Codon usage and expression-based features significantly improve prediction of CRISPR efficiency.

CRISPR is a precise and effective genome editing technology; but despite several advancements during the last decade, our ability to computationally design gRNAs remains limited. Most predictive models have relatively low predictive power and utilize only the sequence of the target site as input. Here we suggest a new category of features, which incorporate the target site genomic position and the presence of genes close to it. We calculate four features based on gene expression and codon usage bias indices. We show, on CRISPR datasets taken from 3 different cell types, that such features perform comparably with 425 state-of-the-art predictive features, ranking in the top 2-12% of features. We trained new predictive models, showing that adding expression features to them significantly improves their r2 by up to 0.04 (relative increase of 39%), achieving average correlations of up to 0.38 on their validation sets; and that these features are deemed important by different feature importance metrics. We believe that incorporating the target site's position, in addition to its sequence, in features such as we have generated here will improve our ability to predict, design and understand CRISPR experiments going forward.

Synonymous and non-synonymous codon substitutions can alleviate dependence on GroEL for folding.

The Escherichia coli GroEL/ES chaperonin system facilitates protein folding in an ATP-driven manner. There are <100 obligate clients of this system in E. coli although GroEL can interact and assist the folding of a multitude of proteins in vitro. It has remained unclear, however, which features distinguish obligate clients from all the other proteins in an E. coli cell. To address this question, we established a system for selecting mutations in mouse dihydrofolate reductase (mDHFR), a GroEL interactor, that diminish its dependence on GroEL for folding. Strikingly, both synonymous and non-synonymous codon substitutions were found to reduce mDHFR's dependence on GroEL. The non-synonymous substitutions increase the rate of spontaneous folding whereas computational analysis indicates that the synonymous substitutions appear to affect translation rates at specific sites.

Strong association between genomic 3D structure and CRISPR cleavage efficiency.

CRISPR is a gene editing technology which enables precise in-vivo genome editing; but its potential is hampered by its relatively low specificity and sensitivity. Improving CRISPR's on-target and off-target effects requires a better understanding of its mechanism and determinants. Here we demonstrate, for the first time, the chromosomal 3D spatial structure's association with CRISPR's cleavage efficiency, and its predictive capabilities. We used high-resolution Hi-C data to estimate the 3D distance between different regions in the human genome and utilized these spatial properties to generate 3D-based features, characterizing each region's density. We evaluated these features based on empirical, in-vivo CRISPR efficiency data and compared them to 425 features used in state-of-the-art models. The 3D features ranked in the top 13% of the features, and significantly improved the predictive power of LASSO and xgboost models trained with these features. The features indicated that sites with lower spatial density demonstrated higher efficiency. Understanding how CRISPR is affected by the 3D DNA structure provides insight into CRISPR's mechanism in general and improves our ability to correctly predict CRISPR's cleavage as well as design sgRNAs for therapeutic and scientific use.

Advanced computational predictive models of miRNA-mRNA interaction efficiency.

The modeling of miRNA-mRNA interactions holds significant implications for synthetic biology and human health. However, this research area presents specific challenges due to the multifaceted nature of mRNA downregulation by miRNAs, influenced by numerous factors including competition or synergism among miRNAs and mRNAs. In this study, we present an improved computational model for predicting miRNA-mRNA interactions, addressing aspects not previously modeled. Firstly, we integrated a novel set of features that significantly enhanced the predictor's performance. Secondly, we demonstrated the cell-specific nature of certain aspects of miRNA-mRNA interactions, highlighting the importance of designing models tailored to specific cell types for improved accuracy. Moreover, we introduce a miRNA binding site interaction model (miBSIM) that, for the first time, accounts for both the distribution of miRNA binding sites along the mRNA and their respective strengths in regulating mRNA stability. Our analysis suggests that distant miRNA sites often compete with each other, revealing the intricate interplay of binding site interactions. Overall, our new predictive model shows a significant improvement of up to 6.43% over previous models in the field. The code of our model is available at https://www.cs.tau.ac.il/~tamirtul/miBSIM.

The Causes for Genomic Instability and How to Try and Reduce Them Through Rational Design of Synthetic DNA.

Genetic engineering has revolutionized our ability to manipulate DNA and engineer organisms for various applications. However, this approach can lead to genomic instability, which can result in unwanted effects such as toxicity, mutagenesis, and reduced productivity. To overcome these challenges, smart design of synthetic DNA has emerged as a promising solution. By taking into consideration the intricate relationships between gene expression and cellular metabolism, researchers can design synthetic constructs that minimize metabolic stress on the host cell, reduce mutagenesis, and increase protein yield. In this chapter, we summarize the main challenges of genomic instability in genetic engineering and address the dangers of unknowingly incorporating genomically unstable sequences in synthetic DNA. We also demonstrate the instability of those sequences by the fact that they are selected against conserved sequences in nature. We highlight the benefits of using ESO, a tool for the rational design of DNA for avoiding genetically unstable sequences, and also summarize the main principles and working parameters of the software that allow maximizing its benefits and impact.

Detecting and understanding meaningful cancerous mutations based on computational models of mRNA splicing.

Cancer research has long relied on non-silent mutations. Yet, it has become overwhelmingly clear that silent mutations can affect gene expression and cancer cell fitness. One fundamental mechanism that apparently silent mutations can severely disrupt is alternative splicing. Here we introduce Oncosplice, a tool that scores mutations based on models of proteomes generated using aberrant splicing predictions. Oncosplice leverages a highly accurate neural network that predicts splice sites within arbitrary mRNA sequences, a greedy transcript constructor that considers alternate arrangements of splicing blueprints, and an algorithm that grades the functional divergence between proteins based on evolutionary conservation. By applying this tool to 12M somatic mutations we identify 8K deleterious variants that are significantly depleted within the healthy population; we demonstrate the tool's ability to identify clinically validated pathogenic variants with a positive predictive value of 94%; we show strong enrichment of predicted deleterious mutations across pan-cancer drivers. We also achieve improved patient survival estimation using a proposed set of novel cancer-involved genes. Ultimately, this pipeline enables accelerated insight-gathering of sequence-specific consequences for a class of understudied mutations and provides an efficient way of filtering through massive variant datasets - functionalities with immediate experimental and clinical applications.

Fighting the battle against evolution: designing genetically modified organisms for evolutionary stability.

Synthetic biology has made significant progress in many areas, but a major challenge that has received limited attention is the evolutionary stability of synthetic constructs made of heterologous genes. The expression of these constructs in microorganisms, that is, production of proteins that are not necessary for the organism, is a metabolic burden, leading to a decrease in relative fitness and make the synthetic constructs unstable over time. This is a significant concern for the synthetic biology community, particularly when it comes to bringing this technology out of the laboratory. In this review, we discuss the issue of evolutionary stability in synthetic biology and review the available tools to address this challenge.

eIF3 mRNA selectivity profiling reveals eIF3k as a cancer-relevant regulator of ribosome content.

eIF3, whose subunits are frequently overexpressed in cancer, regulates mRNA translation from initiation to termination, but mRNA-selective functions of individual subunits remain poorly defined. Using multiomic profiling upon acute depletion of eIF3 subunits, we observed that while eIF3a, b, e, and f markedly differed in their impact on eIF3 holo-complex formation and translation, they were each required for cancer cell proliferation and tumor growth. Remarkably, eIF3k showed the opposite pattern with depletion promoting global translation, cell proliferation, tumor growth, and stress resistance through repressing the synthesis of ribosomal proteins, especially RPS15A. Whereas ectopic expression of RPS15A mimicked the anabolic effects of eIF3k depletion, disruption of eIF3 binding to the 5'-UTR of RSP15A mRNA negated them. eIF3k and eIF3l are selectively downregulated in response to endoplasmic reticulum and oxidative stress. Supported by mathematical modeling, our data uncover eIF3k-l as a mRNA-specific module which, through controlling RPS15A translation, serves as a rheostat of ribosome content, possibly to secure spare translational capacity that can be mobilized during stress.

Altered somatic hypermutation patterns in COVID-19 patients classifies disease severity.

Introduction: The success of the human body in fighting SARS-CoV2 infection relies on lymphocytes and their antigen receptors. Identifying and characterizing clinically relevant receptors is of utmost importance. Methods: We report here the application of a machine learning approach, utilizing B cell receptor repertoire sequencing data from severely and mildly infected individuals with SARS-CoV2 compared with uninfected controls. Results: In contrast to previous studies, our approach successfully stratifies non-infected from infected individuals, as well as disease level of severity. The features that drive this classification are based on somatic hypermutation patterns, and point to alterations in the somatic hypermutation process in COVID-19 patients. Discussion: These features may be used to build and adapt therapeutic strategies to COVID-19, in particular to quantitatively assess potential diagnostic and therapeutic antibodies. These results constitute a proof of concept for future epidemiological challenges.

A computational predictor of the anaerobic mechanical power outputs from a clinical exercise stress test.

We previously were able to predict the anaerobic mechanical power outputs using features taken from a maximal incremental cardiopulmonary exercise stress test (CPET). Since a standard aerobic exercise stress test (with electrocardiogram and blood pressure measurements) has no gas exchange measurement and is more popular than CPET, our goal, in the current paper, was to investigate whether features taken from a clinical exercise stress test (GXT), either submaximal or maximal, can predict the anaerobic mechanical power outputs to the same level as we found with CPET variables. We have used data taken from young healthy subjects undergoing CPET aerobic test and the Wingate anaerobic test, and developed a computational predictive algorithm, based on greedy heuristic multiple linear regression, which enabled the prediction of the anaerobic mechanical power outputs from a corresponding GXT measures (exercise test time, treadmill speed and slope). We found that for submaximal GXT of 85% age predicted HRmax, a combination of 3 and 4 variables produced a correlation of r = 0.93 and r = 0.92 with % error equal to 15 ± 3 and 16 ± 3 on the validation set between real and predicted values of the peak and mean anaerobic mechanical power outputs (p < 0.001), respectively. For maximal GXT (100% of age predicted HRmax), a combination of 4 and 2 variables produced a correlation of r = 0.92 and r = 0.94 with % error equal to 12 ± 2 and 14 ± 3 on the validation set between real and predicted values of the peak and mean anaerobic mechanical power outputs (p < 0.001), respectively. The newly developed model allows to accurately predict the anaerobic mechanical power outputs from a standard, submaximal and maximal GXT. Nevertheless, in the current study the subjects were healthy, normal individuals and therefore the assessment of additional subjects is desirable for the development of a test applicable to other populations.

Determinants of associations between codon and amino acid usage patterns of microbial communities and the environment inferred based on a cross-biome metagenomic analysis.

Codon and amino acid usage were associated with almost every aspect of microbial life. However, how the environment may impact the codon and amino acid choice of microbial communities at the habitat level is not clearly understood. Therefore, in this study, we analyzed codon and amino acid usage patterns of a large number of environmental samples collected from diverse ecological niches. Our results suggested that samples derived from similar environmental niches, in general, show overall similar codon and amino acid distribution as compared to samples from other habitats. To substantiate the relative impact of the environment, we considered several factors, such as their similarity in GC content, or in functional or taxonomic abundance. Our analysis demonstrated that none of these factors can fully explain the trends that we observed at the codon or amino acid level implying a direct environmental influence on them. Further, our analysis demonstrated different levels of selection on codon bias in different microbial communities with the highest bias in host-associated environments such as the digestive system or oral samples and the lowest level of selection in soil and water samples. Considering a large number of metagenomic samples here we showed that microorganisms collected from similar environmental backgrounds exhibit similar patterns of codon and amino acid usage irrespective of the location or time from where the samples were collected. Thus our study suggested a direct impact of the environment on codon and amino usage of microorganisms that cannot be explained considering the influence of other factors.

Translation in the cell under fierce competition for shared resources: a mathematical model.

During translation, mRNAs 'compete' for shared resources. Under stress conditions, during viral infection and also in high-throughput heterologous gene expression, these resources may become scarce, e.g. the pool of free ribosomes is starved, and then the competition may have a dramatic effect on the global dynamics of translation in the cell. We model this scenario using a network that includes m ribosome flow models (RFMs) interconnected via a pool of free ribosomes. Each RFM models ribosome flow along an mRNA molecule, and the pool models the shared resource. We assume that the number of mRNAs is large, so many ribosomes are attached to the mRNAs, and the pool is starved. Our analysis shows that adding an mRNA has an intricate effect on the total protein production. The new mRNA produces new proteins, but the other mRNAs produce less proteins, as the pool that feeds these mRNAs now has a smaller abundance of ribosomes. As the number of mRNAs increases, the marginal utility of adding another mRNA diminishes, and the total protein production rate saturates to a limiting value. We demonstrate our approach using an example of insulin protein production in a cell-free system.

Modulating Gene Expression within a Microbiome Based on Computational Models.

Recent research in the field of bioinformatics and molecular biology has revealed the immense complexity and uniqueness of microbiomes, while also showcasing the impact of the symbiosis between a microbiome and its host or environment. A core property influencing this process is horizontal gene transfer between members of the bacterial community used to maintain genetic variation. The essential effect of this mechanism is the exposure of genetic information to a wide array of members of the community, creating an additional "layer" of information in the microbiome named the "plasmidome". From an engineering perspective, introduction of genetic information to an environment must be facilitated into chosen species which will be able to carry out the desired effect instead of competing and inhibiting it. Moreover, this process of information transfer imposes concerns for the biosafety of genetic engineering of microbiomes as exposure of genetic information into unwanted hosts can have unprecedented ecological impacts. Current technologies are usually experimentally developed for a specific host/environment, and only deal with the transformation process itself at best, ignoring the impact of horizontal gene transfer and gene-microbiome interactions that occur over larger periods of time in uncontrolled environments. The goal of this research was to design new microbiome-specific versions of engineered genetic information, providing an additional layer of compatibility to existing engineering techniques. The engineering framework is entirely computational and is agnostic to the selected microbiome or gene by reducing the problem into the following set up: microbiome species can be defined as wanted or unwanted hosts of the modification. Then, every element related to gene expression (e.g., promoters, coding regions, etc.) and regulation is individually examined and engineered by novel algorithms to provide the defined expression preferences. Additionally, the synergistic effect of the combination of engineered gene blocks facilitates robustness to random mutations that might occur over time. This method has been validated using both computational and experimental tools, stemming from the research done in the iGEM 2021 competition, by the TAU group.

Modeling the effect of rRNA-mRNA interactions and mRNA folding on mRNA translation in chloroplasts.

The process of translation initiation in prokaryotes is mediated by the hybridization of the 16S rRNA of the small ribosomal subunit with the mRNA in a short region called the ribosomal binding site. However, translation initiation in chloroplasts, which have evolved from an ancestral bacterium, is not well understood. Some studies suggest that in many cases it differs from translation initiation in bacteria and involves various novel interactions of the mRNA structures with intracellular factors; however currently, there is no generic quantitative model related to these aspects in chloroplasts. We developed a novel computational pipeline and models that can be used for understanding and modeling translation regulation in chloroplasts. We demonstrate that local folding and co-folding energy of the rRNA and the mRNA correlates with codon usage estimators of expression levels (r = -0.63) and infer predictive models that connect these energies and codon usage to protein levels (with correlation up to 0.71). In addition, we demonstrate that the ends of the transcripts in chloroplasts are populated with various structural elements that may be functional. Furthermore, we report a database of 166 novel structures in the chloroplast transcripts that are predicted to be functional. We believe that the models reported here improve existing understandings of genomic evolution and the biophysics of translation in chloroplasts; as such, they can aid gene expression engineering in chloroplasts for various biotechnological objectives.

MPEPE, a predictive approach to improve protein expression in E. coli based on deep learning.

The expression of proteins in Escherichia coli is often essential for their characterization, modification, and subsequent application. Gene sequence is the major factor contributing expression. In this study, we used the expression data from 6438 heterologous proteins under the same expression condition in E. coli to construct a deep learning classifier for screening high- and low-expression proteins. In conjunction with conserved residue analysis to minimize functional disruption, a mutation predictor for enhanced protein expression (MPEPE) was proposed to identify mutations conducive to protein expression. MPEPE identified mutation sites in laccase 13B22 and the glucose dehydrogenase FAD-AtGDH, that significantly increased both expression levels and activity of these proteins. Additionally, a significant correlation of 0.46 between the predicted high level expression propensity with the constructed models and the protein abundance of endogenous genes in E. coli was also been detected. Therefore, the study provides foundational insights into the relationship between specific amino acid usage, codon usage, and protein expression, and is essential for research and industrial applications.

Modeling the ribosomal small subunit dynamic in Saccharomyces cerevisiae based on TCP-seq data.

Translation Complex Profile Sequencing (TCP-seq), a protocol that was developed and implemented on Saccharomyces cerevisiae, provides the footprints of the small subunit (SSU) of the ribosome (with additional factors) across the entire transcriptome of the analyzed organism. In this study, based on the TCP-seq data, we developed for the first-time a predictive model of the SSU density and analyzed the effect of transcript features on the dynamics of the SSU scan in the 5'UTR. Among others, our model is based on novel tools for detecting complex statistical relations tailored to TCP-seq. We quantitatively estimated the effect of several important features, including the context of the upstream AUG, the upstream ORF length and the mRNA folding strength. Specifically, we suggest that around 50% of the variance related to the read counts (RC) distribution near a start codon can be attributed to the AUG context score. We provide the first large scale direct quantitative evidence that shows that indeed AUG context affects the small sub-unit movement. In addition, we suggest that strong folding may cause the detachment of the SSU from the mRNA. We also identified a number of novel sequence motifs that can affect the SSU scan; some of these motifs affect transcription factors and RNA binding proteins. The results presented in this study provide a better understanding of the biophysical aspects related to the SSU scan along the 5'UTR and of translation initiation in S. cerevisiae, a fundamental step toward a comprehensive modeling of initiation.

Evolutionary Stability Optimizer (ESO): A Novel Approach to Identify and Avoid Mutational Hotspots in DNA Sequences While Maintaining High Expression Levels.

Modern synthetic biology procedures rely on the ability to generate stable genetic constructs that keep their functionality over long periods of time. However, maintenance of these constructs requires energy from the cell and thus reduces the host's fitness. Natural selection results in loss-of-functionality mutations that negate the expression of the construct in the population. Current approaches for the prevention of this phenomenon focus on either small-scale, manual design of evolutionary stable constructs or the detection of mutational sites with unstable tendencies. We designed the Evolutionary Stability Optimizer (ESO), a software tool that enables the large-scale automatic design of evolutionarily stable constructs with respect to both mutational and epigenetic hotspots and allows users to define custom hotspots to avoid. Furthermore, our tool takes the expression of the input constructs into account by considering the guanine-cytosine (GC) content and codon usage of the host organism, balancing the trade-off between stability and gene expression, allowing to increase evolutionary stability while maintaining the high expression. In this study, we present the many features of the ESO and show that it accurately predicts the evolutionary stability of endogenous genes. The ESO was created as an easy-to-use, flexible platform based on the notion that directed genetic stability research will continue to evolve and revolutionize current applications of synthetic biology. The ESO is available at the following link: https://www.cs.tau.ac.il/~tamirtul/ESO/.

Determinants of efficient modulation of ribosomal traffic jams.

mRNA translation is the process which consumes most of the cellular energy. Thus, this process is under strong evolutionary selection for its optimization and rational optimization or reduction of the translation efficiency can impact the cell growth rate. Algorithms for modulating cell growth rate can have various applications in biotechnology, medicine, and agriculture. In this study, we demonstrate that the analysis of these algorithms can also be used for understanding translation. We specifically describe and analyze various generic algorithms, based on comprehensive computational models and whole cell simulations of translation, for introducing silent mutations that can either reduce or increase ribosomal traffic jams along the mRNA. As a result, more or less resources are available, for the cell, promoting improved or reduced cells growth-rate, respectively. We then explore the cost of these algorithms' performance, in terms of their computational time, the number of mutations they introduce, the modified genomic region, the effect on local translation rates, and the properties of the modified genes. Among others, we show that mRNA levels of a gene are much stronger predictors for the effect of its engineering on the ribosomal pool than the ribosomal density of the gene. We also demonstrate that the mutations at the ends of the coding regions have a stronger effect on the ribosomal pool. Furthermore, we report two optimization algorithms that exhibit a tread-off between the number of mutations they introduce and their executing time. The reported results here are fundamental both for understanding the biophysics and evolution of translation, as well as for developing efficient approaches for its engineering.

Accelerating Whole-Cell Simulations of mRNA Translation Using a Dedicated Hardware.

In recent years, intracellular biophysical simulations have been used with increasing frequency not only for answering basic scientific questions but also in the field of synthetic biology. However, since these models include networks of interaction between millions of components, they are extremely time-consuming and cannot run easily on parallel computers. In this study, we demonstrate for the first time a novel approach addressing this challenge by using a dedicated hardware designed specifically to simulate such processes. As a proof of concept, we specifically focus on mRNA translation, which is the process consuming most of the energy in the cell. We design a hardware that simulates translation in Escherichia coli and Saccharomyces cerevisiae for thousands of mRNAs and ribosomes, which is in orders of magnitude faster than a similar software solution. With the sharp increase in the amount of genomic data available today and the complexity of the corresponding models inferred from them, we believe that the strategy suggested here will become common and can be used among others for simulating entire cells with all gene expression steps.

A Mathematical Analysis of HDV Genotypes: From Molecules to Cells.

Hepatitis D virus (HDV) is classified according to eight genotypes. The various genotypes are included in the HDVdb database, where each HDV sequence is specified by its genotype. In this contribution, a mathematical analysis is performed on RNA sequences in HDVdb. The RNA folding predicted structures of the Genbank HDV genome sequences in HDVdb are classified according to their coarse-grain tree-graph representation. The analysis allows discarding in a simple and efficient way the vast majority of the sequences that exhibit a rod-like structure, which is important for the virus replication, to attempt to discover other biological functions by structure consideration. After the filtering, there remain only a small number of sequences that can be checked for their additional stem-loops besides the main one that is known to be responsible for virus replication. It is found that a few sequences contain an additional stem-loop that is responsible for RNA editing or other possible functions. These few sequences are grouped into two main classes, one that is well-known experimentally belonging to genotype 3 for patients from South America associated with RNA editing, and the other that is not known at present belonging to genotype 7 for patients from Cameroon. The possibility that another function besides virus replication reminiscent of the editing mechanism in HDV genotype 3 exists in HDV genotype 7 has not been explored before and is predicted by eigenvalue analysis. Finally, when comparing native and shuffled sequences, it is shown that HDV sequences belonging to all genotypes are accentuated in their mutational robustness and thermodynamic stability as compared to other viruses that were subjected to such an analysis.