DNA Sequence Features Research Articles

DNA sequence properties that predict susceptibility to epiallelic switching

Transgenerationally heritable epialleles are defined by the stable propagation of alternative transcriptional states through mitotic and meiotic cell cycles. Given that the propagation of DNA methylation at CpG sites, mediated in Arabidopsis by MET1, plays a central role in epigenetic inheritance, we examined genomewide DNA methylation in partial and complete loss‐of‐function met1 mutants. We interpreted the data in relation to transgenerational epiallelic stability, which allowed us to classify chromosomal targets of epigenetic regulation into (i) single copy and methylated exclusively at CpGs, readily forming epialleles, and (ii) transposon‐derived, methylated at all cytosines, which may or may not form epialleles. We provide evidence that DNA sequence features such as density of CpGs and genomic repetitiveness of the loci predispose their susceptibility to epiallelic switching. The importance and predictive power of these genetic features were confirmed by analyses of common epialleles in natural Arabidopsis accessions, epigenetic recombinant inbred lines (epiRILs) and also verified in rice.

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.

A genome-integrated massively parallel reporter assay reveals DNA sequence determinants of cis-regulatory activity in neural cells.

Recent large-scale genomics efforts to characterize the cis-regulatory sequences that orchestrate genome-wide expression patterns have produced impressive catalogues of putative regulatory elements. Most of these sequences have not been functionally tested, and our limited understanding of the non-coding genome prevents us from predicting which sequences are bona fide cis-regulatory elements. Recently, massively parallel reporter assays (MPRAs) have been deployed to measure the activity of putative cis-regulatory sequences in several biological contexts, each with specific advantages and distinct limitations. We developed LV-MPRA, a novel lentiviral-based, massively parallel reporter gene assay, to study the function of genome-integrated regulatory elements in any mammalian cell type; thus, making it possible to apply MPRAs in more biologically relevant contexts. We measured the activity of 2,600 sequences in U87 glioblastoma cells and human neural progenitor cells (hNPCs) and explored how regulatory activity is encoded in DNA sequence. We demonstrate that LV-MPRA can be applied to estimate the effects of local DNA sequence and regional chromatin on regulatory activity. Our data reveal that primary DNA sequence features, such as GC content and dinucleotide composition, accurately distinguish sequences with high activity from sequences with low activity in a full chromosomal context, and may also function in combination with different transcription factor binding sites to determine cell type specificity. We conclude that LV-MPRA will be an important tool for identifying cis-regulatory elements and stimulating new understanding about how the non-coding genome encodes information.

Dissecting relative contributions of cis- and trans-determinants to nucleosome distribution by comparing Tetrahymena macronuclear and micronuclear chromatin.

The ciliate protozoan Tetrahymena thermophila contains two types of structurally and functionally differentiated nuclei: the transcriptionally active somatic macronucleus (MAC) and the transcriptionally silent germ-line micronucleus (MIC). Here, we demonstrate that MAC features well-positioned nucleosomes downstream of transcription start sites and flanking splice sites. Transcription-associated trans-determinants promote nucleosome positioning in MAC. By contrast, nucleosomes in MIC are dramatically delocalized. Nucleosome occupancy in MAC and MIC are nonetheless highly correlated with each other, as well as with in vitro reconstitution and predictions based upon DNA sequence features, revealing unexpectedly strong contributions from cis-determinants. In particular, well-positioned nucleosomes are often matched with GC content oscillations. As many nucleosomes are coordinately accommodated by both cis- and trans-determinants, we propose that their distribution is shaped by the impact of these nucleosomes on the mutational and transcriptional landscape, and driven by evolutionary selection.

A synergistic DNA logic predicts genome-wide chromatin accessibility.

Enhancers and promoters commonly occur in accessible chromatin characterized by depleted nucleosome contact; however, it is unclear how chromatin accessibility is governed. We show that log-additive cis-acting DNA sequence features can predict chromatin accessibility at high spatial resolution. We develop a new type of high-dimensional machine learning model, the Synergistic Chromatin Model (SCM), which when trained with DNase-seq data for a cell type is capable of predicting expected read counts of genome-wide chromatin accessibility at every base from DNA sequence alone, with the highest accuracy at hypersensitive sites shared across cell types. We confirm that a SCM accurately predicts chromatin accessibility for thousands of synthetic DNA sequences using a novel CRISPR-based method of highly efficient site-specific DNA library integration. SCMs are directly interpretable and reveal that a logic based on local, nonspecific synergistic effects, largely among pioneer TFs, is sufficient to predict a large fraction of cellular chromatin accessibility in a wide variety of cell types.

Sequence features accurately predict genome-wide MeCP2 binding in vivo.

Methyl-CpG binding protein 2 (MeCP2) is critical for proper brain development and expressed at near-histone levels in neurons, but the mechanism of its genomic localization remains poorly understood. Using high-resolution MeCP2-binding data, we show that DNA sequence features alone can predict binding with 88% accuracy. Integrating MeCP2 binding and DNA methylation in a probabilistic graphical model, we demonstrate that previously reported genome-wide association with methylation is in part due to MeCP2's affinity to GC-rich chromatin, a result replicated using published data. Furthermore, MeCP2 co-localizes with nucleosomes. Finally, MeCP2 binding downstream of promoters correlates with increased expression in Mecp2-deficient neurons.

Prediction of fine-tuned promoter activity from DNA sequence.

The quantitative prediction of transcriptional activity of genes using promoter sequence is fundamental to the engineering of biological systems for industrial purposes and understanding the natural variation in gene expression. To catalyze the development of new algorithms for this purpose, the Dialogue on Reverse Engineering Assessment and Methods (DREAM) organized a community challenge seeking predictive models of promoter activity given normalized promoter activity data for 90 ribosomal protein promoters driving expression of a fluorescent reporter gene. By developing an unbiased modeling approach that performs an iterative search for predictive DNA sequence features using the frequencies of various k-mers, inferred DNA mechanical properties and spatial positions of promoter sequences, we achieved the best performer status in this challenge. The specific predictive features used in the model included the frequency of the nucleotide G, the length of polymeric tracts of T and TA, the frequencies of 6 distinct trinucleotides and 12 tetranucleotides, and the predicted protein deformability of the DNA sequence. Our method accurately predicted the activity of 20 natural variants of ribosomal protein promoters (Spearman correlation r = 0.73) as compared to 33 laboratory-mutated variants of the promoters (r = 0.57) in a test set that was hidden from participants. Notably, our model differed substantially from the rest in 2 main ways: i) it did not explicitly utilize transcription factor binding information implying that subtle DNA sequence features are highly associated with gene expression, and ii) it was entirely based on features extracted exclusively from the 100 bp region upstream from the translational start site demonstrating that this region encodes much of the overall promoter activity. The findings from this study have important implications for the engineering of predictable gene expression systems and the evolution of gene expression in naturally occurring biological systems.

Abstract PR01: Structural mechanism of sequence-specific 5-methylcytosine (5mC) recognition by AP-1 transcription factors

Abstract DNA methylation is generally known to inhibit transcription factor binding. However, this notion has been challenged with identifications of several transcription factors that preferentially activate methylated promoters in specific sequences. AP-1 transcription factor (c-Jun/c-Fos heterodimer), a master regulator of a wide range of cancer-related genes, has been shown to bind a novel consensus sequence (5'-TGACTCG-3') in CpG methylation-dependent manner. In the fluorescence polarization-based DNA binding assay, both c-Jun/c-Fos heterodimer and c-Jun/c-Jun homodimer preferentially bind DNA containing the methylated consensus sequence compared to DNA with unmethylated sequence. Using a homodimer of c-Jun/c-Jun basic leucine zipper (bZIP) DNA binding domain, we have solved the X-ray structure of AP-1 in complex with an oligonucleotide containing the methylated consensus sequence. The structure reveals that a common motif conserved among bZIP proteins is involved in the specific recognition of methylated CpG by one monomer and TpG by the other monomer. Like 5mC, thymine (5-methyluracil) contains a methyl group at C5 position, and the similarity imposed by the methyl C5 preserves key structural features of DNA sequence recognized by AP-1. Our in vitro finding resonates with previous studies of AP-1-like transcription factors that preferentially bind methylated response elements. In light of such evidence, our characterization of AP-1 as a sequence-specific 5mCpG and TpG reader suggests a novel way in which DNA methylation may function to activate transcription in a methylation-dependent manner. Citation Format: Samuel Hong, Xiaodong Cheng. Structural mechanism of sequence-specific 5-methylcytosine (5mC) recognition by AP-1 transcription factors. [abstract]. In: Proceedings of the AACR Special Conference on Chromatin and Epigenetics in Cancer; Sep 24-27, 2015; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2016;76(2 Suppl):Abstract nr PR01.

DNAshapeR: an R/Bioconductor package for DNA shape prediction and feature encoding.

Summary: DNAshapeR predicts DNA shape features in an ultra-fast, high-throughput manner from genomic sequencing data. The package takes either nucleotide sequence or genomic coordinates as input and generates various graphical representations for visualization and further analysis. DNAshapeR further encodes DNA sequence and shape features as user-defined combinations of k-mer and DNA shape features. The resulting feature matrices can be readily used as input of various machine learning software packages for further modeling studies.Availability and implementation: The DNAshapeR software package was implemented in the statistical programming language R and is freely available through the Bioconductor project at https://www.bioconductor.org/packages/devel/bioc/html/DNAshapeR.html and at the GitHub developer site, http://tsupeichiu.github.io/DNAshapeR/.Contact: rohs@usc.eduSupplementary information: Supplementary data are available at Bioinformatics online.

Transcription factor binding site prediction in vivousing DNA sequence and shape features

Transcription factor binding site prediction in vivousing DNA sequence and shape features

Standards for plant synthetic biology: a common syntax for exchange of DNA parts.

Inventors in the field of mechanical and electronic engineering can access multitudes of components and, thanks to standardization, parts from different manufacturers can be used in combination with each other. The introduction of BioBrick standards for the assembly of characterized DNA sequences was a landmark in microbial engineering, shaping the field of synthetic biology. Here, we describe a standard for Type IIS restriction endonuclease-mediated assembly, defining a common syntax of 12 fusion sites to enable the facile assembly of eukaryotic transcriptional units. This standard has been developed and agreed by representatives and leaders of the international plant science and synthetic biology communities, including inventors, developers and adopters of Type IIS cloning methods. Our vision is of an extensive catalogue of standardized, characterized DNA parts that will accelerate plant bioengineering.

Sequence determinants of improved CRISPR sgRNA design.

The CRISPR/Cas9 system has revolutionized mammalian somatic cell genetics. Genome-wide functional screens using CRISPR/Cas9-mediated knockout or dCas9 fusion-mediated inhibition/activation (CRISPRi/a) are powerful techniques for discovering phenotype-associated gene function. We systematically assessed the DNA sequence features that contribute to single guide RNA (sgRNA) efficiency in CRISPR-based screens. Leveraging the information from multiple designs, we derived a new sequence model for predicting sgRNA efficiency in CRISPR/Cas9 knockout experiments. Our model confirmed known features and suggested new features including a preference for cytosine at the cleavage site. The model was experimentally validated for sgRNA-mediated mutation rate and protein knockout efficiency. Tested on independent data sets, the model achieved significant results in both positive and negative selection conditions and outperformed existing models. We also found that the sequence preference for CRISPRi/a is substantially different from that for CRISPR/Cas9 knockout and propose a new model for predicting sgRNA efficiency in CRISPRi/a experiments. These results facilitate the genome-wide design of improved sgRNA for both knockout and CRISPRi/a studies.

Representation of DNA sequences in genetic codon context with applications in exon and intron prediction.

To apply digital signal processing (DSP) methods to analyze DNA sequences, the sequences first must be specially mapped into numerical sequences. Thus, effective numerical mappings of DNA sequences play key roles in the effectiveness of DSP-based methods such as exon prediction. Despite numerous mappings of symbolic DNA sequences to numerical series, the existing mapping methods do not include the genetic coding features of DNA sequences. We present a novel numerical representation of DNA sequences using genetic codon context (GCC) in which the numerical values are optimized by simulation annealing to maximize the 3-periodicity signal to noise ratio (SNR). The optimized GCC representation is then applied in exon and intron prediction by Short-Time Fourier Transform (STFT) approach. The results show the GCC method enhances the SNR values of exon sequences and thus increases the accuracy of predicting protein coding regions in genomes compared with the commonly used 4D binary representation. In addition, this study offers a novel way to reveal specific features of DNA sequences by optimizing numerical mappings of symbolic DNA sequences.

ChIP-nexus enables improved detection of in vivo transcription factor binding footprints.

Understanding how eukaryotic enhancers are bound and regulated by specific combinations of transcription factors is still a major challenge. To better map transcription factor binding genome-wide at nucleotide resolution in vivo, we have developed a robust ChIP-exo protocol called ChIP experiments with nucleotide resolution through exonuclease, unique barcode and single ligation (ChIP-nexus), which utilizes an efficient DNA self-circularization step during library preparation. Application of ChIP-nexus to four proteins—human TBP and Drosophila NFkB, Twist and Max— demonstrates that it outperforms existing ChIP protocols in resolution and specificity, pinpoints relevant binding sites within enhancers containing multiple binding motifs and allows the analysis of in vivo binding specificities. Notably, we show that Max frequently interacts with DNA sequences next to its motif, and that this binding pattern correlates with local DNA sequence features such as DNA shape. ChIP-nexus will be broadly applicable to studying in vivo transcription factor binding specificity and its relationship to cis-regulatory changes in humans and model organisms.

Sequence-based prediction of transcription upregulation by auxin in plants.

Auxin is one of the main regulators of growth and development in plants. Prediction of auxin response based on gene sequence is of high importance. We found the TGTCNC consensus of 111 known natural and artificially mutated auxin response elements (AuxREs) with measured auxin-caused relative increase in genes' transcription levels, so-called either "a response to auxin" or "an auxin response." This consensus was identical to the most cited AuxRE motif. Also, we found several DNA sequence features that correlate with auxin-caused increase in genes' transcription levels, namely: number of matches with TGTCNC, homology score based on nucleotide frequencies at the consensus positions, abundances of five trinucleotides and five B-helical DNA features around these known AuxREs. We combined these correlations using a four-step empirical model of auxin response based on a gene's sequence with four steps, namely: (1) search for AuxREs with no auxin; (2) stop at the found AuxRE; (3) repression of the basal transcription of the gene having this AuxRE; and (4) manifold increase of this gene's transcription in response to auxin. Independently measured increases in transcription levels in response to auxin for 70 Arabidopsis genes were found to significantly correlate with predictions of this equation (r = 0.44, p < 0.001) as well as with TATA-binding protein (TBP)'s affinity to promoters of these genes and with nucleosome packing of these promoters (both, p < 0.025). Finally, we improved our equation for prediction of a gene's transcription increase in response to auxin by taking into account TBP-binding and nucleosome packing (r = 0.53, p < 10(-6)). Fisher's F-test validated the significant impact of both TBP/promoter-affinity and promoter nucleosome on auxin response in addition to those of AuxRE, F = 4.07, p < 0.025. It means that both TATA-box and nucleosome should be taken into account to recognize transcription factor binding sites upon DNA sequences: in the case of the TATA-less nucleosome-rich promoters, recognition scores must be higher than in the case of the TATA-containing nucleosome-free promoters at the same transcription activity.

PNImodeler: web server for inferring protein-binding nucleotides from sequence data.

BackgroundInteractions between DNA and proteins are essential to many biological processes such as transcriptional regulation and DNA replication. With the increased availability of structures of protein-DNA complexes, several computational studies have been conducted to predict DNA binding sites in proteins. However, little attempt has been made to predict protein binding sites in DNA.ResultsFrom an extensive analysis of protein-DNA complexes, we identified powerful features of DNA and protein sequences which can be used in predicting protein binding sites in DNA sequences. We developed two support vector machine (SVM) models that predict protein binding nucleotides from DNA and/or protein sequences. One SVM model that used DNA sequence data alone achieved a sensitivity of 73.4%, a specificity of 64.8%, an accuracy of 68.9% and a correlation coefficient of 0.382 with a test dataset that was not used in training. Another SVM model that used both DNA and protein sequences achieved a sensitivity of 67.6%, a specificity of 74.3%, an accuracy of 71.4% and a correlation coefficient of 0.418.ConclusionsPredicting binding sites in double-stranded DNAs is a more difficult task than predicting binding sites in single-stranded molecules. Our study showed that protein binding sites in double-stranded DNA molecules can be predicted with a comparable accuracy as those in single-stranded molecules. Our study also demonstrated that using both DNA and protein sequences resulted in a better prediction performance than using DNA sequence data alone. The SVM models and datasets constructed in this study are available at http://bclab.inha.ac.kr/pnimodeler.

4D Graphical representation research of DNA sequences

Graphical representation of DNA sequences is a key component in studying biological problems. In order to gain new insights in DNA sequences, this paper combined the digitized methods of single-base, base pairs and coding in triplet bases with the times of base appearing, and then a novel 4D graphical representation method of DNA sequences was put forward. It was a one-to-one correspondence of the arbitrary DNA sequence and 4D graphical representation, that avoided causing non-unique 4D graphical representation and overlapping lines. The method could reflect the biological information features of DNA sequence more comprehensively and effectively without any losses. Based on the 4D graphical representation, we used the geometric center of 4D graphical representation as eigenvalue of DNA sequences analyses, which kept the original features of the data, and then established the Euclidean distances and included angles between vectors' terminal point for similarity analyses of the first extron of the beta-globulin gene among 11 species. Finally, we established the graph of systematic hierarchical cluster analysis of 11 species to observe more easily the relationship between species. A positive outcome was reached, and the results were in accord with biological taxonomy, which also supported the rationality and effectiveness of the novel 4D graphical representation.

How evolution of genomes is reflected in exact DNA sequence match statistics.

Genome evolution is shaped by a multitude of mutational processes, including point mutations, insertions, and deletions of DNA sequences, as well as segmental duplications. These mutational processes can leave distinctive qualitative marks in the statistical features of genomic DNA sequences. One such feature is the match length distribution (MLD) of exactly matching sequence segments within an individual genome or between the genomes of related species. These have been observed to exhibit characteristic power law decays in many species. Here, we show that simple dynamical models consisting solely of duplication and mutation processes can already explain the characteristic features of MLDs observed in genomic sequences. Surprisingly, we find that these features are largely insensitive to details of the underlying mutational processes and do not necessarily rely on the action of natural selection. Our results demonstrate how analyzing statistical features of DNA sequences can help us reveal and quantify the different mutational processes that underlie genome evolution.

Genome-wide characterization and prediction of Arabidopsis thaliana replication origins

Identification of replication origins is crucial for the faithful duplication of genomic DNA. The frequencies of single nucleotides and dinucleotides, GC/AT bias and GC/AT profile in the vicinity of Arabidopsis thaliana replication origins were analyzed in the present work. The guanine content or cytosine content is higher in origin of replication (Ori) than in non-Ori. The SS (S=G or C) dinucleotides are favoured in Ori whereas WW (W=A or T) dinucleotides are favoured in non-Ori. GC/AT bias and GC/AT profile in Ori are significantly different from that in non-Ori. Furthermore, by inputting DNA sequence features into support vector machine, we distinguished between the Ori and non-Ori regions in A. thaliana. The total prediction accuracy is about 69.5% as evaluated by the 10-fold cross-validation. This result suggested that apart from DNA sequence, deciphering the selection of replication origin must integrate many other factors including nucleosome positioning, DNA methylation, histone modification, etc. In addition, by comparing predictive performance we found that the predictive accuracy of SVM using sequence features on the context of WS language is significantly better than that of RY language. Furthermore, the same conclusion was also obtained in S. cerevisiae and D. melanogaster.

Coregulation of Transcription Factor Binding and Nucleosome Occupancy through DNA Features of Mammalian Enhancers

Transcription factors (TFs) preferentially bind sites contained in regions of computationally predicted high nucleosomal occupancy, suggesting that nucleosomes are gatekeepers of TF binding sites. However, because of their complexity mammalian genomes contain millions of randomly occurring, unbound TF consensus binding sites. We hypothesized that the information controlling nucleosome assembly may coincide with the information that enables TFs to bind cis-regulatory elements while ignoring randomly occurring sites. Hence, nucleosomes would selectively mask genomic sites that can be contacted by TFs and thus be potentially functional. The hematopoietic pioneer TF Pu.1 maintained nucleosome depletion at macrophage-specific enhancers that displayed a broad range of nucleosome occupancy in other cell types and in reconstituted chromatin. We identified a minimal set of DNA sequence and shape features that accurately predicted both Pu.1 binding and nucleosome occupancy genome-wide. These data reveal a basic organizational principle of mammalian cis-regulatory elements whereby TF recruitment and nucleosome deposition are controlled by overlapping DNA sequence features.