Transcription factors (TFs) influence cell fate by interpreting the regulatory DNA within a genome. TFs recognize DNA in a specific manner; the mechanisms underlying this specificity have been identified for many TFs based on 3D structures of protein-DNA complexes. More recently, structural views have been complemented with data from high-throughput in vitro and in vivo explorations of the DNA-binding preferences of many TFs. Together, these approaches have greatly expanded our understanding of TF-DNA interactions. However, the mechanisms by which TFs select in vivo binding sites and alter gene expression remain unclear. Recent work has highlighted the many variables that influence TF-DNA binding, while demonstrating that a biophysical understanding of these many factors will be central to understanding TF function.

DNA sequence is a major determinant of the binding specificity of transcription factors (TFs) for their genomic targets. However, eukaryotic cells often express, at the same time, TFs with highly similar DNA binding motifs but distinct in vivo targets. Currently, it is not well understood how TFs with seemingly identical DNA motifs achieve unique specificities in vivo. Here, we used custom protein-binding microarrays to analyze TF specificity for putative binding sites in their genomic sequence context. Using yeast TFs Cbf1 and Tye7 as our case studies, we found that binding sites of these bHLH TFs (i.e., E-boxes) are bound differently in vitro and in vivo, depending on their genomic context. Computational analyses suggest that nucleotides outside E-box binding sites contribute to specificity by influencing the three-dimensional structure of DNA binding sites. Thus, the local shape of target sites might play a widespread role in achieving regulatory specificity within TF families.

We present a method and web server for predicting DNA structural features in a high-throughput (HT) manner for massive sequence data. This approach provides the framework for the integration of DNA sequence and shape analyses in genome-wide studies. The HT methodology uses a sliding-window approach to mine DNA structural information obtained from Monte Carlo simulations. It requires only nucleotide sequence as input and instantly predicts multiple structural features of DNA (minor groove width, roll, propeller twist and helix twist). The results of rigorous validations of the HT predictions based on DNA structures solved by X-ray crystallography and NMR spectroscopy, hydroxyl radical cleavage data, statistical analysis and cross-validation, and molecular dynamics simulations provide strong confidence in this approach. The DNAshape web server is freely available at http://rohslab.cmb.usc.edu/DNAshape/.

DNA binding specificities of transcription factors (TFs) are a key component of gene regulatory processes. Underlying mechanisms that explain the highly specific binding of TFs to their genomic target sites are poorly understood. A better understanding of TF−DNA binding requires the ability to quantitatively model TF binding to accessible DNA as its basic step, before additional in vivo components can be considered. Traditionally, these models were built based on nucleotide sequence. Here, we integrated 3D DNA shape information derived with a high-throughput approach into the modeling of TF binding specificities. Using support vector regression, we trained quantitative models of TF binding specificity based on protein binding microarray (PBM) data for 68 mammalian TFs. The evaluation of our models included cross-validation on specific PBM array designs, testing across different PBM array designs, and using PBM-trained models to predict relative binding affinities derived from in vitro selection combined with deep sequencing (SELEX-seq). Our results showed that shape-augmented models compared favorably to sequence-based models. Although both k-mer and DNA shape features can encode interdependencies between nucleotide positions of the binding site, using DNA shape features reduced the dimensionality of the feature space. In addition, analyzing the feature weights of DNA shape-augmented models uncovered TF family-specific structural readout mechanisms that were not revealed by the DNA sequence. As such, this work combines knowledge from structural biology and genomics, and suggests a new path toward understanding TF binding and genome function. Significance Genomes provide an abundance of putative binding sites for each transcription factor (TF). However, only small subsets of these potential targets are functional. TFs of the same protein family bind to target sites that are very similar but not identical. This distinction allows closely related TFs to regulate different genes and thus execute distinct functions. Because the nucleotide sequence of the core motif is often not sufficient for identifying a genomic target, we refined the description of TF binding sites by introducing a combination of DNA sequence and shape features, which consistently improved the modeling of in vitro TF−DNA binding specificities. Although additional factors affect TF binding in vivo, shape-augmented models reveal binding specificity mechanisms that are not apparent from sequence alone.

Transcription factors regulate gene expression, but how these proteins recognize and specifically bind to their DNA targets is still debated. Machine learning models are effective means to reveal interaction mechanisms. Here we studied the ability of a quantum machine learning approach to classify and rank binding affinities. Using simplified data sets of a small number of DNA sequences derived from actual binding affinity experiments, we trained a commercially available quantum annealer to classify and rank transcription factor binding. The results were compared to state-of-the-art classical approaches for the same simplified data sets, including simulated annealing, simulated quantum annealing, multiple linear regression, LASSO, and extreme gradient boosting. Despite technological limitations, we find a slight advantage in classification performance and nearly equal ranking performance using the quantum annealer for these fairly small training data sets. Thus, we propose that quantum annealing might be an effective method to implement machine learning for certain computational biology problems.