Short Sequences Research Articles

Cas9 interaction with the tracrRNA nexus modulates the repression of type II-A CRISPR-cas genes.

Immune responses need to be regulated to prevent autoimmunity. CRISPR-Cas systems provide adaptive immunity in prokaryotes through the acquisition of short DNA sequences from invading viruses (bacteriophages), known as spacers. Spacers are inserted into the CRISPR locus and serve as templates for the transcription of guides used by RNA-guided nucleases to recognize complementary nucleic acids of the invaders and start the CRISPR immune response. In type II-A CRISPR systems, Cas9 uses the guide RNA to cleave target DNA sequences in the genome of infecting phages, and the tracrRNA to bind the promoter of cas genes and repress their transcription. We previously isolated a Cas9 mutant carrying the I473F substitution that increased the frequency of spacer acquisition by 2-3 orders of magnitude, leading to a fitness cost due to higher levels of autoimmunity. Here, we investigated the molecular basis underlying these findings. We found that the I473F mutation decreases the association of Cas9 to tracrRNA, limiting its repressor function, leading to high levels of expression of cas genes, which in turn increase the strength of the type II-A CRISPR-Cas immune response. We obtained similar results for a related type II-A system, and therefore our findings highlight the importance of the interaction between Cas9 and its tracrRNA cofactor in tuning the immune response to balanced levels that enable phage defense but avoid autoimmunity.

Applications of CRISPR in Parasitology.

Clustered Regions of Interspersed Palindromic Repeat (CRISPR)-based techniques have been utilized in various research areas, including agriculture, biotechnology, and medicine. With the use of a short sequence guide RNA and CRISPR-associated (Cas) protein, this technique allows for robust, site-specific manipulation of the genome, aiding researchers in making important biomedical discoveries and scientific advancements. In this review, we explored the applications of CRISPR/Cas systems in the field of parasitology for the identification and validation of novel functional genes, diagnosis of parasitic infections, reduction of parasite virulence, and the disruption of disease transmission. We also discussed how CRISPR can be used for the development of therapeutics, vaccines, and drug discovery. Furthermore, the challenges and future perspectives of this technology are also highlighted.

GWarrange: a pre- and post- genome-wide association studies pipeline for detecting phenotype-associated genome rearrangement events.

The use of k-mers to capture genetic variation in bacterial genome-wide association studies (bGWAS) has demonstrated its effectiveness in overcoming the plasticity of bacterial genomes by providing a comprehensive array of genetic variants in a genome set that is not confined to a single reference genome. However, little attempt has been made to interpret k-mers in the context of genome rearrangements, partly due to challenges in the exhaustive and high-throughput identification of genome structure and individual rearrangement events. Here, we present GWarrange, a pre- and post-bGWAS processing methodology that leverages the unique properties of k-mers to facilitate bGWAS for genome rearrangements. Repeat sequences are common instigators of genome rearrangements through intragenomic homologous recombination, and they are commonly found at rearrangement boundaries. Using whole-genome sequences, repeat sequences are replaced by short placeholder sequences, allowing the regions flanking repeats to be incorporated into relatively short k-mers. Then, locations of flanking regions in significant k-mers are mapped back to complete genome sequences to visualise genome rearrangements. Four case studies based on two bacterial species (Bordetella pertussis and Enterococcus faecium) and a simulated genome set are presented to demonstrate the ability to identify phenotype-associated rearrangements. GWarrange is available at https://github.com/DorothyTamYiLing/GWarrange.

ProcaryaSV: structural variation detection pipeline for bacterial genomes using short-read sequencing

BackgroundStructural variations play an important role in bacterial genomes. They can mediate genome adaptation quickly in response to the external environment and thus can also play a role in antibiotic resistance. The detection of structural variations in bacteria is challenging, and the recognition of even small rearrangements can be important. Even though most detection tools are aimed at and benchmarked on eukaryotic genomes, they can also be used on prokaryotic genomes. The key features of detection are the ability to detect small rearrangements and support haploid genomes. Because of the limiting performance of a single detection tool, combining the detection abilities of multiple tools can lead to more robust results. There are already available workflows for structural variation detection for long-reads technologies and for the detection of single-nucleotide variation and indels, both aimed at bacteria. Yet we are unaware of structural variations detection workflows for the short-reads sequencing platform. Motivated by this gap we created our workflow. Further, we were interested in increasing the detection performance and providing more robust results.ResultsWe developed an open-source bioinformatics pipeline, ProcaryaSV, for the detection of structural variations in bacterial isolates from paired-end short sequencing reads. Multiple tools, starting with quality control and trimming of sequencing data, alignment to the reference genome, and multiple structural variation detection tools, are integrated. All the partial results are then processed and merged with an in-house merging algorithm. Compared with a single detection approach, ProcaryaSV has improved detection performance and is a reproducible easy-to-use tool.ConclusionsThe ProcaryaSV pipeline provides an integrative approach to structural variation detection from paired-end next-generation sequencing of bacterial samples. It can be easily installed and used on Linux machines. It is publicly available on GitHub at https://github.com/robinjugas/ProcaryaSV.

Spinocerebellar ataxia type 4 is caused by a GGC expansion in the ZFHX3 gene and is associated with prominent dysautonomia and motor neuron signs.

Spinocerebellar ataxia 4 (SCA4), characterized in 1996, features adult-onset ataxia, polyneuropathy, and linkage to chromosome 16q22.1; its underlying mutation has remained elusive. To explore the radiological and neuropathological abnormalities in the entire neuroaxis in SCA4 and search for its mutation. Three Swedish families with undiagnosed ataxia went through clinical, neurophysiological, and neuroimaging tests, including PET studies and genetic investigations. In four cases, neuropathological assessments of the neuroaxis were performed. Genetic testing included short read whole genome sequencing, short tandem repeat analysis with ExpansionHunter de novo, and long read sequencing. Novel features for SCA4 include dysautonomia, motor neuron affection, and abnormal eye movements. We found evidence of anticipation; neuroimaging demonstrated atrophy in the cerebellum, brainstem, and spinal cord. [18F]FDG-PET demonstrated brain hypometabolism and [11C]Flumazenil-PET reduced binding in several brain lobes, insula, thalamus, hypothalamus, and cerebellum. Moderate to severe loss of Purkinje cells in the cerebellum and of motor neurons in the anterior horns of the spinal cord along with pronounced degeneration of posterior tracts was also found. Intranuclear, mainly neuronal, inclusions positive for p62 and ubiquitin were sparse but widespread in the CNS. This finding prompted assessment for nucleotide expansions. A polyglycine stretch encoding GGC expansions in the last exon of the zink finger homeobox 3 gene was identified segregating with disease and not found in 1000 controls. SCA4 is a neurodegenerative disease caused by a novel GGC expansion in the coding region of ZFHX3, and its spectrum is expanded to include dysautonomia and neuromuscular manifestations.

A Short Sequence Targets Transmembrane Proteins to Primary Cilia.

Primary cilia are finger-like sensory organelles that extend from the bodies of most cell types and have a distinct lipid and protein composition from the plasma membrane. This partitioning is maintained by a diffusion barrier that restricts the entry of non-ciliary proteins, and allows the selective entry of proteins harboring a ciliary targeting sequence (CTS). However, CTSs are not stereotyped and previously reported sequences are insufficient to drive efficient ciliary localisation across diverse cell types. Here, we describe a short peptide sequence that efficiently targets transmembrane proteins to primary cilia in all tested cell types, including human neurons. We generate human-induced pluripotent stem cell (hiPSC) lines stably expressing a transmembrane construct bearing an extracellular HaloTag and intracellular fluorescent protein, which enables the bright, specific labeling of primary cilia in neurons and other cell types to facilitate studies of cilia in health and disease. We demonstrate the utility of this resource by developing an image analysis pipeline for the automated measurement of primary cilia to detect changes in their length associated with altered signaling or disease state.

The non-negligible non-specific adsorption of oligonucleotides in target-immobilized Mag-SELEX

Target-immobilized magnetic beads-based Systematic Evolution of Ligands by Exponential Enrichment (target-immobilized Mag-SELEX) has emerged as a powerful tool for aptamer selection owing to its convenience, efficiency, and versatility. However, in this study we systematically investigated non-specific adsorption in target-immobilized Mag-SELEX and found that the non-specific adsorption of the oligonucleotides to target-labeled magnetic beads was comparable to that of the screening libraries, indicating a substantial portion of captured sequences likely stem from non-specific adsorption. Longer nucleic acid sequences (80 nt and above, such as polyA80 and yeast tRNA) were found to attenuate this non-specific adsorption, with more complex higher-order structures demonstrating greater efficacy, while dNTP and short sequences such as primer sequences (20 nt), polyT(59), or polyA(59), did not possess this capability. Various evidence suggested that hydrophobic interactions and other weak interactions may be the primary underlying cause of non-specific adsorption. Additionally, surface modification of magnetic beads with polar molecule polyethylene glycol (PEG) also yielded a significant reduction in non-specific adsorption. In conclusion, our research underscores the critical importance of closely monitoring non-specific adsorption in target-immobilized Mag-SELEX.

A comparative evaluation of measures to assess randomness in human-generated sequences

Whether and how well people can behave randomly is of interest in many areas of psychological research. The ability to generate randomness is often investigated using random number generation (RNG) tasks, in which participants are asked to generate a sequence of numbers that is as random as possible. However, there is no consensus on how best to quantify the randomness of responses in human-generated sequences. Traditionally, psychologists have used measures of randomness that directly assess specific features of human behavior in RNG tasks, such as the tendency to avoid repetition or to systematically generate numbers that have not been generated in the recent choice history, a behavior known as cycling. Other disciplines have proposed measures of randomness that are based on a more rigorous mathematical foundation and are less restricted to specific features of randomness, such as algorithmic complexity. More recently, variants of these measures have been proposed to assess systematic patterns in short sequences. We report the first large-scale integrative study to compare measures of specific aspects of randomness with entropy-derived measures based on information theory and measures based on algorithmic complexity. We compare the ability of the different measures to discriminate between human-generated sequences and truly random sequences based on atmospheric noise, and provide a systematic analysis of how the usefulness of randomness measures is affected by sequence length. We conclude with recommendations that can guide the selection of appropriate measures of randomness in psychological research.

Bidirectional transfer of human cytomegalovirus strains in donor and recipient seropositive lung transplant patients

AbstractDonor and recipient human cytomegalovirus (HCMV) seropositive (D+R+) lung transplant recipients (LTRs) often harbor multiple strains of HCMV, likely due to transmitted donor (D) strains and reactivated recipient (R) strains. To date, the extent and timely occurrence of each likely source in shaping the post‐transplantation (post‐Tx) strain population is unknown. Here, we deciphered the D and R origin of the post‐Tx HCMV strain composition in blood, bronchoalveolar lavage (BAL), and CD45+ BAL cell subsets. We investigated either D and/or R formalin‐fixed paraffin‐embedded blocks or fresh D lung tissue from four D+R+ LTRs obtained before transplantation. HCMV strains were characterized by short amplicon deep sequencing. In two LTRs, we show that the transplanted lung is reseeded by R strains within the first 6 months after transplantation, likely by infiltrating CD14+ CD163+/− alveolar macrophages. In three LTRs, we demonstrate both rapid D‐strain dissemination and persistence in the transplanted lung for >1 year post‐Tx. Broad inter‐host diversity contrasts with intra‐host genotype sequence stability upon transmission, during follow‐up and across compartments. In D+R+ LTRs, HCMV strains of both, D and R origin can emerge first and dominate long‐term in subsequent episodes of infection, indicating replication of both sources despite pre‐existing immunity.

Predicting symptom severity in PSTVd-infected tomato plants using the PSTVd genome sequence.

Viroids, one of the smallest known infectious agents, induce symptoms of varying severity, ranging from latent to severe, based on the combination of viroid isolates and host plant species. Because viroids are transmissible between plant species, asymptomatic viroid-infected plants may serve as latent sources of infection for other species that could exhibit severe symptoms, occasionally leading to agricultural and economic losses. Therefore, predicting the symptoms induced by viroids in host plants without biological experiments could remarkably enhance control measures against viroid damage. Here, we developed an algorithm using unsupervised machine learning to predict the severity of disease symptoms caused by viroids (e.g., potato spindle tuber viroid; PSTVd) in host plants (e.g., tomato). This algorithm, mimicking the RNA silencing mechanism thought to be linked to viroid pathogenicity, requires only the genome sequences of the viroids and host plants. It involves three steps: alignment of synthetic short sequences of the viroids to the host plant genome, calculation of the alignment coverage, and clustering of the viroids based on coverage using UMAP and DBSCAN. Validation through inoculation experiments confirmed the effectiveness of the algorithm in predicting the severity of disease symptoms induced by viroids. As the algorithm only requires the genome sequence data, it may be applied to any viroid and plant combination. These findings underscore a correlation between viroid pathogenicity and the genome sequences of viroid isolates and host plants, potentially aiding in the prevention of viroid outbreaks and the breeding of viroid-resistant crops.

Saccharomycetaceae: delineation of fungal genera based on phylogenomic analyses, genomic relatedness indices and genomics-based synapomorphies.

A correct classification of fungi, including yeasts, is of prime importance to understand fungal biodiversity and to communicate about this diversity. Fungal genera are mainly defined based on phenotypic characteristics and the results of single or multigene-based phylogenetic analyses. However, because yeasts often have less phenotypic characters, their classification experienced a strong move towards DNA-based data, from short ribosomal sequences to multigene phylogenies and more recently to phylogenomics. Here, we explore the usefulness of various genomics-based parameters to circumscribe fungal genera more correctly taking the yeast domain as an example. Therefore, we compared the results of a phylogenomic analysis, average amino acid identity (AAI) values, the presence of conserved signature indels (CSIs), the percentage of conserved proteins (POCP) and the presence-absence patterns of orthologs (PAPO). These genome-based metrics were used to investigate their usefulness in demarcating 13 hitherto relatively well accepted genera in Saccharomycetaceae, namely Eremothecium, Grigorovia, Kazachstania, Kluyveromyces, Lachancea, Nakaseomyces, Naumovozyma, Saccharomyces, Tetrapisispora, Torulaspora, Vanderwaltozyma, Zygosaccharomyces and Zygotorulaspora. As a result, most of these genera are supported by the genomics-based metrics, but the genera Kazachstania, Nakaseomyces and Tetrapisispora were shown to be genetically highly diverse based on the above listed analyses. Considering the results obtained for the presently recognized genera, a range of 80-92 % POCP values and a range of 60-70 % AAI values might be valuable thresholds to discriminate genera in Saccharomycetaceae. Furthermore, the genus-specific genes identified in the PAPO analysis and the CSIs were found to be useful as synapomorphies to characterize and define genera in Saccharomycetaceae. Our results indicate that the combined monophyly-based phylogenomic analysis together with genomic relatedness indices and synapomorphies provide promising approaches to delineating yeast genera and likely those of filamentous fungi as well. The genera Kazachstania, Nakaseomyces and Tetrapisispora are revised and we propose eight new genera and 41 new combinations. Citation: Liu F, Hu Z-D, Yurkov A, et al. 2024. Saccharomycetaceae: delinaeation of fungal genera based on phylogenomic analyses, genomic relatedness indices and genomics-based synapomorphies. Persoonia 52: 1-21. https://doi.org/10.3767/persoonia.2024.52.01.

Floria: fast and accurate strain haplotyping in metagenomes.

Shotgun metagenomics allows for direct analysis of microbial community genetics, but scalable computational methods for the recovery of bacterial strain genomes from microbiomes remains a key challenge. We introduce Floria, a novel method designed for rapid and accurate recovery of strain haplotypes from short and long-read metagenome sequencing data, based on minimum error correction (MEC) read clustering and a strain-preserving network flow model. Floria can function as a standalone haplotyping method, outputting alleles and reads that co-occur on the same strain, as well as an end-to-end read-to-assembly pipeline (Floria-PL) for strain-level assembly. Benchmarking evaluations on synthetic metagenomes show that Floria is > 3× faster and recovers 21% more strain content than base-level assembly methods (Strainberry) while being over an order of magnitude faster when only phasing is required. Applying Floria to a set of 109 deeply sequenced nanopore metagenomes took <20 min on average per sample and identified several species that have consistent strain heterogeneity. Applying Floria's short-read haplotyping to a longitudinal gut metagenomics dataset revealed a dynamic multi-strain Anaerostipes hadrus community with frequent strain loss and emergence events over 636 days. With Floria, accurate haplotyping of metagenomic datasets takes mere minutes on standard workstations, paving the way for extensive strain-level metagenomic analyses. Floria is available at https://github.com/bluenote-1577/floria, and the Floria-PL pipeline is available at https://github.com/jsgounot/Floria_analysis_workflow along with code for reproducing the benchmarks.

Utility Analyses of AVITI Sequencing Chemistry.

DNA sequencing is a critical tool in modern biology. Over the last two decades, it has been revolutionized by the advent of massively parallel sequencing, leading to significant advances in the genome and transcriptome sequencing of various organisms. Nevertheless, challenges with accuracy, lack of competitive options and prohibitive costs associated with high throughput parallel short-read sequencing persist. Here, we conduct a comparative analysis using matched DNA and RNA short-reads assays between Element Biosciences' AVITI and Illumina's NextSeq 550 chemistries. Similar comparisons were evaluated for synthetic long-read sequencing for RNA and targeted single-cell transcripts between the AVITI and Illumina's NovaSeq 6000. For both DNA and RNA short-read applications, the study found that the AVITI produced significantly higher per sequence quality scores. For PCR-free DNA libraries, we observed an average 89.7% lower experimentally determined error rate when using the AVITI chemistry, compared to the NextSeq 550. For short-read RNA quantification, AVITI platform had an average of 32.5% lower error rate than that for NextSeq 550. With regards to synthetic long-read mRNA and targeted synthetic long read single cell mRNA sequencing, both platforms' respective chemistries performed comparably in quantification of genes and isoforms. The AVITI displayed a marginally lower error rate for long reads, with fewer chemistry-specific errors and a higher mutation detection rate. These results point to the potential of the AVITI platform as a competitive candidate in high-throughput short read sequencing analyses when juxtaposed with the Illumina NextSeq 550.

New specific primer matK and rbcL region for DNA barcode pitcher plant Nepenthes spathulata

Abstract. Utama MN, Etikawati N, Sugiyarto, Susilowati A. 2024. New specific primer matK and rbcL region for DNA barcode pitcher plant Nepenthes spathulata. Biodiversitas 25: 2515-2523. The matK and rbcL genes were proposed as the preferred plant barcoding loci by The Consortium for the Barcode of Life (CBOL). DNA barcoding efficiently identifies samples at the species level using short standard DNA sequences. The identification disclosure needs is an important basis given the confusion in systematics in Nepenthes spathulata. Therefore, this study aims to design matK and rbcL region-specific primers for DNA barcoding pitcher plants. The primary design uses Primer3Plus with a blueprint from the GenBank KDQ007081.1 for matK and MH346374.1 for rbcL. Following testing primary candidates with Oligo Analyzer and Primer-BLAST, the primers MkNs1, and RbNs4 were selected as the optimal criteria based on the values of ?G and primer specificity. Optimization of annealing temperature with PCR gradient shows that the temperature range 52-57°C produces a good band for both regions and corresponds to the matK and rbcL region with product sizes of 800-900 bp from seven samples, which could be observed from the gel electrophoresis. Sequence similarity using blast-N, the matK sequence has similarity to the N. spathulata matK gene of 99.64% and 99.63% for the rbcL gene of N. ventricosa x N. alata. This MkNs1 and RbNs4 primer can be used to discover the identity of N. spathulata and the Nepenthes genus.

Amplification editing enables efficient and precise duplication of DNA from short sequence to megabase and chromosomal scale

Duplication is a foundation of molecular evolution and a driver of genomic and complex diseases. Here, we develop a genome editing tool named Amplification Editing (AE) that enables programmable DNA duplication with precision at chromosomal scale. AE can duplicate human genomes ranging from 20bp to 100 Mb, a size comparable to human chromosomes. AE exhibits activity across various cell types, encompassing diploid, haploid, and primary cells. AE exhibited up to 73.0% efficiency for 1 Mb and 3.4% for 100 Mb duplications, respectively. Whole-genome sequencing and deep sequencing of the junctions of edited sequences confirm the precision of duplication. AE can create chromosomal microduplications within disease-relevant regions in embryonic stem cells, indicating its potential for generating cellular and animal models. AE is a precise and efficient tool for chromosomal engineering and DNA duplication, broadening the landscape of precision genome editing from an individual genetic locus to the chromosomal scale.

A deep learning model for predicting the state of energy in lithium-ion batteries based on magnetic field effects

The state of energy (SOE) is one of the most critical state indicators in battery management systems. However, its nonlinear characteristics present significant challenges in obtaining accurate SOE. Especially when applying different magnetic field strengths to perform battery charging and discharging tests, the change in battery energy becomes more complex due to the influence of the magnetization effect. In this paper, a deep learning network, combining an improved Informer and long short-term memory network (LSTM), was developed to estimate the SOE of lithium-ion batteries under different magnetic fields. First, we improve the decoder structure by adding a convolutional module using residual connections with trainable weight parameters to capture hidden states with more details.The improved decoder does not require label history information for decoding, which improves the generalization ability of the model. Finally, the output of the Informer network is a higher-dimensional hidden feature that is input into the LSTM network layer to output the SOE prediction value, which improves the original Informer network's ability to integrate sequences. Experiments with magnetic field and public datasets show the improved Informer-LSTM network achieves 0.31 % MAE, 0.42 % RMSE, and 1.79 % maximum error in SOE estimation, outperforming others in short sequence predictions.

Bioengineered self-assembled nanofibrils for high-affinity SARS-CoV-2 capture and neutralization

The recent coronavirus disease 2019 (COVID-19) pandemic caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has spurred intense research efforts to develop new materials with antiviral activity. In this study, we genetically engineered amyloid-based nanofibrils for capturing and neutralizing SARS-CoV-2. Building upon the amyloid properties of a short Sup35 yeast prion sequence, we fused it to SARS-CoV-2 receptor-binding domain (RBD) capturing proteins, LCB1 and LCB3. By tuning the reaction conditions, we achieved the spontaneous self-assembly of the Sup35-LCB1 fusion protein into a highly homogeneous and well-dispersed amyloid-like fibrillar material. These nanofibrils exhibited high affinity for the SARS-CoV-2 RBD, effectively inhibiting its interaction with the angiotensin-converting enzyme 2 (ACE2) receptor, the primary entry point for the virus into host cells. We further demonstrate that this functional nanomaterial entraps and neutralizes SARS-CoV-2 virus-like particles (VLPs), with a potency comparable to that of therapeutic antibodies. As a proof of concept, we successfully fabricated patterned surfaces that selectively capture SARS-CoV-2 RBD protein on wet environments. Collectively, these findings suggest that these protein-only nanofibrils hold promise as disinfecting coatings endowed with selective SARS-CoV-2 neutralizing properties to combat viral spread or in the development of sensitive viral sampling and diagnostic tools.

Pseudo-pac site sequences used by phage P22 in generalized transduction of Salmonella.

Salmonella enterica Serovar Typhimurium (Salmonella) and its bacteriophage P22 are a model system for the study of horizontal gene transfer by generalized transduction. Typically, the P22 DNA packaging machinery initiates packaging when a short sequence of DNA, known as the pac site, is recognized on the P22 genome. However, sequences similar to the pac site in the host genome, called pseudo-pac sites, lead to erroneous packaging and subsequent generalized transduction of Salmonella DNA. While the general genomic locations of the Salmonella pseudo-pac sites are known, the sequences themselves have not been determined. We used visualization of P22 sequencing reads mapped to host Salmonella genomes to define regions of generalized transduction initiation and the likely locations of pseudo-pac sites. We searched each genome region for the sequence with the highest similarity to the P22 pac site and aligned the resulting sequences. We built a regular expression (sequence match pattern) from the alignment and used it to search the genomes of two P22-susceptible Salmonella strains-LT2 and 14028S-for sequence matches. The final regular expression successfully identified pseudo-pac sites in both LT2 and 14028S that correspond with generalized transduction initiation sites in mapped read coverages. The pseudo-pac site sequences identified in this study can be used to predict locations of generalized transduction in other P22-susceptible hosts or to initiate generalized transduction at specific locations in P22-susceptible hosts with genetic engineering. Furthermore, the bioinformatics approach used to identify the Salmonella pseudo-pac sites in this study could be applied to other phage-host systems.

Designing Cell-Type-Specific Promoter Sequences Using Conservative Model-Based Optimization.

Gene therapies have the potential to treat disease by delivering therapeutic genetic cargo to disease-associated cells. One limitation to their widespread use is the lack of short regulatory sequences, or promoters, that differentially induce the expression of delivered genetic cargo in target cells, minimizing side effects in other cell types. Such cell-type-specific promoters are difficult to discover using existing methods, requiring either manual curation or access to large datasets of promoter-driven expression from both targeted and untargeted cells. Model-based optimization (MBO) has emerged as an effective method to design biological sequences in an automated manner, and has recently been used in promoter design methods. However, these methods have only been tested using large training datasets that are expensive to collect, and focus on designing promoters for markedly different cell types, overlooking the complexities associated with designing promoters for closely related cell types that share similar regulatory features. Therefore, we introduce a comprehensive framework for utilizing MBO to design promoters in a data-efficient manner, with an emphasis on discovering promoters for similar cell types. We use conservative objective models (COMs) for MBO and highlight practical considerations such as best practices for improving sequence diversity, getting estimates of model uncertainty, and choosing the optimal set of sequences for experimental validation. Using three relatively similar blood cancer cell lines (Jurkat, K562, and THP1), we show that our approach discovers many novel cell-type-specific promoters after experimentally validating the designed sequences. For K562 cells, in particular, we discover a promoter that has 75.85% higher cell-type-specificity than the best promoter from the initial dataset used to train our models.

From Protein to a Bioactive Peptide: Potent IL‐6 Peptide Antagonist Generated by a Novel Method

ABSTRACTHere, we describe a new technology, designed to accelerate peptide discovery by quick identification and optimization of the residues critical for protein–protein interactions or ligand binding. We called it PepFusion. It is based on ligation of short DNA sequences generated from known ligand‐binding regions. We tested it by selecting peptide antagonists of interleukin‐6 (IL‐6), a key mediator of inflammatory diseases such as rheumatoid arthritis (RA), Crohn's disease, and Castleman disease (CD). The PepFusion library demonstrated superiority over a random library by yielding a peptide with low micromolar affinity for IL‐6, whereas the random library failed. The affinity of the peptide from the PepFusion library was further enhanced by additional rounds of mutagenesis leading to peptide variants with low nanomolar IL‐6 affinity. In addition to generating high‐affinity peptides, our method opens the way to solve the problem of the false positive sequences, which are common with all display technologies.