A framework for mutational signature analysis based on DNA shape parameters

The mutation risk of a DNA locus depends on its oligonucleotide context. In turn, mutability of oligonucleotides varies across individuals, due to exposure to mutagenic agents or due to variable efficiency and/or accuracy of DNA repair. Such variability is captured by mutational signatures, a mathematical construct obtained by a deconvolution of mutation frequency spectra across individuals. There is a need to enhance methods for inferring mutational signatures to make better use of sparse mutation data (e.g., resulting from exome sequencing of cancers), to facilitate insight into underlying biological mechanisms, and to provide more accurate mutation rate baselines for inferring positive and negative selection. We propose a conceptualization of mutational signatures that represents oligonucleotides via descriptors of DNA conformation: base pair, base pair step, and minor groove width parameters. We demonstrate how such DNA structural parameters can accurately predict mutation occurrence due to DNA repair failures or due to exposure to diverse mutagens such as radiation, chemical exposure, and the APOBEC cytosine deaminase enzymes. Furthermore, the mutation frequency of DNA oligomers classed by structural features can accurately capture systematic variability in mutagenesis of >1,000 tumors originating from diverse human tissues. A nonnegative matrix factorization was applied to mutation spectra stratified by DNA structural features, thereby extracting novel mutational signatures. Moreover, many of the known trinucleotide signatures were associated with an additional spectrum in the DNA structural descriptor space, which may aid interpretation and provide mechanistic insight. Overall, we suggest that the power of DNA sequence motif-based mutational signature analysis can be enhanced by drawing on DNA shape features.

Local DNA shape is a general principle of transcription factor binding specificity in Arabidopsis thaliana

Understanding gene expression will require understanding where regulatory factors bind genomic DNA. The frequently used sequence-based motifs of protein-DNA binding are not predictive, since a genome contains many more binding sites than are actually bound and transcription factors of the same family share similar DNA-binding motifs. Traditionally, these motifs only depict sequence but neglect DNA shape. Since shape may contribute non-linearly and combinational to binding, machine learning approaches ought to be able to better predict transcription factor binding. Here we show that a random forest machine learning approach, which incorporates the 3D-shape of DNA, enhances binding prediction for all 216 tested Arabidopsis thaliana transcription factors and improves the resolution of differential binding by transcription factor family members which share the same binding motif. We observed that DNA shape features were individually weighted for each transcription factor, even if they shared the same binding sequence.

Comprehensive understanding of Tn5 insertion preference improves transcription regulatory element identification

Tn5 transposase, which can efficiently tagment the genome, has been widely adopted as a molecular tool in next-generation sequencing, from short-read sequencing to more complex methods such as assay for transposase-accessible chromatin using sequencing (ATAC-seq). Here, we systematically map Tn5 insertion characteristics across several model organisms, finding critical parameters that affect its insertion. On naked genomic DNA, we found that Tn5 insertion is not uniformly distributed or random. To uncover drivers of these biases, we used a machine learning framework, which revealed that DNA shape cooperatively works with DNA motif to affect Tn5 insertion preference. These intrinsic insertion preferences can be modeled using nucleotide dependence information from DNA sequences, and we developed a computational pipeline to correct for these biases in ATAC-seq data. Using our pipeline, we show that bias correction improves the overall performance of ATAC-seq peak detection, recovering many potential false-negative peaks. Furthermore, we found that these peaks are bound by transcription factors, underscoring the biological relevance of capturing this additional information. These findings highlight the benefits of an improved understanding and precise correction of Tn5 insertion preference.

Examining the Effects of Netropsin on the Curvature of DNA A-Tracts Using Electrophoresis

A-tracts are sequences of repeated adenine bases that, under the proper conditions, are capable of mediating DNA curvature. A-tracts occur naturally in the regulatory regions of many organisms, yet their biological functions are not fully understood. Orienting multiple A-tracts together constructively or destructively in a phase has the potential to create different shapes in the DNA helix axis. One means of detecting these molecular shape differences is from altered DNA mobilities measured using electrophoresis. The small molecule netropsin binds the minor groove of DNA, particularly at AT-rich sequences including A-tracts. Here, we systematically test the hypothesis that netropsin binding eliminates the curvature of A-tracts by measuring the electrophoretic mobilities of seven 98-base pair DNA samples containing different numbers and arrangements of centrally located A-tracts under varying conditions with netropsin. We find that netropsin binding eliminates the mobility difference between the DNA fragments with different A-tract arrangements in a concentration-dependent manner. This work provides evidence for the straightening of A-tracts upon netropsin binding and illustrates an artificial approach to re-sculpt DNA shape.

Construction of transcript regulation mechanism prediction models based on binding motif environment of transcription factor AoXlnR in Aspergillus oryzae

Recent study revealed that there are thousands of genes that remain unaffected by increased AoXlnR expression, despite the presence of one or more AoXlnR-binding motifs in their promoter region. Given this knowledge, we designed this study to construct several predictive models for determining whether a gene can exhibit a differential response to changes in AoXlnR expression. These models were constructed using 3D DNA shape information determined using the sequence around the AoXlnR binding motifs with classification as functional or nonfunctional. These models were created using a support vector machine followed by the evaluations designed to determine whether these DNA shape-based models can correctly classify functional motifs in terms of area under the receiver operating characteristic curve. The results showed that the differential expression levels of genes located downstream of the AoXlnR motif are closely related to specific DNA shape information around the binding motifs. Furthermore, we found that the parameters contributing to differential expressions differed depending on the number of motifs in the promoter region by comparing the prediction models using regions with only one binding DNA motif and those with multiple binding DNA motifs.

MeCP2 is a microsatellite binding protein that protects CA repeats from nucleosome invasion

The Rett syndrome protein MeCP2 was described as a methyl-CpG-binding protein, but its exact function remains unknown. Here we show that mouse MeCP2 is a microsatellite binding protein that specifically recognizes hydroxymethylated CA repeats. Depletion of MeCP2 alters chromatin organization of CA repeats and lamina-associated domains and results in nucleosome accumulation on CA repeats and genome-wide transcriptional dysregulation. The structure of MeCP2 in complex with a hydroxymethylated CA repeat reveals a characteristic DNA shape, with considerably modified geometry at the 5-hydroxymethylcytosine, which is recognized specifically by Arg133, a key residue whose mutation causes Rett syndrome. Our work identifies MeCP2 as a microsatellite DNA binding protein that targets the 5hmC-modified CA-rich strand and maintains genome regions nucleosome-free, suggesting a role for MeCP2 dysfunction in Rett syndrome.

Genome information processing by the INO80 chromatin remodeler positions nucleosomes

The fundamental molecular determinants by which ATP-dependent chromatin remodelers organize nucleosomes across eukaryotic genomes remain largely elusive. Here, chromatin reconstitutions on physiological, whole-genome templates reveal how remodelers read and translate genomic information into nucleosome positions. Using the yeast genome and the multi-subunit INO80 remodeler as a paradigm, we identify DNA shape/mechanics encoded signature motifs as sufficient for nucleosome positioning and distinct from known DNA sequence preferences of histones. INO80 processes such information through an allosteric interplay between its core- and Arp8-modules that probes mechanical properties of nucleosomal and linker DNA. At promoters, INO80 integrates this readout of DNA shape/mechanics with a readout of co-evolved sequence motifs via interaction with general regulatory factors bound to these motifs. Our findings establish a molecular mechanism for robust and yet adjustable +1 nucleosome positioning and, more generally, remodelers as information processing hubs that enable active organization and allosteric regulation of the first level of chromatin.