Simple DNA Sequences Research Articles

Evidence for compensatory evolution within pleiotropic regulatory elements.

Pleiotropy, measured as expression breadth across tissues, is one of the best predictors for protein sequence and expression conservation. In this study, we investigated its effect on the evolution of cis-regulatory elements (CREs). To this end, we carefully reanalyzed the Epigenomics Roadmap data for nine fetal tissues, assigning a measure of pleiotropic degree to nearly half a million CREs. To assess the functional conservation of CREs, we generated ATAC-seq and RNA-seq data from humans and macaques. We found that more pleiotropic CREs exhibit greater conservation in accessibility, and the mRNA expression levels of the associated genes are more conserved. This trend of higher conservation for higher degrees of pleiotropy persists when analyzing the transcription factor binding repertoire. In contrast, simple DNA sequence conservation of orthologous sites between species tends to be even lower for pleiotropic CREs than for species-specific CREs. Combining various lines of evidence, we propose that the lack of sequence conservation in functionally conserved pleiotropic CREs is owing to within-element compensatory evolution. In summary, our findings suggest that pleiotropy is also a good predictor for the functional conservation of CREs, even though this is not reflected in the sequence conservation of pleiotropic CREs.

GENet: A Graph-Based Model Leveraging Histone Marks and Transcription Factors for Enhanced Gene Expression Prediction.

Understanding the regulatory mechanisms of gene expression is a crucial objective in genomics. Although the DNA sequence near the transcription start site (TSS) offers valuable insights, recent methods suggest that analyzing only the surrounding DNA may not suffice to accurately predict gene expression levels. We developed GENet (Gene Expression Network from Histone and Transcription Factor Integration), a novel approach that integrates essential regulatory signals from transcription factors and histone modifications into a graph-based model. GENet extends beyond simple DNA sequence analysis by incorporating additional layers of genetic control, which are vital for determining gene expression. Our method markedly enhances the prediction of mRNA levels compared to previous models that depend solely on DNA sequence data. The results underscore the significance of including comprehensive regulatory information in gene expression studies. GENet emerges as a promising tool for researchers, with potential applications extending from fundamental biological research to the development of medical therapies.

Mismatch repair enzymes regulate telomere recombination in Saccharomycescerevisiae

DNA mismatch repair (MMR) is a crucial mechanism that ensures chromosome stability and prevents the development of various human cancers. Apart from its role in correcting mismatches during DNA replication, MMR also plays a significant role in regulating recombination between non-identical sequences, a process known as homeologous recombination. Telomeres, the protective ends of eukaryotic chromosomes, possess sequences that are not perfectly homologous. While telomerase primarily maintains telomere length in the yeast Saccharomyces cerevisiae, recombination between telomeres becomes a major pathway for length maintenance in cells lacking telomerase. This study investigates the participation of MMR in telomere recombination. Our findings reveal that mutations in MMR genes activate type I recombination. Notably, among the MMR proteins, MutSα (Msh2 and Msh6) and MutLα (Mlh1 and Pms1) exerted the most pronounced effects on telomere recombination. We also found that yeast cells containing simple human telomeric TTAGGG DNA sequences preferentially utilize type II recombination to maintain their telomeres, highlighting the influence of the heterogeneous nature of yeast telomeric sequences on type II recombination. Furthermore, our observations indicate that MMR activity is indispensable for its impact on telomere recombination. Collectively, these results contribute to a more comprehensive understanding of the role of MMR in telomere recombination.

Examination of phase-variable haemoglobin-haptoglobin binding proteins in non-typeable Haemophilus influenzae reveals a diverse distribution of multiple variants.

Non-typeable Haemophilus influenzae (NTHi) is a major human pathogen for which there is no globally licensed vaccine. NTHi has a strict growth requirement for iron and encodes several systems to scavenge elemental iron and heme from the host. An effective NTHi vaccine would target conserved, essential surface factors, such as those involved in iron acquisition. Haemoglobin–haptoglobin binding proteins (Hgps) are iron-uptake proteins localized on the outer-membrane of NTHi. If the Hgps are to be included as components of a rationally designed subunit vaccine against NTHi, it is important to understand their prevalence and diversity. Following analysis of all available Hgp sequences, we propose a standardized grouping method for Hgps, and demonstrate increased diversity of these proteins than previously determined. This analysis demonstrated that genes encoding variants HgpB and HgpC are present in all strains examined, and almost 40% of strains had a duplicate, nonidentical hgpB gene. Hgps are also phase-variably expressed; the encoding genes contain a CCAA(n) simple DNA sequence repeat tract, resulting in biphasic ON–OFF switching of expression. Examination of the ON–OFF state of hgpB and hgpC genes in a collection of invasive NTHi isolates demonstrated that 58% of isolates had at least one of hgpB or hgpC expressed (ON). Varying expression of a diverse repertoire of hgp genes would provide strains a method of evading an immune response while maintaining the ability to acquire iron via heme. Structural analysis of Hgps also revealed high sequence variability at the sites predicted to be surface exposed, demonstrating a further mechanism to evade the immune system—through varying the surface, immune-exposed regions of the membrane anchored protein. This information will direct and inform the choice of candidates to include in a vaccine against NTHi.

Can one detect intermediate denaturation states of DNA sequences by following the equilibrium open-close dynamic fluctuations of a single base pair?

Melting of DNA sequences may occur through a few major intermediate states, whose influence on the melting curve has been discussed previously, while their effect on the kinetics has not been explored thoroughly. Here, we chose a simple DNA sequence, forming a hairpin in its native (zipped) state, and study it using molecular dynamic (MD) simulations and a model integrating the Gaussian network model with bond-binding energies-the Gaussian binding energy (GBE) model. We find two major partial denaturation states, a bubble state and a partial unzipping state. We demonstrate the influence of these two states on the closing-opening base pair dynamics, as probed by a tagged bond auto-correlation function (ACF). We argue that the latter is measured by fluorescence correlation spectroscopy experiments, in which one base of the pair is linked to a fluorescent dye, while the complementary base is linked to a quencher, similar to the experiment reported by Altan-Bonnet et al. [Phys. Rev. Lett. 90, 138101 (2003)]. We find that tagging certain base pairs at temperatures around the melting temperature results in a multi-step relaxation of the ACF, while tagging other base pairs leads to an effectively single-step relaxation, albeit non-exponential. Only the latter type of relaxation has been observed experimentally, and we suggest which of the other base pairs should be tagged in order to observe multi-step relaxation. We demonstrate that this behavior can be observed with other sequences and argue that the GBE can reliably predict these dynamics for very long sequences, where MD simulations might be limited.

Characterization of the Phase-Variable Autotransporter Lav Reveals a Role in Host Cell Adherence and Biofilm Formation in Nontypeable Haemophilus influenzae.

ABSTRACTLav is an autotransporter protein found in pathogenic Haemophilus and Neisseria species. Lav in nontypeable Haemophilus influenzae (NTHi) is phase-variable: the gene reversibly switches ON-OFF via changes in length of a locus-located GCAA(n) simple DNA sequence repeat tract. The expression status of lav was examined in carriage and invasive collections of NTHi, where it was predominantly not expressed (OFF). Phenotypic study showed lav expression (ON) results in increased adherence to human lung cells and denser biofilm formation. A survey of Haemophilus species genome sequences showed lav is present in ∼60% of NTHi strains, but lav is not present in most typeable H. influenzae strains. Sequence analysis revealed a total of five distinct variants of the Lav passenger domain present in Haemophilus spp., with these five variants showing a distinct lineage distribution. Determining the role of Lav in NTHi will help understand the role of this protein during distinct pathologies.

Chronic Exposure to Cadmium and Antioxidants Does Not Affect the Dynamics of Expanded CAG•CTG Trinucleotide Repeats in a Mouse Cell Culture System of Unstable DNA.

More than 30 human disorders are caused by the expansion of simple sequence DNA repeats, among which triplet repeats remain the most frequent. Most trinucleotide repeat expansion disorders affect primarily the nervous system, through mechanisms of neurodysfunction and/or neurodegeneration. While trinucleotide repeat tracts are short and stably transmitted in unaffected individuals, disease-associated expansions are highly dynamic in the germline and in somatic cells, with a tendency toward further expansion. Since longer repeats are associated with increasing disease severity and earlier onset of symptoms, intergenerational repeat size gains account for the phenomenon of anticipation. In turn, higher levels of age-dependent somatic expansion have been linked with increased disease severity and earlier age of onset, implicating somatic instability in the onset and progression of disease symptoms. Hence, tackling the root cause of symptoms through the control of repeat dynamics may provide therapeutic modulation of clinical manifestations. DNA repair pathways have been firmly implicated in the molecular mechanism of repeat length mutation. The demonstration that repeat expansion depends on functional DNA mismatch repair (MMR) proteins, points to MMR as a potential therapeutic target. Similarly, a role of DNA base excision repair (BER) in repeat expansion has also been suggested, particularly during the removal of oxidative lesions. Using a well-characterized mouse cell model system of an unstable CAG•CTG trinucleotide repeat, we tested if expanded repeat tracts can be stabilized by small molecules with reported roles in both pathways: cadmium (an inhibitor of MMR activity) and a variety of antioxidants (capable of neutralizing oxidative species). We found that chronic exposure to sublethal doses of cadmium and antioxidants did not result in significant reduction of the rate of trinucleotide repeat expansion. Surprisingly, manganese yielded a significant stabilization of the triplet repeat tract. We conclude that treatment with cadmium and antioxidants, at doses that do not interfere with cell survival and cell culture dynamics, is not sufficient to modify trinucleotide repeat dynamics in cell culture.

Generation of novelties in the genus Ocimum as a result of natural hybridization: A morphological, genetical and chemical appraisal

The genus Ocimum is a boutique of a diverse set of specialized metabolites such as terpenoids and phenylpropanoids. Each Ocimum species and its cultivars represent a characteristic chemical profile. The present study explored the two interspecific Ocimum hybrids originating through a serendipitous natural cross between O. kilimandscharicum and O. basilicum. These two novel Ocimum hybrids exhibited intermediate morphological features of two parental species. Inter simple sequence repeats (ISSR) analysis and DNA barcoding with the plastid non-coding trnH-psbA intergenic spacer region reaffirmed unambiguous parental identification and differentiation of these natural hybrids from other available Ocimum species. Consequently, gas chromatography-mass spectrometry-based metabolite profiling of two hybrids identified them as specific chemotypes with the presence of a unique blend of specialized metabolites from the parental species, which are either rich in terpenes or phenylpropanoids. Additionally, expression analysis of key genes from terpenoid and phenylpropanoid pathways corroborated with differential metabolite accumulation. Thus, these two Ocimum hybrids represented the novel chemotypes, which could be useful in commercial cultivation to produce novel essential oil and bioactive constituents. Further, the resulting metabolite diversity could have potential industrial applications in the areas of healthcare, bioremediation, and crop protection.

Moraxella catarrhalis phase-variable loci show differences in expression during conditions relevant to disease.

Moraxella catarrhalis is a human-adapted, opportunistic bacterial pathogen of the respiratory mucosa. Although asymptomatic colonization of the nasopharynx is common, M. catarrhalis can ascend into the middle ear, where it is a prevalent causative agent of otitis media in children, or enter the lower respiratory tract, where it is associated with acute exacerbations of chronic obstructive pulmonary disease in adults. Phase variation is the high frequency, random, reversible switching of gene expression that allows bacteria to adapt to different host microenvironments and evade host defences, and is most commonly mediated by simple DNA sequence repeats. Bioinformatic analysis of five closed M. catarrhalis genomes identified 17 unique simple DNA sequence repeat tracts that were variable between strains, indicating the potential to mediate phase variable expression of the associated genes. Assays designed to assess simple sequence repeat variation under conditions mimicking host infection demonstrated that phase variation of uspA1 (ubiquitous surface protein A1) from high to low expression occurs over 72 hours of biofilm passage, while phase variation of uspA2 (ubiquitous surface protein A2) to high expression variants occurs during repeated exposure to human serum, as measured by mRNA levels. We also identify and confirm the variable expression of two novel phase variable genes encoding a Type III DNA methyltransferase (modO), and a conserved hypothetical permease (MC25239_RS00020). These data reveal the repertoire of phase variable genes mediated by simple sequence repeats in M. catarrhalis and demonstrate that modulation of expression under conditions mimicking human infection is attributed to changes in simple sequence repeat length.

DNA sequence repeats identify numerous Type I restriction‐modification systems that are potential epigenetic regulators controlling phase‐variable regulons; phasevarions

Over recent years several examples of randomly switching methyltransferases, associated with Type III restriction‐modification (R‐M) systems, have been described in pathogenic bacteria. In every case examined, changes in simple DNA sequence repeats result in variable methyltransferase expression and result in global changes in gene expression, and differentiation of the bacterial cell into distinct phenotypes. These epigenetic regulatory systems are called phasevarions, phase‐variable regulons, and are widespread in bacteria, with 17.4% of Type III R‐M system containing simple DNA sequence repeats. A distinct, recombination‐driven random switching system has also been described in Streptococci in Type I R‐M systems that also regulate gene expression. Here, we interrogate the most extensive and well‐curated database of R‐M systems, REBASE, by searching for all possible simple DNA sequence repeats in the hsdRMS genes that encode Type I R‐M systems. We report that 7.9% of hsdS, 2% of hsdM, and of 4.3% of hsdR genes contain simple sequence repeats that are capable of mediating phase variation. Phase variation of both hsdM and hsdS genes will lead to differential methyltransferase expression or specificity, and thereby the potential to control phasevarions. These data suggest that in addition to well characterized phasevarions controlled by Type III mod genes, and the previously described Streptococcal Type I R‐M systems that switch via recombination, approximately 10% of all Type I R‐M systems surveyed herein have independently evolved the ability to randomly switch expression via simple DNA sequence repeats.

How many came home? Evaluating ex situ conservation of green turtles in the Cayman Islands.

Ex situ management is an important conservation tool that allows the preservation of biological diversity outside natural habitats while supporting survival in the wild. Captive breeding followed by re-introduction is a possible approach for endangered species conservation and preservation of genetic variability. The Cayman Turtle Centre Ltd was established in 1968 to market green turtle (Chelonia mydas) meat and other products and replenish wild populations, thought to be locally extirpated, through captive breeding. We evaluated the effects of this re-introduction programmme using molecular markers (13 microsatellites, 800-bp D-loop and simple tandem repeat mitochondrial DNA sequences) from captive breeders (N=257) and wild nesting females (N=57) (sampling period: 2013-2015). We divided the captive breeders into three groups: founders (from the original stock), and then two subdivisions of F1 individuals corresponding to two different management strategies, cohort 1995 ("C1995") and multicohort F1 ("MCF1"). Loss of genetic variability and increased relatedness was observed in the captive stock over time. We found no significant differences in diversity among captive and wild groups, and similar or higher levels of haplotype variability when compared to other natural populations. Using parentage and sibship assignment, we determined that 90% of the wild individuals were related to the captive stock. Our results suggest a strong impact of the re-introduction programmme on the present recovery of the wild green turtle population nesting in the Cayman Islands. Moreover, genetic relatedness analyses of captive populations are necessary to improve future management actions to maintain genetic diversity in the long term and avoid inbreeding depression.

Genetic diversity of ‘Candidatus Liberibacter asiaticus’ in Brazil analyzed in different geographic regions and citrus varieties

The bacterium ‘Candidatus Liberibacter asiaticus’ - CLas, the agent associated with citrus huanglongbing disease was first reported in Sao Paulo State in 2004 and has spread throughout the citrus-growing regions of Sao Paulo and further to the states of Minas Gerais and Parana. However, little information is available regarding the genetic diversity of CLas since its introduction. Understanding the genetic diversity of this bacterium is important for tracing migration routes and for identifying evolutionary selection forces that may affect the genetic diversity of this pathogen. Total DNA from 199 HLB-diseased citrus species trees was sampled from geographic regions in Sao Paulo, Parana and Minas Gerais states and the CLas isolates were genotyped by simple sequence DNA repeats (SSR). Nei’s genetic diversity index was observed to be low in all populations (HNei = 0.11–0.26). Wright’s fixation index (FST), which measures population genetic differentiation, did not differ significantly between CLas populations from Sao Paulo State and from Minas Gerais, but significant values (FST = 0.118–0.191) of the Parana CLas population distinguish it from the others. Interestingly, higher values (FST = 0.275–0.445) were observed for the CLas populations obtained from different citrus species compared to sweet orange, suggesting that the citrus genotypes could be driven the genetic diversity of CLas. Clustering analysis supports the FST results that split the CLas samples into three genetically distinct populations. These results indicate that genetically homogeneous populations of CLas infect sweet orange plants in various regions of Sao Paulo State and Minas Gerais, but not Parana, suggesting that different introduction events may have occurred for the Sao Paulo and Parana states.

Efficient coralline algal psbA mini barcoding and High Resolution Melt (HRM) analysis using a simple custom DNA preparation

Coralline algae form extensive maerl and rhodolith habitats that support a rich biodiversity. Calcium carbonate harvesting as well as trawling activities threatens this ecosystem. Eleven species were recorded so far as maerl-forming in NE Atlantic, but identification based on morphological characters is unreliable. As for most red algae, we now use molecular characters to resolve identification of these taxa. However, obtaining DNA sequences requires time and resource demanding methods. The purpose of our study was to improve methods for achieving simple DNA extraction, amplification, sequencing and sequence analysis to allow robust identification of maerl species and other coralline algae. Our novel and easy DNA preparation method for coralline algae was of sufficient quality for qPCR amplification and sequencing of all 47 tested samples. The new psbA qPCR assay successfully amplified a 350 bp fragment identifying six species and uncovering two new Operational Taxonomic Units. Molecular results were corroborated with anatomical examination using i.e. scanning electron microscopy. Finally, the qPCR assay was coupled with High Resolution Melt analysis that successfully differentiated the closely related species Lithothamnion erinaceum and L. cf. glaciale. This DNA preparation and qPCR technique should vitalize coralline research by reducing time and cost associated with molecular systematics.

Ionic Polypeptide Polymers with Unusual β-Sheet Stability.

Incorporating charged amino acid side chains in polypeptide polymer backbones to improve solubility usually leads to reduced secondary structuring. Here we show that highly water soluble (>15 mg.mL-1) β-sheetscan be obtained via nucleotide monophosphate grafting onto simple poly(γ-propargyl- L-glutamate) backbone. This synthetic methodology has been applied to the synthesis of thymidine-based nucleopolypeptides presenting stable β-sheet conformation in aqueous solutions with pH values comprised between 4 and 8. These polymeric analogues of nucleoproteins exhibited selective interaction with simple DNA sequences displaying adenine.

A survey of Type III restriction-modification systems reveals numerous, novel epigenetic regulators controlling phase-variable regulons; phasevarions.

Many bacteria utilize simple DNA sequence repeats as a mechanism to randomly switch genes on and off. This process is called phase variation. Several phase-variable N6-adenine DNA-methyltransferases from Type III restriction-modification systems have been reported in bacterial pathogens. Random switching of DNA methyltransferases changes the global DNA methylation pattern, leading to changes in gene expression. These epigenetic regulatory systems are called phasevarions — phase-variable regulons. The extent of these phase-variable genes in the bacterial kingdom is unknown. Here, we interrogated a database of restriction-modification systems, REBASE, by searching for all simple DNA sequence repeats in mod genes that encode Type III N6-adenine DNA-methyltransferases. We report that 17.4% of Type III mod genes (662/3805) contain simple sequence repeats. Of these, only one-fifth have been previously identified. The newly discovered examples are widely distributed and include many examples in opportunistic pathogens as well as in environmental species. In many cases, multiple phasevarions exist in one genome, with examples of up to 4 independent phasevarions in some species. We found several new types of phase-variable mod genes, including the first example of a phase-variable methyltransferase in pathogenic Escherichia coli. Phasevarions are a common epigenetic regulation contingency strategy used by both pathogenic and non-pathogenic bacteria.

Low genetic variation of invasive Fallopia spp. in their northernmost European distribution range.

Knowledge about the reproduction strategies of invasive species is fundamental for effective control. The invasive Fallopia taxa (Japanese knotweed s.l.) reproduce mainly clonally in Europe, and preventing spread of vegetative fragments is the most important control measure. However, high levels of genetic variation within the hybrid F. × bohemica indicate that hybridization and seed dispersal could be important. In Norway in northern Europe, it is assumed that these taxa do not reproduce sexually due to low temperatures in the autumn when the plants are flowering. The main objective of this study was to examine the genetic variation of invasive Fallopia taxa in selected areas in Norway in order to evaluate whether the taxa may reproduce by seeds in their most northerly distribution range in Europe. Fallopia stands from different localities in Norway were analyzed with respect to prevalence of taxa, and genetic variation within and between taxa was studied using amplified fragment length polymorphism (AFLP). Taxonomic identification based on morphology corresponded with identification based on simple sequence repeats (SSR) and DNA ploidy levels (8× F. japonica, 6× F. × bohemica and 4× F. sachalinensis). No genetic variation within F. japonica was detected. All F. × bohemica samples belonged to a single AFLP genotype, but one sample had a different SSR genotype. Two SSR genotypes of F. sachalinensis were also detected. Extremely low genetic variation within the invasive Fallopia taxa indicates that these taxa do not reproduce sexually in the region, suggesting that control efforts can be focused on preventing clonal spread. Climate warming may increase sexual reproduction of invasive Fallopia taxa in northern regions. The hermaphrodite F. × bohemica is a potential pollen source for the male‐sterile parental species. Targeted eradication of the hybrid can therefore reduce the risk of increased sexual reproduction under future warmer climate.

Telomeres and telomerase in basal Metazoa

Telomeres are protein-bound tandemly repeated simple DNA sequences placed at the chromosome ends, their role being essential for maintaining genome integrity. Severe telomere damage can trigger potential carcinogenic events such as chromosome fusions, whilst telomere shortening results (at least in mammals) in a protective mechanism known as 'replicative senescence'. In most Metazoa, telomere shortening is avoided by the ribonucleoprotein telomerase. In this brief overview, we focused on the evolutionary conservation of telomeres and telomerase in basal Metazoa (Ctenophora, Porifera, Placozoa, Cnidaria). In all these taxa, telomerase seems to be active and telomeres show the canonical TTAGGG sequence. Presence of telomerase activity and absence of telomere shortening is in accordance with the lack of senescence seen in basal Metazoa, although a clear correspondence between the “demographic senescence” and “replicative senescence” remains to be elucidated. Nonetheless, basal Metazoa can be useful as model organisms in studies on the evolution of telomere biology, Evo-Devo investigations on the control of telomerase activity, the emergence of senescence, as well as telomere-independent effects of telomerase.

Ethidium Bromide Modifies The Agarose Electrophoretic Mobility of CAG•CTG Alternative DNA Structures Generated by PCR.

The abnormal expansion of unstable simple sequence DNA repeats can cause human disease through a variety of mechanisms, including gene loss-of-function, toxic gain-of-function of the encoded protein and toxicity of the repeat-containing RNA transcript. Disease-associated unstable DNA repeats display unusual biophysical properties, including the ability to adopt non-B-DNA structures. CAG•CTG trinucleotide sequences, in particular, have been most extensively studied and they can fold into slipped-stranded DNA structures, which have been proposed as mutation intermediates in repeat size expansion. Here, we describe a simple assay to detect unusual DNA structures generated by PCR amplification, based on their slow electrophoretic migration in agarose and on the effects of ethidium bromide on the mobility of structural isoforms through agarose gels. Notably, the inclusion of ethidium bromide in agarose gels and running buffer eliminates the detection of additional slow-migrating DNA species, which are detected in the absence of the intercalating dye and may be incorrectly classified as mutant alleles with larger than actual expansion sizes. Denaturing and re-annealing experiments confirmed the slipped-stranded nature of the additional DNA species observed in agarose gels. Thus, we have shown that genuine non-B-DNA conformations are generated during standard PCR amplification of CAG•CTG sequences and detected by agarose gel electrophoresis. In contrast, ethidium bromide does not change the multi-band electrophoretic profiles of repeat-containing PCR products through native polyacrylamide gels. These data have implications for the analysis of trinucleotide repeat DNA and possibly other types of unstable repetitive DNA sequences by standard agarose gel electrophoresis in diagnostic and research protocols. We suggest that proper sizing of CAG•CTG PCR products in agarose gels should be performed in the presence of ethidium bromide.

Interference between Triplex and Protein Binding to Distal Sites on Supercoiled DNA

We have explored the interdependence of the binding of a DNA triplex and a repressor protein to distal recognition sites on supercoiled DNA minicircles using MD simulations. We observe that the interaction between the two ligands through their influence on their DNA template is determined by a subtle interplay of DNA mechanics and electrostatics, that the changes in flexibility induced by ligand binding play an important role and that supercoiling can instigate additional ligand-DNA contacts that would not be possible in simple linear DNA sequences.

Improved Prediction of Non-methylated Islands in Vertebrates Highlights Different Characteristic Sequence Patterns.

Non-methylated islands (NMIs) of DNA are genomic regions that are important for gene regulation and development. A recent study of genome-wide non-methylation data in vertebrates by Long et al. (eLife 2013;2:e00348) has shown that many experimentally identified non-methylated regions do not overlap with classically defined CpG islands which are computationally predicted using simple DNA sequence features. This is especially true in cold-blooded vertebrates such as Danio rerio (zebrafish). In order to investigate how predictive DNA sequence is of a region’s methylation status, we applied a supervised learning approach using a spectrum kernel support vector machine, to see if a more complex model and supervised learning can be used to improve non-methylated island prediction and to understand the sequence properties of these regions. We demonstrate that DNA sequence is highly predictive of methylation status, and that in contrast to existing CpG island prediction methods our method is able to provide more useful predictions of NMIs genome-wide in all vertebrate organisms that were studied. Our results also show that in cold-blooded vertebrates (Anolis carolinensis, Xenopus tropicalis and Danio rerio) where genome-wide classical CpG island predictions consist primarily of false positives, longer primarily AT-rich DNA sequence features are able to identify these regions much more accurately.