DNA Sequence Design Research Articles

Multi-Mode Data Organization and File Retrieval Based on a Primer Library in Large-Scale Digital DNA Storage

At present, the polymerase chain reaction (PCR) amplification-based file retrieval method is the most commonly used and effective means of DNA file retrieval. The number of orthogonal primers limits the number of files that can be accurately accessed, which in turn affects the density in a single oligo pool of digital DNA storage. In this paper, a multi-mode DNA sequence design method based on PCR file retrieval in a single oligonucleotide pool is proposed for high-capacity DNA data storage. Firstly, by analyzing the maximum number of orthogonal primers at each predicted primer length, it was found that the relationship between primer length and the maximum available primer number does not increase linearly, and the maximum number of orthogonal primers is on the order of 104. Next, this paper analyzes the maximum address space capacity of DNA sequences with different types of primer binding sites for file mapping. In the case where the capacity of the primer library is R (where R is even), the number of address spaces that can be mapped by the single-primer DNA sequence design scheme proposed in this paper is four times that of the previous one, and the two-level primer DNA sequence design scheme can reach [R2∙(R2-1)]2 times. Finally, a multi-mode DNA sequence generation method is designed based on the number of files to be stored in the oligonucleotide pool, in order to meet the requirements of the random retrieval of target files in an oligonucleotide pool with large-scale file numbers. The performance of the primers generated by the orthogonal primer library generator proposed in this paper is verified, and the average Gibbs free energy of the most stable heterodimer formed between the orthogonal primers produced is –1 kcal∙(mol∙L−1)−1 (1 kcal = 4.184 kJ). At the same time, by selectively PCR-amplifying the DNA sequences of the two-level primer binding sites for random access, the target sequence can be accurately read with a minimum of 103 reads, when the primer binding site sequences at different positions are mutually different. This paper provides a pipeline for orthogonal primer library generation and multi-mode mapping schemes between files and primers, which can help achieve precise random access to files in large-scale DNA oligo pools.

Modeling 0.6 million genes for the rational design of functional cis-regulatory variants and de novo design of cis-regulatory sequences

Rational design of plant cis-regulatory DNA sequences without expert intervention or prior domain knowledge is still a daunting task. Here, we developed PhytoExpr, a deep learning framework capable of predicting both mRNA abundance and plant species using the proximal regulatory sequence as the sole input. PhytoExpr was trained over 17 species representative of major clades of the plant kingdom to enhance its generalizability. Via input perturbation, quantitative functional annotation of the input sequence was achieved at single-nucleotide resolution, revealing an abundance of predicted high-impact nucleotides in conserved noncoding sequences and transcription factor binding sites. Evaluation of maize HapMap3 single-nucleotide polymorphisms (SNPs) by PhytoExpr demonstrates an enrichment of predicted high-impact SNPs in cis-eQTL. Additionally, we provided two algorithms that harnessed the power of PhytoExpr in designing functional cis-regulatory variants, and de novo creation of species-specific cis-regulatory sequences through in silico evolution of random DNA sequences. Our model represents a general and robust approach for functional variant discovery in population genetics and rational design of regulatory sequences for genome editing and synthetic biology.

Transfer learning for cross-context prediction of protein expression from 5'UTR sequence.

Model-guided DNA sequence design can accelerate the reprogramming of living cells. It allows us to engineer more complex biological systems by removing the need to physically assemble and test each potential design. While mechanistic models of gene expression have seen some success in supporting this goal, data-centric, deep learning-based approaches often provide more accurate predictions. This accuracy, however, comes at a cost - a lack of generalization across genetic and experimental contexts that has limited their wider use outside the context in which they were trained. Here, we address this issue by demonstrating how a simple transfer learning procedure can effectively tune a pre-trained deep learning model to predict protein translation rate from 5' untranslated region (5'UTR) sequence for diverse contexts in Escherichia coli using a small number of new measurements. This allows for important model features learnt from expensive massively parallel reporter assays to be easily transferred to new settings. By releasing our trained deep learning model and complementary calibration procedure, this study acts as a starting point for continually refined model-based sequence design that builds on previous knowledge and future experimental efforts.

In silico design of DNA sequences for in vivo nucleosome positioning.

The computational design of synthetic DNA sequences with designer in vivo properties is gaining traction in the field of synthetic genomics. We propose here a computational method which combines a kinetic Monte Carlo framework with a deep mutational screening based on deep learning predictions. We apply our method to build regular nucleosome arrays with tailored nucleosomal repeat lengths (NRL) in yeast. Our design was validated in vivo by successfully engineering and integrating thousands of kilobases long tandem arrays of computationally optimized sequences which could accommodate NRLs much larger than the yeast natural NRL (namely 197 and 237bp, compared to the natural NRL of ∼165bp). RNA-seq results show that transcription of the arrays can occur but is not driven by the NRL. The computational method proposed here delineates the key sequence rules for nucleosome positioning in yeast and should be easily applicable to other sequence properties and other genomes.

H-ACO with Consecutive Bases Pairing Constraint for Designing DNA Sequences.

DNA computing is a novel computing method that does not rely on traditional computers. The design of DNA sequences is a crucial step in DNA computing, and the quality of the sequence design directly affects the results of DNA computing. In this paper, a new constraint called the consecutive base pairing constraint is proposed to limit specific base pairings in DNA sequence design. Additionally, to improve the efficiency and capability of DNA sequence design, the Hierarchy-ant colony (H-ACO) algorithm is introduced, which combines the features of multiple algorithms and optimizes discrete numerical calculations. Experimental results show that the H-ACO algorithm performs well in DNA sequence design. Finally, this paper compares a series of constraint values and NUPACK simulation data with previous design results, and the DNA sequence set designed in this paper has more advantages.

DNA Sequences Under Multiple Guanine-Cytosine (GC) Base Pairs Constraint.

DNA computing is a new computing method that has high efficiency in solving large-scale nonlinear and Non-deterministic Polynomial complete problems. The design of DNA sequences is an important step in DNA computing, and the quality of the DNA sequences directly affects the accuracy of DNA computing results. Efficiently designing high-quality DNA sequences is currently a significant challenge. In order to improve the efficiency of DNA sequence design, a sparrow evolutionary search algorithm (SESA) is proposed by us. It inherits the fast convergence of the sparrow search algorithm and avoids the situation that the sparrow search algorithm is prone to fall into a local optimum, which greatly improves the search performance of the algorithm on discrete numerical problems. In order to improve the quality of DNA sequence, a new constraint, multiple GC constraint, has been proposed in this paper. Simulated experiments in NUPACK show that this constraint can greatly improve the quality of the DNA sequences designed by us. Compared with previous results, our DNA sequences have better stability.

Enhanced DNA sequence design with learning PSO

Enhanced DNA sequence design with learning PSO

Programming crystallization kinetics of self-assembled DNA crystals with 5-methylcytosine modification

Self-assembled DNA crystals offer a precise chemical platform at the ångström-scale for DNA nanotechnology, holding enormous potential in material separation, catalysis, and DNA data storage. However, accurately controlling the crystallization kinetics of such DNA crystals remains challenging. Herein, we found that atomic-level 5-methylcytosine (5mC) modification can regulate the crystallization kinetics of DNA crystal by tuning the hybridization rates of DNA motifs. We discovered that by manipulating the axial and combination of 5mC modification on the sticky ends of DNA tensegrity triangle motifs, we can obtain a series of DNA crystals with controllable morphological features. Through DNA-PAINT and FRET-labeled DNA strand displacement experiments, we elucidate that atomic-level 5mC modification enhances the affinity constant of DNA hybridization at both the single-molecule and macroscopic scales. This enhancement can be harnessed for kinetic-driven control of the preferential growth direction of DNA crystals. The 5mC modification strategy can overcome the limitations of DNA sequence design imposed by limited nucleobase numbers in various DNA hybridization reactions. This strategy provides a new avenue for the manipulation of DNA crystal structure, valuable for the advancement of DNA and biomacromolecular crystallography.

Prediction of Mycotoxin Response of DNA-Wrapped Nanotube Sensor with Machine Learning

DNA-wrapped single-walled carbon nanotubes (SWCNTs) have demonstrated great versatility in their use as optical sensors. SWCNTs emit a near-infrared fluorescence that is responsive to even the slightest changes in the nanotube environment, enabling sensors that can respond to single-molecule fluctuations within the vicinity of their surfaces. The fluorescence response and surface interactions of these sensors are determined by the DNA wrapping sequence. However, the lack of information on the relationship between the DNA sequence and its effect on the SWCNT fluorescence remains a bottleneck for designing sequences that are specific to analytes of interest. We have recently demonstrated the use of directed volution to control the fluorescence response of SWCNTs through DNA design. Iterative cycles of DNA mutation, screening, and selection allowed us to evolve sequences that yield DNA-wrapped SWCNT sensors with a desired fluorescence response to mycotoxins.In this work, we apply the screening results of the DNA libraries used in this approach to train machine learning (ML) algorithms. Artificial neural network (ANN) and support vector machine (SVM) methods were used to predict the response of ssDNA-SWCNT sensors to a specific mycotoxin. The reliability of these models was further assessed through cross-validation. The ANN and SVM models with cross-validation were able to accurately classify the various DNA sequences as yielding either a high or low fluorescence response with an accuracy of 73 and 81%, respectively. The models were further used to predict the performance of alternative DNA sequences outside the initial training dataset, using the Hierarchy and k-means++ clustering methods to evaluate the similarity and dissimilarity of each DNA sequence. Compared to the SVM model, the ANN model showed an improved ability to predict high responses for dissimilar DNA sequences. We further applied a combinatorial approach based on SVM and ANN models to design new DNA sequences for improving sensor performance. The success of this approach was validated experimentally, demonstrating the rational design of improved sensors with 95% prediction accuracy.The application of ML algorithms to directed evolution libraries of DNA thus allows one to accurately map the performances of these sensors within a particular sequence space. The computational success of this mapping provides a framework for replacing current empirical approaches with the rational design of DNA sequences for SWCNT sensing. Keywords: Nano Biosensor, DNA-SWCNT, Mycotoxin, Machine learning, Directed evolution method.

Hold out the genome: a roadmap to solving the cis-regulatory code.

Gene expression is regulated by transcription factors that work together to read cis-regulatory DNA sequences. The 'cis-regulatory code' - how cells interpret DNA sequences to determine when, where and how much genes should be expressed - has proven to be exceedingly complex. Recently, advances in the scale and resolution of functional genomics assays and machine learning have enabled substantial progress towards deciphering this code. However, the cis-regulatory code will probably never be solved if models are trained only on genomic sequences; regions of homology can easily lead to overestimation of predictive performance, and our genome is too short and has insufficient sequence diversity to learn all relevant parameters. Fortunately, randomly synthesized DNA sequences enable testing a far larger sequence space than exists in our genomes, and designed DNA sequences enable targeted queries to maximally improve the models. As the same biochemical principles are used to interpret DNA regardless of its source, models trained on these synthetic data can predict genomic activity, often better than genome-trained models. Here we provide an outlook on the field, and propose a roadmap towards solving the cis-regulatory code by a combination of machine learning and massively parallel assays using synthetic DNA.

Towards Chinese text and DNA shift encoding scheme based on biomass plasmid storage.

DNA, as the storage medium in organisms, can address the shortcomings of existing electromagnetic storage media, such as low information density, high maintenance power consumption, and short storage time. Current research on DNA storage mainly focuses on designing corresponding encoders to convert binary data into DNA base data that meets biological constraints. We have created a new Chinese character code table that enables exceptionally high information storage density for storing Chinese characters (compared to traditional UTF-8 encoding). To meet biological constraints, we have devised a DNA shift coding scheme with low algorithmic complexity, which can encode any strand of DNA even has excessively long homopolymer. The designed DNA sequence will be stored in a double-stranded plasmid of 744bp, ensuring high reliability during storage. Additionally, the plasmid's resistance to environmental interference ensuring long-term stable information storage. Moreover, it can be replicated at a lower cost.

DNA Programmable Peroxidase Activity of Oxidized Polypyrrole Quantum Dots

AbstractSignificant progress has recently been made in the application of oxidized graphene (GO) quantum dots as enzyme mimics in various biomedical fields due to their bio‐compatibility and excellent solubility in physiological media. However, their catalytic performance and controllability are barely satisfactory. Here, this study constructs oxygen‐functionalized polypyrrole quantum dots (o‐ppy QDs) with excellent peroxidase activity in a mild condition. Compared with oxidized graphene QDs, o‐ppy QDs exhibit superior catalytic efficiency (120 times higher than HRP). More importantly, it is found that guanine (G) and adenine (A) bases possess higher binding affinities to o‐ppy QDs. G base is able to significantly increase the peroxidase activity while A base decreases the activity, providing a fascinating method to precisely regulate the catalytic activity of o‐ppy QDs in a programmable manner by the design of DNA sequences. The enhancement on the peroxidase by G base regulation is attributed to the existence of carbonyl group that promotes its catalytic activity, while A base tends to block the original carbonyl group on o‐ppy QDs. Based on this feature, a colorimetric and fluorescent dual‐mode biosensor for detecting DNA methylation is developed. This study holds significant theoretical and practical implications for the development of nanozymes and precise regulation of their catalytic activity.

Prediction of Mycotoxin Response of DNA-Wrapped Nanotube Sensor with Machine Learning

DNA-wrapped single-walled carbon nanotubes (SWCNTs) have demonstrated great versatility in their use as optical sensors. SWCNTs emit a near-infrared fluorescence that is responsive to even the slightest changes in the nanotube environment, enabling sensors that can respond to single-molecule fluctuations within the vicinity of their surfaces. The fluorescence response and surface interactions of these sensors are determined by the DNA wrapping sequence. However, the lack of information on the relationship between the DNA sequence and its effect on the SWCNT fluorescence remains a bottleneck for designing sequences that are specific to analytes of interest. We have recently demonstrated the use of directed evolution to control the fluorescence response of SWCNTs through DNA design. Iterative cycles of DNA mutation, screening, and selection allowed us to evolve sequences that yield DNA-wrapped SWCNT sensors with a desired fluorescence response to mycotoxins.In this work, we apply the screening results of the DNA libraries used in this approach to train machine learning (ML) algorithms. Artificial neural network (ANN) and support vector machine (SVM) methods were used to predict the response of ssDNA-SWCNT sensors to a specific mycotoxin. The reliability of these models was further assessed through cross-validation. The ANN and SVM models with cross-validation were able to accurately classify the various DNA sequences as yielding either a high or low fluorescence response with an accuracy of 73 and 81%, respectively. The models were further used to predict the performance of alternative DNA sequences outside the initial training dataset, using the Hierarchy and k-means ++ clustering methods to evaluate the similarity and dissimilarity of each DNA sequence. Compared to the SVM model, the ANN model showed an improved ability to predict high responses for dissimilar DNA sequences. We further applied a combinatorial approach based on SVM and ANN models to design new DNA sequences for improving sensor performance. The success of this approach was validated experimentally, demonstrating the rational design of improved sensors with 95% prediction accuracy.The application of ML algorithms to directed evolution libraries of DNA thus allows one to accurately map the performances of these sensors within a particular sequence space. The computational success of this mapping provides a framework for replacing current empirical approaches with the rational design of DNA sequences for SWCNT sensing. Keywords: Nano Biosensor, DNA-SWCNT, Mycotoxin, Machine learning, Directed evolution method

Improved Bare Bones Particle Swarm Optimization for DNA Sequence Design.

DNA computing has efficient computational power, but requires high requirements on the DNA sequences used for coding, and reliable DNA sequences can effectively improve the quality of DNA encoding. And designing reliable DNA sequences is an NP problem, because it requires finding DNA sequences that satisfy multiple sets of conflicting constraints from a large solution space. To better solve the DNA sequence design problem, we propose an improved bare bones particle swarm optimization algorithm (IBPSO). The algorithm uses dynamic lensing opposition-based learning to initialize the population to improve population diversity and enhance the ability of the algorithm to jump out of local optima; An evolutionary strategy based on signal-to-noise ratio(SNR) distance is designed to balance the exploration and exploitation of the algorithm; Then an invasive weed optimization algorithm with niche crowding(NCIWO) is used to eliminate low-quality solutions and improve the search efficiency of the algorithm. In addition, we introduce the triplet-bases unpaired constraint to further improve the quality of DNA sequences. Finally, the effectiveness of the improved strategy is demonstrated by ablation experiments; and the DNA sequences designed by our algorithm are of higher quality compared with those generated by the four advanced algorithms.

TReSR: A PCR-compatible DNA sequence design method for engineering proteins containing tandem repeats.

Protein tandem repeats (TRs) are motifs comprised of near-identical contiguous sequence duplications. They are found in approximately 14% of all proteins and are implicated in diverse biological functions facilitating both structured and disordered protein-protein and protein-DNA interactions. These functionalities make protein TR domains an attractive component for the modular design of protein constructs. However, the repetitive nature of DNA sequences encoding TR motifs complicates their synthesis and mutagenesis by traditional molecular biology workflows commonly employed by protein engineers and synthetic biologists. To address this challenge, we developed a computational protocol to significantly reduce the complementarity of DNA sequences encoding TRs called TReSR (for Tandem Repeat DNA Sequence Redesign). The utility of TReSR was demonstrated by constructing a novel constitutive repressor synthesized by duplicating the LacI DNA binding domain into a single-chain TR construct by assembly PCR. Repressor function was evaluated by expression of a fluorescent reporter delivered on a single plasmid encoding a three-component genetic circuit. The successful application of TReSR to construct a novel TR-containing repressor with a DNA sequence that is amenable to PCR-based construction and manipulation will enable the incorporation of a wide range of TR-containing proteins for protein engineering and synthetic biology applications.

Programming Nucleation and Growth in Colloidal Crystals Using DNA.

Colloidal crystal engineering with DNA has advanced beyond controlling the lattice symmetry and parameters of ordered crystals to now tuning crystal habit and size. However, the predominately used slow-cooling procedure that enables faceted crystal habits also limits control over crystal size and uniformity because nucleation and growth cannot be separated. Here, we explore how DNA sequence design can be used to deliberately separate nucleation and growth in a given crystallization process. Specifically, two batches of complementary particles are created with one batch exhibiting perfectly complementary base pairs while the other has a strategically introduced mismatch. This design enables the weaker binding "growth" particles to participate in heterogeneous growth on the nucleates formed from the stronger binding "seed" particles, effectively eliminating secondary nucleation pathways. By eliminating secondary nucleation events, this approach improves crystal uniformity, as measured by polydispersity (from PDI = 0.201 to 0.091). By using this approach with two different particle cores (gold and silver), we show how core-shell colloidal crystals can be synthesized in a one-pot fashion. This work shows how tuning DNA interaction strength can profoundly impact crystal size, uniformity, and structure, parameters central to using such materials as device components.

Improved Multi-Strategy Matrix Particle Swarm Optimization for DNA Sequence Design

The efficiency of DNA computation is closely related to the design of DNA coding sequences. For the purpose of obtaining superior DNA coding sequences, it is necessary to choose suitable DNA constraints to prevent potential conflicting interactions in different DNA sequences and to ensure the reliability of DNA sequences. An improved matrix particle swarm optimization algorithm, referred to as IMPSO, is proposed in this paper to optimize DNA sequence design. In addition, this paper incorporates centroid opposition-based learning to fully preserve population diversity and develops and adapts a dynamic update on the basis of signal-to-noise ratio distance to search for high-quality solutions in a sufficiently intelligent manner. The results show that the proposal of this paper achieves satisfactory results and can obtain higher computational efficiency.

Beyond nature's base pairs: machine learning-enabled design of DNA-stabilized silver nanoclusters.

Sequence-encoded biomolecules such as DNA and peptides are powerful programmable building blocks for nanomaterials. This paradigm is enabled by decades of prior research into how nucleic acid and amino acid sequences dictate biomolecular interactions. The properties of biomolecular materials can be significantly expanded with non-natural interactions, including metal ion coordination of nucleic acids and amino acids. However, these approaches present design challenges because it is often not well-understood how biomolecular sequence dictates such non-natural interactions. This Feature Article presents a case study in overcoming challenges in biomolecular materials with emerging approaches in data mining and machine learning for chemical design. We review progress in this area for a specific class of DNA-templated metal nanomaterials with complex sequence-to-property relationships: DNA-stabilized silver nanoclusters (AgN-DNAs) with bright, sequence-tuned fluorescence colors and promise for biophotonics applications. A brief overview of machine learning concepts is presented, and high-throughput experimental synthesis and characterization of AgN-DNAs are discussed. Then, recent progress in machine learning-guided design of DNA sequences that select for specific AgN-DNA fluorescence properties is reviewed. We conclude with emerging opportunities in machine learning-guided design and discovery of AgN-DNAs and other sequence-encoded biomolecular nanomaterials.

Diversifying Design of Nucleic Acid Aptamers Using Unsupervised Machine Learning.

Inverse design of short single-stranded RNA and DNA sequences (aptamers) is the task of finding sequences that satisfy a set of desired criteria. Relevant criteria may be, for example, the presence of specific folding motifs, binding to molecular ligands, sensing properties, and so on. Most practical approaches to aptamer design identify a small set of promising candidate sequences using high-throughput experiments (e.g., SELEX) and then optimize performance by introducing only minor modifications to the empirically found candidates. Sequences that possess the desired properties but differ drastically in chemical composition will add diversity to the search space and facilitate the discovery of useful nucleic acid aptamers. Systematic diversification protocols are needed. Here we propose to use an unsupervised machine learning model known as the Potts model to discover new, useful sequences with controllable sequence diversity. We start by training a Potts model using the maximum entropy principle on a small set of empirically identified sequences unified by a common feature. To generate new candidate sequences with a controllable degree of diversity, we take advantage of the model's spectral feature: an "energy" bandgap separating sequences that are similar to the training set from those that are distinct. By controlling the Potts energy range that is sampled, we generate sequences that are distinct from the training set yet still likely to have the encoded features. To demonstrate performance, we apply our approach to design diverse pools of sequences with specified secondary structure motifs in 30-mer RNA and DNA aptamers.

Electrochemiluminescence Cascade Amplification Platform for Detection of Dual-microRNA and Operation of Concatenated Logic Circuit.

The detection of multiple biomarkers is of great significance to the accurate diagnosis of diseases. Herein, in this work, we constructed an electrochemiluminescence (ECL) cascade amplification platform for dual acute myocardial infarction (AMI)-related microRNA detection. The Zn2+-dependent DNAzyme digestion reaction initiated by miR-133a and the duplex-specific nuclease (DSN) cleavage circuit initiated by miR-499 were carried out independently to form a fuel hairpin DNA and active initiator strand, respectively, to trigger a hybridization chain reaction, which constituted a two-input-regulated "AND" logic circuit based on single ECL signal output. The use of single signal probe (Ru(bpy)32+) avoided the time-consuming and costly process of multiple signal molecule labeling or modification. The independent operation of the DNAzyme digestion reaction and DSN-assisted target recycling improved the detection efficiency of the system. In addition, the detection of each miRNA had undergone a cascade amplification process, which improved the detection sensitivity for each target. Furthermore, benefitting from the strong complexation of EDTA with Zn2+ and the flexible design of DNA sequences, the two-input "AND" logic gate was extended to a four-input "INHIBIT-AND-INHIBIT" concatenated logic circuit, which broadens the application of the ECL method in logic gates. We anticipate that this cascading amplification strategy can be widely applied in accurate diagnosis of AMI and the construction of ECL-based logic devices.