordFramework) The present inventors have Use methylation signals from three gene regions from the data (at_signal, at_gene_seq, package="GenomicWordFramework")

[0058] Each base contains a complex number (a+ib), and the cyclic group is the eighth root of unity.

number

number

number

number

number

number

[0059] In the GWF R package, the coding of methylation signals is achieved by the function signalEncoding. In the present case, we used the generator of the group as follows:

number

[0060] The signal that an encoding is ready for export can be obtained using the function getEncoding. [ka]

[0061] We can get the entire matrix of coded signals: [ka]

[0062] The matrix carrying the signals can be exported to Python or MatLab. [ka]

[0063] Full exportable numeric matrix (only 2 rows and first 20 columns shown): [ka]

[0064] Embodiment The following non-limiting numbered embodiments also form part of the present disclosure.

[0065] 1. An extension for a general-purpose or statistical programming language, the extension comprising: 1. An algorithm or algorithms for analyzing methylation signals on a stretch of a DNA sequence, the DNA sequence comprising: i. methylation information, and ii. comprising an algorithm(s) that characterizes the physicochemical information surrounding each methylated cytosine; The algorithm(s) A distance matrix over a set of selected regions of the DNA sequence can be estimated; A hierarchical cluster on a set of selected regions can be analyzed; The set of selected regions can be classified into a specified number of clusters; An extension, including one or more features that can align multiple DNA sequences from a cluster to a methylation motif.

[0066] 2. The extension of embodiment 1, wherein the extension is written in the R statistical language.

[0067] 3. An extension as described in embodiment 1 or 2, further comprising digital signal processing (DSP) tools available in another programming language.

[0068] 4. The extension function described in embodiment 3, wherein the other programming language is C++, Python, or MatLab.

[0069] 5. The extended function of any one of embodiments 1 to 4, further comprising encoding the physicochemical properties of the DNA sequence and DNA bases into a numerical or complex signal further comprising a DSP.

[0070] 6. The enhancement function of embodiment 5, wherein said coding is based on a group structure.

[0071] 7. An extension of embodiment 6, wherein the group structure is an Abelian group.

[0072] 8. An extension function according to any one of embodiments 1 to 7, wherein the set of selected regions is classified into groups of at least 100 clusters.

[0073] 9. An extension function according to any one of the preceding embodiments, wherein clusters having less than 10 regions are discarded.

[0074] 10. An extension of any one of embodiments 1 to 9, wherein multiple DNA sequences are aligned using multiple sequence comparison with the log-expectation algorithm.

[0075] 11. The extension function of embodiment 5, further comprising exporting numeric or complex signals to C++, Python, or MatLab.

[0076] 12. An extension function according to any one of embodiments 1 to 12, wherein the methylation motifs are further classified for downstream analysis.

[0077] 13. An extension of embodiment 12, wherein the methylation motifs are further classified using a clustering algorithm.

[0078] 14. An extension of embodiment 13, wherein the clustering algorithm is a distance-based k-medoids clustering algorithm.

[0079] 15. An extension of any one of embodiments 1 to 14, further comprising encoding the physicochemical signals of methylation and DNA bases into one binary, numeric, or complex signal.

[0080] 16. The extension function of embodiment 15, wherein said coding is based on a group structure.

[0081] 17. The extension of embodiment 16, wherein the group structure is an Abelian group.

[0082] 18. An extension method described in any one of embodiments 15 to 17, further comprising performing power spectrum analysis on the encoded methylation and physicochemical signals.

[0083] 19. An extension function according to embodiment 18, wherein the power spectral analysis is wavelet power spectral analysis (WPS).

[0084] 20. The extension method of embodiment 19, further comprising calculating a correlogram based on the WPS.

[0085] 21. A computerized heuristic comprising: Higher order DNA base interdependence on methylated cytosines and A statistically non-random base distribution; and a computerized heuristic comprising:

[0086] 22. The computerized heuristic of embodiment 21, wherein the higher order DNA interdependencies and the base distributions result from analysis of at least one hierarchical cluster on a region of DNA sequences and aligning a plurality of DNA sequences from the cluster to methylation motifs.

[0087] 23. A method for analyzing methylation signals on a stretch of a DNA sequence, the method comprising: Analyzing hierarchical clusters on regions of DNA sequences; hierarchically classifying the set of selected regions into a specified number of clusters; aligning potential DNA sequence motifs from the cluster; and and applying digital signal processing to the encoded methylation and physicochemical signals.

[0088] 24. The method of embodiment 23, wherein the set of selected regions is classified into groups of at least 100 clusters.

[0089] 25. The method of embodiment 23 or 24, wherein clusters having fewer than 10 regions are discarded.

[0090] 26. The method of any one of embodiments 23 to 25, wherein the encoded methylation and physicochemical signals are encoded based on group structure.

[0091] 27. The method of embodiment 26, wherein the group structure is an Abelian group.

[0092] 28. A method according to any one of embodiments 23 to 27, wherein potential DNA sequences are aligned using multiple sequence comparison with the log-expectation algorithm.

[0093] 29. The method of any one of embodiments 23 to 28, further comprising performing power spectrum analysis on the encoded methylation and physicochemical signals.

[0094] 30. The method of embodiment 29, wherein the DSP is applied through the use of the Genome Word Framework (GWF) R package.

[0095] 31. The method of embodiment 30, wherein the GWF R package includes one or more clustering algorithms.

[0096] 32. The method of embodiment 30 or 31, further comprising exporting the GWF R package to be analyzed using other DSP tools in different programming languages.

[0097] 33. The method of embodiment 32, wherein the different programming languages ​​are C++, Python, or MatLab.

[0098] 34. The method of any one of embodiments 23 to 33, further comprising using methylation analysis to identify differentially methylated positions (DMPs) that form part of the methylation signal.

[0099] 35. The method of any one of embodiments 23 to 34, further comprising evaluating the deviation of each of the multiple sequence alignments (MSAs) from a random Monte Carlo simulated MSA.

[0100] 36. The method of any one of embodiments 23 to 35, further comprising analyzing genes having epigenetic regulatory function to diagnose or cure a disease.

[0101] 37. The method of embodiment 36, wherein the disease is autism or cancer.

[0102] 38. A computerized heuristic for analyzing genetic data, comprising: (1) The information divergence (ID) statistic estimated for each gene that has at least one differentially methylated position (DMP) on the gene or on the promoter region; (2) Principal Component Analysis (PCA), in which the first k components having 1% or more of the total sample variance are considered in downstream analysis; (3) Calculation of a correlation matrix with pairwise gene correlations, represented as a vector of PCs; (4) Analysis of correlation matrices for networks; (5) the contribution of each gene to the discrimination of the phenotype assessed in terms of the proportion of cumulative variance from the total sample variance that the gene carries.

[0103] 39. The computerized heuristic of embodiment 38, wherein the ID is selected from the group consisting of Hellinger divergence / distance, J-divergence, and total variation distance.

[0104] 40. The computerized heuristic of embodiment 39, wherein the ID is J-divergence.

[0105] 41. A computerized heuristic according to any one of embodiments 38 to 40, in which PCA is applied using a function such that genes are represented as k-dimensional vectors of PCs, and further, the square of each coordinate has a vector contribution with respect to variance to treatment discrimination from the control group.

[0106] 42. A computerized heuristic according to any one of embodiments 38 to 41, wherein PCA is applied using the pcaLDA function.

[0107] 43. A computerized heuristic described in any one of embodiments 38 to 42, wherein the correlation matrix comprises a weighted correlation network (WCN), the WCN and the network are analyzed, and the network is a protein-protein interaction (PPI) network.

[0108] 44. The computerized heuristic of embodiment 43, further comprising comparing results from the WCN and PPI networks with identified consistent relationships and contributions of epigenetic genes to phenotypes.

[0109] 45. A computerized heuristic described in any one of embodiments 38 to 44, further comprising a magnitude calculated as the Euclidean norm of the gene represented as a vector of k PCs.

[0110] 46. ​​A computerized heuristic described in any one of embodiments 38 to 45, wherein the network is related to one or more biological processes.

[0111] 47. A method for reducing computational load, the method comprising: Using methylation analysis to identify differentially methylated genes (DMGs) on a stretch of DNA sequence; Integrating the DMG into a gene network; and identifying a network hub.

[0112] 48. The method of embodiment 47, wherein network hubs are identified through protein-protein interaction (PPI) network analysis and weighted correlation network (WCN) analysis.

[0113] 49. The method of embodiment 48, further comprising comparing the results from the WCN and PPI networks with the identified consistent relationships and contributions of epigenetic genes to the phenotype.

[0114] 50. The method of any one of embodiments 47 to 49, wherein the network hub is associated with one or more biological processes.

[0115] Example 1: Arabidopsis thaliana data set For specific application to raw genomic data, the concepts, algorithms, and formulas must be implemented in some computational language. Although it should be understood that a wide variety of computational languages ​​can be utilized, this example chooses the R statistical language and relies on an R package named GenomicWordFramework. The steps and results obtained in this example are obtained from the application of the heuristics to a specific (and small) experimental data set. Results for a larger data set (in humans) are shown in a later example. [Table 1]

[0116] The goal of the package is to derive objects that may be useful for further application of DSP and ML tools available in other R packages, however some basic DSP tool applications are also provided.

[0117] Signal analysis of Arabidopsis thaliana experimental data sets An example of using empirical methylation signal data is illustrated using the dataset included in the package. The experimental dataset has methylation levels from Arabidopsis Columbia-0 ecotype (Col-0) and msh1 mutant (dwarf phenotype). Methylation data derived in previous applications of methylation analysis are included as datasets with the package. DMP datasets can be loaded from the package. [ka]

[0118] Obtaining genomic signals from their DNA sequences Here, we annotated the genes given in the GTF file downloaded from ftp: / / ftp.ensemblgenomes.org / pub / plants / release-49 / gtf / arabidopsis_thaliana / with their Arabidopsis coordinates. [ka]

[0119] The coordinates of the gene are used to obtain the corresponding DNA sequence with the signal. For simplicity, the DNA sequence is available with the package. [Table 2]

[0120] The sequence of this example is provided with the package. [ka]

[0121] Then, the binary signals on gene regions are obtained with the function getSignalAtRegions. Methylation signals are provided for three genes in five different Col-0 samples and five msh1 mutant (dwarf) samples. Only DMPs are included.

[0122] In different samples, methylation can be located at different positions, so there is a complete_region=true set to get information from the whole region, and a zero_signal=true set to allow positions with no signal. [Table 3]

[0123] This dataset is included in the R package. data(at_signal)

[0124] Results of the Examples Additional results supporting our heuristics for identifying methylation motifs in the Arabidopsis thaliana methylome are provided below.

[0125] The msh1 epi lineage comprises a distinct phenotypically based non-genetic condition. Engineering the msh1 mutant resulted in four distinct phenotypes, states 1 and 2 were characterized by delayed growth, delayed flowering, and a sustained stress response, while states 3 and 4 produced enhanced growth vigor and larger seed set over wild type (WT) (Figure 1). State 4 resulted from direct or reciprocal crosses of Col-0 msh1 mutants (state 1) or memory lines (state 2) × Col-0 WT in Arabidopsis, and the generation of epi-F2 and epi-F3 families. Similar results were obtained regardless of the direction of crossing. Progeny populations showed a more variable distribution of growth-enhanced phenotypes within the F2 generation than occurred in the grafted progeny of state 3, with individual epi line populations being classified as either enhanced vegetative growth, larger seed set, or both (Figure 2).

[0126] Four F3 populations were traced from the epi lines derived from the cross. Epi8 and Epi24 were sibling lines from one WT×msh1 mating event, and Epi10 and Epi19 were sibling lines from a second WT×msh1 mating. All four F3 epi lines showed uniform phenotypes within each population, but significant variation among the four populations (Figure 2). The Epi line populations showed increased aboveground vegetative growth and belowground root development (Figures 2A, 2E). Three of the populations, Epi8, 10, and 19, had significantly greater total leaf area than the WT control (Figures 2B, 2F), and all four populations had greater dry leaf weight than the WT (Figure 2G). The four populations also showed greater dry root weight than the WT (Figure 2H).

[0127] Enhanced reproductive growth occurred in three epi lines, Epi8, 10, and 19, although Epi8 showed early flowering (Figure 2C). Importantly, full-sib populations Epi8 and Epi24 from the same original cross exhibited distinct phenotypes. Epi24 had the highest seed weight and lowest leaf area among the epi lines, whereas Epi8 had the highest leaf area and lowest seed weight. These observations indicate that growth vigor can vary in vegetative versus reproductive allocation within closely related populations. A similar phenomenon can occur in tomato epi lines (Figure 3), where sibling epi line populations also differed in vegetative growth enhancement (Figure 3H).

[0128] The epi lineage phenotype reverted to wild type by the fifth or sixth (S5, S6) generation. Over these successive generations, sporadic occurrences of reversion to a state resembling the memory (state 2) phenotype (Figures 4A, 4B) were observed and were characterized by reduced growth rate, altered leaf morphology, and enhanced stress response. This reversion was not observed in the grafted progeny. Reversion frequencies varied among populations and occurred among species (Table 1). [Table 4] *n, i, and e refer to mild (normal), intermediate, and extreme phenotypes of the msh1 memory lines used in crosses to generate the epi lineage population. (TDNA) refers to the msh1 TDNA insertion mutant used in crosses as a control.

[0129] Plant characteristics of the four msh1-derived states in Arabidopsis were replicated in tomato (Figure 3) and soybean lines used in this study. Cross-species similarities included epi line reversion (Figure 4; Table 1). What is now recognized as epi line reversion was also previously reported as instability in msh1-modified sorghum. The observation of reversion phenomena across different species implies that the epi line state 4 and memory state 2 states are closely related. Reversion only occurred in epi line populations derived from crosses to msh1 memory lines (state 2) parents, but not to msh1 null mutants (state 1) parents (Table 1). One interpretation of this observation is that the msh1 signal is weaker in memory lines than null mutants, resulting in a more unstable epi line outcome.

[0130] msh1 states 1-4 comprise distinct epigenetic phases by whole genome methylome analysis. Significant changes in DNA methylation were detected in the four Arabidopsis epi lines (F3), with gene-associated changes detected primarily in the CG context (Figure 5). Similar DNA methylation changes occurred in Arabidopsis, tomato, and soybean epi lines, which showed variable hyper- and hypomethylation in the CHH context within and between the three epi lines (Figure 6). Comparison of soybean Epi-F4 and Epi-F6, in which the Epi-F4 line exhibited enhanced growth vigor that was reduced to a wild-type phenotype by the F6 generation, allowed us to identify potential growth vigor-associated DNA methylation changes (Figure 6C).

[0131] High-resolution methylome analysis was used to estimate the relationship of genetic differential methylation with plant phenotypic variation. This procedure incorporates signal detection and machine learning for the identification of high-probability, treatment-associated methylation changes within genic regions. Hierarchical clustering of methylome data from individual plants from all four states of Arabidopsis thaliana was used to estimate the changes in methylation levels (calculated as Hellinger divergence) at differentially methylated positions (DMPs) in genic regions. Results showed clustering of individual plants (biological replicates) from the same population, with plants from different states segregating into six branches distinct from the WT control group (Figure 7). Despite originating from the same cross, Epi8 and Epi24 segregated into two subgroups (Figure 7A), appearing to be consistent with their distinct phenotypes (Figure 2). The data suggested that Epi24 gene DNA methylation repatterning was more closely related to that of the grafted progeny (HEG; state 3) than its full-sibling Epi8 (Figure 7A). Epi10 and Epi19 also formed two closely related clusters consistent with their phenotypic similarity. In contrast, hierarchical clustering of transposable element-associated DMPs in the four epi lineages yielded phylogenetic outcomes such that Epi8 and Epi24 were now clustered together, as were Epi10 and Epi19 in a single subcluster, and both clusters were separated from the HEG state 3 (Figure 7C).

[0132] Similar results emerged in hierarchical cluster analysis of tomato gene methylome data. The tomato msh1 mutant (state 1), HEG (state 3) and epi lineage data sets (state 4) formed three branches consistent with unique gene methylation effects (Figure 4b). Two epi lines with distinct plant growth vigor phenotypes, Epi high (H; enhanced biomass over WT) and Epi low (L; biomass equivalent to WT), were separated into two subgroups within the same main cluster (Figure 7B). The observations indicate that gene-associated methylation repatterning can be identified to reflect plant phenotypes among lines predicted to be genetically homogeneous.

[0133] Methylation data support epi-lineage revertants more closely related to msh1 memory state 2. Epi-lineage (state 4) revertant samples in Arabidopsis were included in the methylome analysis to test for evidence of methylation repatterning in revertant versus non-revertant full-sib samples. Principal components with linear discriminant analysis (PCA-LDA) of these data sets using Hellinger divergence produced a clear clustering of non-revertants from revertant individuals derived from the same progeny population. Epigenomic features discriminating between revertant and non-revertant full-sib progeny likely arose spontaneously. The revertant methylome data set showed a closer relationship to data from msh1 memory state 2 (Figure 4C). These msh1 memory plants were derived from an independent population. The relationship of revertant methylome profiles with msh1 memory would be consistent with their demonstrated similarity in plant phenotype (Figure 4). These results support the methylome features 18 Our results support previous adaptations of msh1 memory that are unique in M. elegans and identify low-frequency conversion of epigenetic state 4 with enhanced growth vigor as similar to state 2, a persistent stress-response phenotype.

[0134] Epi lineage methylome datasets are altered in growth-related gene networks. Differentially methylated genes (DMGs) for each epi lineage population were identified by applying generalized linear regression analysis (GLM) to test the significance of differences between group DMP counts (WT vs. epi lineages) in genes. This analysis identified 3204, 2860, 3208, and 2797 DMGs in Epi8, Epi10, Epi19, and Epi24, respectively. To investigate the concordance of DMG functional relationships in each population, gene network-based enrichment analysis was performed. DMGs from different epi lineages shared enrichment for specific functional networks. For example, full-sibs Epi8 and Epi24 shared common enrichment networks for response to auxin, response to red or far-red light, response to nutrient levels, photoperiodism, detection of abiotic stimuli, response to stress, and catabolic processes (Figure 8A), whereas Epi10 and Epi19 shared enrichment networks for response to cyclopentenone, response to auxin, cellular response to auxin stimulation, and auxin-activated signaling pathways (Figure 8B). This outcome reflects the nonrandom quality of the DMP dataset and implies that methylation machinery targeted loci and their respective networks can be identified.

[0135] Using the methylome analysis method applied to the msh1 dataset, we analyzed previously reported methylome data for Arabidopsis F1 heterotic crosses of ecotypes C24 and Ler. The enrichment networks that emerged from the F1 hybrid (C24 × Ler, and Ler × C24) data showed remarkable compatibility with those identified in the msh1-derived epi line data, highlighting pathways for responses to red or far-red light, responses to auxin, regulation of growth, cellular responses to auxin stimulation, and auxin-activated signaling pathways (Figure 8C). DMG overlap between the hybrid data and the epi lines ranged from 47% overlap with Epi10 to 54% overlap with Epi8. This apparent alignment of methylome-specific gene pathways between heterotic hybrids and msh1 epi lines supports a link between phenotype and methylome repatterning, linking enhanced vigor to altered auxin signaling. The observations were also consistent with the profiles of DMG-derived enriched networks in tomato and soybean epi lineage datasets, identifying prominent auxin-related networks (Figures 8D-8E). The soybean study highlighted phenotypic changes, with Epi-F4 lines measurably enhanced growth vigor that was reduced back to wild-type growth by the Epi-F6 generation. Therefore, we analyzed discriminating methylome features between Epi-F4 and Epi-F6 datasets as a direct comparison to identify vigor-related methylome features within a single lineage.

[0136] Examination of gene expression changes in the Arabidopsis epi lineage population involved sampling three different tissues for RNAseq analysis. Epi8 and Epi10 were contrasted with wild type for differential gene expression in leaf tissues by sampling similar to that performed for the methylome study, and Epi8 and Epi24 were further compared by analysis of inflorescence stem and root tissues. From leaf tissue, 1884 differentially expressed genes (DEGs) were identified in Epi8 compared to wild type and 992 in Epi10. Epi8 and Epi24 analysis revealed 1991 DEGs from inflorescence stems in Epi8 and 1650 for Epi24, and 1133 DEGs from roots in Epi8 and 1111 in Epi24 compared to wild type.

[0137] Network enrichment analysis of the derived DEG datasets showed shared pathways that were altered in response to msh1 effects in leaf, flower stem and root tissues. The most enriched of these included abiotic and biotic stress responses, with circadian rhythm and plant hormone response related networks also significantly enriched in the three tissues. Regulation of transcription was particularly prominent in the flower stem, where MSH1 accumulates (Figure 9). These outcomes appeared consistent with the epi lineage DMG results, as integration of the DEG and DMG datasets helped to amplify msh1-related network enrichment compared to either dataset alone (Figure 8).

[0138] Differential methylation and expression analysis identified central gene hubs for the epi lineage state. To explore the interactions of the DMG and DEG datasets in the epi lineage, DMG and DEG data were input into Cytoscape to construct protein-protein interaction (PPI) maps, followed by K-means cluster analysis to identify putative core networks with central gene hubs. The K-means cluster machine learning algorithm uses betweenness centrality, closeness centrality, average shortest path length, clustering coefficient, degree, and eccentricity as parameters, allowing for the identification of clusters containing the most concentrated nodes (proteins) in the PPI network.

[0139] In Arabidopsis Epi8, a total of 3647 unique loci from DMGs and DEGs were used in the analysis, resulting in a PPI network formed by 430 genes. Functional enrichment analysis of these putative hub genes using the STRING database functional enrichment tool revealed a PPI network of 153 hub genes and associated functional networks (Figure 10A). These core networks included developmental processes, response to hormones (especially auxin), response to low temperature, chromosome organization, mRNA processing, spliceosome, and ribosome biogenesis. Comparison with Epi24, where a total of 3523 unique loci from DMGs and DEGs formed a much larger PPI network of 346 genes, indicated that some of these same processes were represented (Figure 10B). However, Epi24 analysis also revealed more pronounced changes in development-related gene expression, and most notably robust and overlapping effects in ribosome biogenesis-related gene expression (Figure 10B). Ribosome biogenesis is essential for growth responses. Because Epi8 and Epi24 are full siblings from the same mating, these different traits are predicted to be related to their different phenotypes, with Epi8 being enhanced in vegetative growth and Epi24 showing significantly increased reproductive vigor (Figure 2).

[0140] Analysis of soybean epiphylogenetic data involved assigning an Arabidopsis ortholog to each soybean gene identified with BLASTP, obtaining a soybean epiphylogenetic PPI network of 109 core hub genes and their functional networks (Figure 11A). The resulting enriched networks included developmental processes, ribosomes, chromatin organization, mRNA processing, and response to light stimuli. These results showed cross-species overlap in the identified pathways (Figure 11A), thus suggesting that msh1 directional effects led to parallel phenotypic outcomes in different species through similar networks. For core hub genes, the overlap of annotated genes between soybean and Arabidopsis was 16% (18 genes).

[0141] To assess the relationship of the PPI network output for epilines to the graft-derived HEGs, a core-hub PPI network was constructed for the tomato HEG (state 3) phenotype. The network included ribosome biogenesis, developmental processes, and chromatin organization (Figure 11B). Identified core hub proteins shared in the epilines and HEG datasets from soybean, Arabidopsis, and tomato included TOR, MOM, RFC1, UBQ family, NRPD family, and HAC family proteins. Although there was cross-species conservation of some overlapping hubs, detailed cross-species comparisons were limited by differences in gene annotations.

[0142] Epi state comparison in Arabidopsis reveals conserved msh1 epigenomic targets within biologically meaningful gene networks. To investigate the relationship of genetic methylation repatterning between the four distinct msh1-derived states, DMG overlap was assessed. Figure 12A shows that 871 DMGs were shared between the four msh1 states, including 17.6% of DMGs in msh1 (state 1), 39.7% in memory (state 2), 31.6% in grafted progeny (HEG, state 3), and 31.1% in epi lineages (state 4, Epi24). The overlap established a conserved msh1 "core" DMG dataset (Figure 12A). Each state was also classified by state-specific DMGs, which accounted for 33.8% in state 1, 15.4% in state 2, 10.2% in state 3, and 15.2% in state 4 (Figure 12A). Of the 871 “core” msh1 -associated DMGs, 531 (61%) could be placed within known networks.

[0143] Components of the RdDM pathway were shown to be required for the induction of msh1 state 1, the transition from state 1 to state 2, and the generation of state 3 after grafting. msh1 Grafted progeny from and Col-0 / Col-0 dcl1 / dcl3 / dcl4 / msh1 The grafted (state 3) contrasted methylome dataset helped classify DMGs as RdDM(sRNA) dependent. These subtracted datasets confirmed that 674 of the 871 core DMG loci (77%) were predicted to be DCL2, DCL3, DCL4 dependent by obtaining overlap between the 871 DMG core dataset and the msh1 and dcl2 / dcl3 / dcl4 / msh1 datasets. Figure 12B shows the PPI core hub network of DMGs with those predicted as DCL2, DCL3, DCL4 dependent with cross-hatching. These outcomes provide evidence of a relationship between msh1-directed epigenetic effects and the impact of the RdDM pathway on targeted methylome repatterning.

[0144] Although 871 core DMGs were shared among the four msh1 states, methylation changes were also found within these 871 loci to help distinguish the four states. Figure 12C shows the results of hierarchical clustering for individual plants from the four epigenetic states with separation into four distinct clades by using only DMP methylation information across the 871 msh1 core genes. Based on the PPI network for the 871 DMGs, K-means clustering was performed to identify putative central hubs and their functional enrichment analysis. Figure 12D shows the resulting network of 67 candidate hubs involved in response to stress, developmental (growth) processes, gene expression, spliceosome, histone modification, and chromosome organization networks. More than 80% of the 67 hubs were predicted to be RdDM targets, including several central players in plant development and environmental stress responses, such as the cell growth regulator TARGET OF RAPAMYCIN (TOR), SWI2 / SNF2-like, the meristem identity gene SPLAYED (SYD), its close homologue BRAHMA (BRM), and the chromatin remodelers PICKLE (PKL) and PICKLE-RELATED 1 (PKR1 / CHR4).

[0145] Transposable elements and sRNAs associate with candidate RdDM target loci in 871 msh1 DMGs. The RdDM pathway is known to actively target transposable element (TE) sequences, prompting the investigation of TE association with msh1-responsive loci. Looking at the 871 DMGs common between msh1 conditions, association of these loci was detected in TE and sRNA (20-24 nt) clusters at a higher than genome-wide level (61% DMGs within 2 kb of TE vs. 47% genome-wide; 77% DMGs within 2 kb of sRNA cluster vs. 46% genome-wide). This enrichment of TE and sRNA cluster proximity increases further when dcl2-, dcl3-, and dcl4-sensitive DMGs (65% DMGs within 2 kb of TE; 81% DMGs within 2 kb of sRNA cluster) are subsetted. However, when focusing on the 67 hub DMGs derived by k-means clustering, only sRNA cluster enrichment was found (49% of DMGs within 2 kb of TEs; 72% of DMGs within 2 kb of sRNA clusters). Figure 12D shows only TE-associated genes (7), sRNA-associated genes (22), and genes associated with both TEs and sRNAs (26). Thus, although TEs can affect the methylation status of proximal genes and act as RdDM targets, sRNA cluster association could define RdDM targets for msh1 effects, regardless of TE proximity.

[0146] We found some possible evidence of association between TE families and DMG proximity. Comparing the number of members between observed and expected TE families revealed significant over-representation in L1 and Gypsy families and under-representation in Helitron families in both 871 DMGs common between msh1 states and 674 DMGs sensitive to dcl2, dcl3, dcl4. Further investigation is needed to clarify any biological significance of these associations. Based on various analyses, four criteria helped to classify the RdDM target loci in our study (Figure 12D). Positive loci were DMGs confirmed in all four msh1 states (1), including putative network hubs based on k-means clustering (2), with genetic evidence for dcl2, dcl3, dcl4 sensitivity (3), and / or association with sRNA clusters (4).

[0147] Detailed methylation analysis of selected RdDM target genes in four distinct non-genetic conditions reveals sequence motifs encompassing dcl2 / 3 / 4-sensitive DMPs. Annotation of the 67 candidate hub loci supported their association with the phenotypic effects observed in the four msh1 conditions. In addition to gene networks for changes in gene expression and chromatin behavior, major overlapping networks appeared to reflect the observed transition between stress response and growth underpinning the msh1 and memory stress phenotypes versus growth vigor in the HEG and epi lineage conditions (Figure S12D). The methylation signatures of the 67 candidate hub loci discriminating between the four msh1 conditions support a model of precisely targeted repatterning (Figure S12C).

[0148] Figure 13 and Tables 2 and 3 show the analysis results for seven selected candidate RdDM target loci: TARGET OF RAPAMYCIN (TOR), a cell growth regulator, SPLAYED (SYD), a chromatin remodeling component, UBIQUITIN-SPECIFIC PROTEASE 26 (UBP26), required for heterochromatin silencing, NUCLEAR RNA POLYMERASE D1A (NRPD1), the largest subunit of RNA polymerase IV, RNA POLYMERASE II LARGE SUBUNIT (NRPB1), involved in transcription, SU(VAR)3-9-RELATED PROTEIN 5 (SUVR5), a gene involved in H3K9me2 deposition and gene silencing, and UP-FRAMESHIFT1 (UPF1), a gene involved in RNA interference and splicing. [Table 5] *+ indicates presence of DMP in at least two samples, - indicates absence of DMP, ! indicates dcl2, dcl3, dcl4 sensitivity. **Numbers in brackets refer to the number of epi lineages displaying DMP in at least two samples. [Table 6] *+ indicates presence of DMP in at least two samples, - indicates absence of DMP, ! indicates dcl2, dcl3, dcl4 sensitivity. **Numbers in brackets refer to the number of epi lineages displaying DMP in at least two samples.

[0149] Differential methylation analysis within the seven loci revealed evidence of condition-specific repatterning (Tables 2 and 3). Methylation changes at each locus were associated with identifiable sequence motifs spanning approximately 14 nucleotides. Two sample composite motifs are shown in FIG. 13A, and examples of UBP26 and UPF1 condition-specific methylation repatterning within cluster 11 are shown in FIG. 13B and 13C. Systematic analysis of 67 loci revealed similar condition-specific changes. Although DMPs were assayed individually, multiple DMPs could be detected within a single identified motif domain. DMPs were identified across replicate datasets, with some sites shown to be dcl2 / dcl3 / dcl4 sensitive in condition 1 and condition 3 (FIG. 13C; Tables 2 and 3). It was only possible to test dcl2 / dcl3 / dcl4 sensitive in two conditions, encompassing the msh1 rootstock. Tables 2 and 3 summarize the scoring outcomes for cluster 11 and 18 motifs within the seven selected genes.

[0150] The identified DMPs did not show obvious patterns of exon, intron or junction localization, and each gene contained multiple motif sites. Evaluation of the cluster motifs encompassing the DMPs within the 67 msh1 core hub loci revealed evidence of higher order dependencies in many cases. Multiple sequence alignment (MSA) of a given DNA motif can reveal dependencies between two nucleotides located at different positions within the motif, reflected in their frequency of co-occurrence. First order dependencies refer to adjacent nucleotides, typically found in CG methylation contexts, second order for nucleotides spaced two nucleotides apart, and higher order for nucleotides with intermediate distances of more than two nucleotides. The relationships are derived from studies of Markov dependencies in DNA sequences, which is the basis for the application of hidden Markov modeling of motif findings.

[0151] For the identified motifs, individual consensus nucleotides were evident at variable distances from the target cytosine, which was nucleotide 7 (on the plus or minus strand) within each motif. For example, the motif from cluster 65 showed an invariant T at position 14 and a consensus A at position 12, while the motif from cluster 66 showed an invariant G at position 14 and a consensus AG pair at positions 2 and 3, respectively (Figure 13D). In motif 76, only G was present at positions 6 and 9 with a consensus T at position 5, while motifs 82 and 86 showed a consensus or invariant T at position 2. These observations support the hypothesis that the target methylation sites within these genes are characterized by invariant or nearly invariant sequence features.

[0152] Small regions (14 bp) encompassing DMPs in at least three samples were identified and considered as DNA methylation motif candidates in the 67 identified msh1 core hub genes. A distance matrix on the set of selected regions was estimated using the function dist.dna from the ape R package (version 5.5). Hierarchical cluster analysis on the set of selected regions (using the previous estimated distance matrix) was achieved using the function hclust from the stats R package (version 4.1.1) and classified into 100 clusters. The UPGMA approach was applied as an agglomeration algorithm. Clusters with less than 10 discarded regions. DNA multiple sequence alignment on each cluster of sequences was achieved using the MUSCLE algorithm implemented on the Bioconductor R package muscle (version 3.14). The motifs presented in Figure 13A were obtained from the specific analysis of the genes TOR, SYD, NRPB1, NRPD1, UBP26, SUVR5, and UPF1.

[0153] Example 2: Detection of signal peaks A rapid method for discovering DNA sequence features associated with methylation is to locate methylation sites across samples. This approach relies on evidence that methylation occurs at specific, non-random sites (a type of DNA sequence motif). [Table 7]

[0154] This data is included in the package. [ka]

[0155] The function signal_peaks provides potential DNA sequence motifs with coordinates centered on the methylation peak covering 3 up and 4 down bp around the signal. Methylation signals were required to be present in at least three samples (cut point = 3L).

[0156] Clustering of potential DNA sequence motifs DNA multiple sequence alignment MUSCLE and hierarchical clustering are applied to identify clusters of DNA sequence motifs (Figures 14A-14D). [Table 8]

[0157] A cluster motif encompassing DMP on gene AT1G50030 (TOR) revealed evidence of higher order dependence: in a multiple sequence alignment (MSA) of a given DNA motif, base Y at a given site k depends on base X at the preceding site j if frequent bases X and Y are observed together in the MSA. In particular, when kj=1, k is a first order dependence typically found in CG methylation contexts, when kj=2, k is a second order dependence, and when kj>2, k is a higher order dependence.

[0158] For these motifs, individual consensus nucleotides were evident at variable distances from the target cytosine, which was nucleotide 7 (on the "+" or "-" strand) within each motif. For example, motif from cluster 1 (Figure 14A) showed only a T at position 1 (maximum score of 2 bits), a consensus G at position 9, and Ts (about 1.5 bits) at positions 12 and 15, while motif 17 (Figure 14D) showed only a T at its position 1. In motif 12 (Figure 14C), a significant consensus for A is found at positions 5 and 10 with a methylated C at position 7.

[0159] To obtain DNA sequence motifs by genes. [ka]

[0160] Application of Digital Signal Processing (DSP) The DSP can be applied to the previously acquired signals. Next, the power spectrum from the signals from chromosomes 2 and 3 is calculated.

[0161] A function can be applied that codes the previously detected binary signals of 0 and 1 from the DNA sequence into a user-defined numerical code. Possible codings can be binary, real, and complex. Coding of DNA methylation sequences using complex numbers is also supported in GWF. Coding using normal real numbers is supported as well. The basic idea is to code the physicochemical properties of DNA bases. In this scenario, by applying different DSP tools, periodicities and correlations on the coded signals targeting the superposition of methylation and physicochemical signals can be searched. Currently, DSP analysis of complex signals with R is not good. However, methylation signals can be coded in GWF and then exported to, for example, MatLab or Pythom, where DSP analysis can be achieved.

[0162] Given two objects, one with a signal and the other with a DNA sequence, the function performs the coding set by the user. Using this function, previously detected binary signals of 0 and 1 from a DNA sequence can be recoded into a numerical code defined by the user. In particular, it is possible to incorporate information about the physicochemical properties of adjacent DNA bases, i.e. the number of hydrogen bonds and the chemical type of the base. This can be the default used by this function. 1.Methylation status:1 (methylated) and 0 (unmathylated) 2. Hydrogen bonds:1 (three hydrogen bonds) and 0 (two hydrogen bonds) 3.Chemical type:1 (pyrimidine) and 0 (pyrimidine)

[0163] So each base becomes a binary string: 1. the signal mC, methylated pyrimidine with three hydrogen bonds:111 2. C, pyrimidine with three hydrogen bonds:011 3. G, purine with three hydrogen bonds:010 4. T, pyrimidine with two hydrogen bonds:001 5. A, purine with two hydrogen bonds:000 [ka]

[0164] The signal is then divided into non-overlapping window intervals of 90 bits (30 bp). [ka]

[0165] There are 592 potential DNA sequence motifs. [ka]

[0166] Overlap can be searched between the motif and a 30 bp wide sliding window. [ka]

[0167] That is, 20 genomic word frameworks from 592 lie entirely within the 30 bp region on gene AT1G50030. However, in the dwarf sample "dw2", there are only three significant motifs covered by this region. [ka]

[0168] The three motifs are embedded at the beginning of the signal region being examined. [ka]

[0169] Sequences from clusters 12 and 9 (FIGS. 14B, 14C) differ only by a base shift (3 bits). [ka]

[0170] Power spectrum analysis on experimental data sets The power spectrum of the binary signal can be obtained from the SignalMatrix-class object using the function plot_power_spectral (Figure 15). The dots help visually identify peaks where spectral differences between control (wt) (Figure 15A) and treatment (msh1 mutant, dwarf phenotype) (Figure 15B) are found. [Table 9]

[0171] The 1 / 3 (0.33) peak indicates that the region under investigation is a protein coding region (Figure 15A). Methylation may or may not affect the period 1 / 3. In protein coding regions, peaks other than the 1 / 3 peak arise from periodicities derived from the physicochemical properties of the DNA bases, which in the present case are encoded in a binary vector representing the base sequence.

[0172] spectrogram A spectrogram is a visual representation of the spectrum of the frequency of a signal as it changes over time. In our case, the axis "time" is represented by bits or single DNA base positions. Spectrograms can be obtained from any coded signal region using the function "spectrogram" from the R package phonTools.

[0173] To prepare methylation signals as numerical signals. [ka]

[0174] Spectrograms for the same genes from the control (Figure 16A, upper panel) and treated (Figure 16A, lower panel) are then easily calculated. [Table 10]

[0175] The methylation signal resolves the power spectral energy around a periodicity at approximately 1 / 3 frequency (dashed line in FIG. 16A) and bin 16. In the third region, the following was found: [Table 11]

[0176] In this case, the methylation effect is around bins 6-10, and the energy power is downshifted to 1 / 3 frequency (Figure 16B). In essence, the effect of methylation on signal power and periodicity depends on the DNA sequence context. This is an expected result that has been ignored in conventional methylation analysis so far.

[0177] Wavelet Power Spectrum The wavelet coefficients provide information about the correlation between the wavelet (at a particular scale) and the data array (at a particular location): more positive amplitude means more positive correlation, while more negative amplitude means more negative correlation.

[0178] The wavelet power spectrum provides a useful method for determining the distribution of energy in a data array. By examining plots for areas within the wavelet power spectrum (WPS) of large power, one can determine which features of a signal are important and which features can be ignored. Here, the term "energy" is not arbitrary, but is borrowed from applications in human-built communication systems. The level of energy represented in the WPS is proportional to the energy dissipated in transmitting a binary signal of the same size to a given receiver through a human-built communication grid.

[0179] Furthermore, with the advancement of information theory and its application to biomolecular processes, it is well known that to achieve a single methylation change, all methyltransferases / demethylases must dissipate a minimum amount of energy to process the information associated with the change. This energy is determined by Landauer's principle, according to which a molecular machine must dissipate a minimum amount of energy at each step of a genetic logic operation, including proofreading, ε=k B A minimum energy of T ln2 (approximately 3×10-21 Joules / bit at room temperature) must be dissipated.

[0180] Wavelet power spectrum analysis on experimental data sets Wavelet power spectrum analysis of the previously estimated at_signal_diff dataset (Figure 17) is achieved by: [Table 12]

[0181] Some methylation motifs have the same methylation state in the same region from both groups, Col-0 and msh1 Dwarf. This is the case for the regions: AT1G50030.3 (Figure 17C) and AT1G50030.6. The changes observed in the wavelet power (WPS) spectrum are linked to the base where the signal is located.

[0182] In the case of AT1G50030.1 (Figure 17A), the WPS indicates that the change is made at 60 bits, meaning the base at position 20 = 60 / 3. The table below confirms our predictions. The methylated cytosine is located at base 20 in the minus strand from Col-0. The reported base in the plus strand is G, which is coded for by the binary string 110 when the complementary base C is methylated. The sample base position is unmethylated in msh1 Dwarf-2 (G:010). [ka]

[0183] Similarly, a correlogram can be calculated based on the WPS (Figure 18). [Table 13]

[0184] In addition to the expected correlation decomposition in the region 55-60 bits, there is a loss of correlation between the energy spectra in the regions 30-35 bits and 55-60 bits (FIG. 18).

[0185] Example 3: Human Data Set Analysis of methylation signals achieved on 81 genes related to human nervous system development.

[0186] Figure 19 shows the subnetwork of 81 DMG hubs. A PPI network was constructed on the set of 751 DMGs identified by principal component analysis and analyzed with the STRING Cytoscape App. Circles without cross-hatching highlight DMGs involved in biological processes including nervous system, nervous system development, synapse, neuronal projection, ion transport, CNS disease, axon guidance, neurogenesis, ion / cation binding, ion / cation transmembrane transport, voltage-gated channels, and TRP channels.

[0187] The entire set of DMGs was derived from 751 DMGs, which were selected according to their contribution to the classification of patients into two groups: "typical" and "autistic". Specifically, 751 DMGs contribute more than 1% of the total variance to the major principal components from the PCA. These DMGs were analyzed with the STRING Cytoscape App, and major sub-networks of hubs were identified applying a K-means clustering approach using network centrality indices as variables. All DMGs were obtained by methylation analysis.

[0188] Figure 20 is a network enrichment analysis identified in the main network of DMG hubs from Figure 19. The methylation motifs identified in 81 DMG members of the main sub-network of hubs derived from the network of genes associated with autism can be achieved by the methods described herein.

[0189] From the foregoing, it can be seen that the present disclosure accomplishes at least all of its stated objectives.

[0190] Glossary Unless otherwise defined, all technical and scientific terms used above have the same meaning as commonly understood by one of ordinary skill in the art to which the embodiments of the present disclosure pertain.

[0191] The terms "a," "an," and "the" include both singular and plural referents.

[0192] The term "or" is equivalent to "and / or" and means any one member or combination of members of a particular list.

[0193] The terms "invention," "the present invention," "disclosure," or "this disclosure" are not intended to refer to any single embodiment of a particular invention, but rather encompass all possible embodiments described in the specification and claims.

[0194] The term "about" as used herein refers to slight variations in a quantity for any quantifiable variable. Unintentional errors may occur, for example, from the use of typical measuring techniques or equipment, or from differences in manufacture, source, or purity of the components.

[0195] The term "substantially" refers to a large or significant degree. Thus, given the appropriate context, "substantially" can refer to a plurality, a majority, and / or a vast majority of the quantifiable variables in question.

[0196] The term "generally" encompasses both "about" and "substantially."

[0197] The term "configured" describes a structure that is capable of performing a task or adopting a particular configuration. The term "configured" may be used interchangeably with other similar phrases such as built, arranged, adapted, manufactured, etc.

[0198] Terms characterizing order, position, and / or orientation are not limiting and are merely referenced according to the figures presented.

[0199] "Phenotype" refers to the set of observable characteristics of an individual that results from the interaction of its genotype and the environment.

[0200] "Epigenetic" refers to resulting from non-genetic influences on gene expression.

[0201] An "R package" is an extension to the R statistical programming language. R packages typically contain code, data, and documentation in a standardized collection format that can be installed by users of R via a centralized software repository such as CRAN.

[0202] "Expression" is the degree to which a phenotype is expressed by individuals possessing a particular genotype.

[0203] The "scope" of the present invention is defined by the appended claims, along with the full scope of equivalents to which such claims are entitled. The scope of the present disclosure is further limited as including any possible modifications to any of the aspects and / or embodiments disclosed herein that result in other embodiments, combinations, subcombinations, etc. that will be apparent to those skilled in the art.

Claims

1. An extension for a general purpose programming language or a statistical programming language, said extension comprising:

1. An algorithm or algorithms for analyzing methylation signals on a stretch of a DNA sequence, said DNA sequence comprising: iii. Methylation information, and iv. includes an algorithm(s) that features physicochemical information around each methylated cytosine; The algorithm(s) A distance matrix on a set of selected regions of the DNA sequence can be estimated; analysing hierarchical clusters on the set of selected regions; The set of selected regions may be classified into a specified number of clusters; An extension function that includes one or more functions that can align multiple DNA sequences from the cluster to a methylation motif.

2. The extension of claim 1 , wherein the extension is written in the R statistical language.

3. The enhanced functionality of claim 2 further comprising digital signal processing (DSP) tools available in another programming language.

4. The extension function of claim 3 , wherein the other programming language is C++, Python, or MatLab.

5. The enhanced function of claim 3, further comprising encoding the DNA sequence and physicochemical properties of the DNA bases into a numerical or complex signal further comprising the DSP.

6. The enhanced function of claim 5 , wherein the encoding is based on a group structure.

7. The extension method of claim 6 , wherein the group structure is an Abelian group.

8. The method of claim 1 , wherein the set of selected regions is classified into groups of at least 100 clusters.

9. The enhancement of claim 1 , wherein clusters with fewer than 10 regions are discarded.

10. The method of claim 1 , wherein the plurality of DNA sequences are aligned using multiple sequence comparison with a log-expectation algorithm.

11. The extension of claim 5 further comprising exporting the numeric or complex signal to C++, Python, or MatLab.

12. The method of claim 1, wherein the methylation motifs are further classified for downstream analysis.

13. The method of claim 12, wherein the methylation motifs are further classified using a clustering algorithm.

14. The extension of claim 13 , wherein the clustering algorithm is a distance-based k-medoids clustering algorithm.

15. The method of claim 1 further comprising encoding the physicochemical signals of methylation and DNA bases into one binary, numeric or complex signal.

16. The enhanced feature of claim 15 , wherein the encoding is based on a group structure.

17. The extension of claim 16, wherein the group structure is an Abelian group.

18. The enhanced method of claim 15, further comprising performing a power spectrum analysis on the encoded methylation and physicochemical signals.

19. The enhancement according to claim 18, wherein the power spectrum analysis is wavelet power spectrum analysis (WPS).

20. The enhanced method of claim 19 further comprising: calculating a correlogram based on the WPS.

21. Higher order DNA base interdependence on methylated cytosines; A statistically non-random base distribution; and a computerized heuristic comprising:

22. 22. The computerized heuristic of claim 21, wherein the higher order DNA interdependencies and the base distributions result from analysis of at least one hierarchical cluster on a region of DNA sequences and aligning a plurality of DNA sequences from the cluster to methylation motifs.

23. A method for analyzing methylation signals on a stretch of a DNA sequence, said method comprising the steps of: analyzing hierarchical clusters on regions of said DNA sequence; hierarchically classifying the set of selected regions into a specified number of clusters; aligning potential DNA sequence motifs from said clusters; and applying digital signal processing to the encoded methylation and physicochemical signals.

24. 24. The method of claim 23, wherein the set of selected regions is classified into groups of at least 100 clusters.

25. 24. The method of claim 23, wherein clusters with fewer than 10 regions are discarded.

26. The method of claim 23, wherein the encoded methylation and physicochemical signals are encoded based on group structure.

27. 27. The method of claim 26, wherein the group structure is an Abelian group.

28. 24. The method of claim 23, wherein the potential DNA sequences are aligned using multiple sequence comparison with a log-expectation algorithm.

29. 24. The method of claim 23, further comprising performing a power spectrum analysis on the encoded methylation and physicochemical signals.

30. 30. The method of claim 29, wherein the DSP is applied through the use of the Genome Word Framework (GWF) R package.

31. The method of claim 30 , wherein the GWF R package includes one or more clustering algorithms.

32. 31. The method of claim 30, further comprising exporting the GWFR package to be analyzed with other DSP tools in different programming languages.

33. 33. The method of claim 32, wherein the different programming languages ​​are C++, Python, or MatLab.

34. 24. The method of claim 23, further comprising using methylation analysis to identify differentially methylated positions (DMPs) that form part of the methylation signal.

35. 24. The method of claim 23, further comprising evaluating the deviation of each of the multiple sequence alignments (MSAs) from a random Monte Carlo simulated MSA.

36. The method of claim 23, further comprising analyzing genes having epigenetic regulatory functions to diagnose or cure a disease.

37. 37. The method of claim 36, wherein the disease is autism or cancer.

38. 1. A computerized heuristic for analyzing genetic data, comprising: (1) The information divergence (ID) statistic estimated for each gene that has at least one differentially methylated position (DMP) on the gene or on the promoter region; (2) Principal Component Analysis (PCA), in which the first k components having 1% or more of the total sample variance are considered in downstream analysis; (3) Computing a correlation matrix with pairwise gene correlations, represented as a vector of PCs; (4) analyzing the correlation matrix for the network; and (5) The contribution of each gene to the discrimination of the phenotype assessed in terms of the proportion of cumulative variance from the total sample variance that the gene carries.

39. 39. The computerized heuristic of claim 38, wherein the ID is selected from the group consisting of Hellinger divergence / distance, J-divergence, and total variation distance.

40. 40. The computerized heuristic of claim 39, wherein the ID is J-divergence.

41. 39. The computerized heuristic of claim 38, wherein the PCA is applied using a function such that genes are represented as k-dimensional vectors of PCs, and further wherein the square of each coordinate has the contribution of said vector with respect to variance to treatment discrimination from a control group.

42. 39. The computerized heuristic of claim 38, wherein the PCA is applied using a pcaLDA function.

43. 39. The computerized heuristic of claim 38, wherein the correlation matrix comprises a weighted correlation network (WCN), the WCN and the network being analyzed, and the network is a protein-protein interaction (PPI) network.

44. 44. The computerized heuristic of claim 43, further comprising comparing results from the WCN and the PPI network with identified consistent relationships and epigenetic gene contributions to the phenotype.

45. 39. The computerized heuristic of claim 38, further comprising a magnitude calculated as the Euclidean norm of the gene represented as a vector of k PCs.

46. 39. The computerized heuristic of claim 38, wherein the network is associated with one or more biological processes.

47. 1. A method for reducing computational load, the method comprising: using methylation analysis to identify differentially methylated genes (DMGs) on a stretch of DNA sequence; Integrating the DMG into a gene network; identifying a network hub.

48. 48. The method of claim 47, wherein the network hubs are identified through protein-protein interaction (PPI) network analysis and weighted correlation network (WCN) analysis.

49. 49. The method of claim 48, further comprising comparing the results from the WCN and the PPI network with the identified consistent relationships and contributions of epigenetic genes to phenotype.

50. 48. The method of claim 47, wherein the network hub is associated with one or more biological processes.

Images