bioRxiv : the preprint server for biology最新文献

Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR)-associated proteins (Cas) systems have revolutionized genome editing by providing high precision and versatility. However, most genome editing applications rely on a limited number of well-characterized Cas9 and Cas12 variants, constraining the potential for broader genome engineering applications. In this study, we extensively explored Cas9 and Cas12 proteins and developed CasGen, a novel transformer-based deep generative model with margin-based latent space regularization to enhance the quality of newly generative Cas9 and Cas12 proteins. Specifically, CasGen employs a strategies that combine classification to filter out non-Cas sequences, Bayesian optimization of the latent space to guide functionally relevant designs, and thorough structural validation using AlphaFold-based analyses to ensure robust protein generation. We collected a comprehensive dataset with 3,021 Cas9, 597 Cas12, and 597 Non-Cas protein sequences from reputable biological databases such as InterPro and PDB. To validate the generated proteins, we performed sequence alignment using the BLAST tool to ensure novelty and filter out highly similar sequences to existing Cas proteins. Structural prediction using AlphaFold2 and AlphaFold3 confirmed that the generated proteins exhibit high structural similarity to known Cas9 and Cas12 variants, with TM-scores between 0.70 and 0.85 and root-mean-square deviation (RMSD) values below 2.00 Å. Sequence identity analysis further demonstrated that the generated Cas9 orthologs exhibited 28% to 55% identity with known variants, while Cas12a variants show up to 48% identity. Our results demonstrate that the proposed Cas generative model has significant potential to expand the genome editing toolkit by designing diverse Cas proteins that retain functional integrity. The developed deep generative approach offers a promising avenue for synthetic biology and therapeutic applications, enableling the development of more precise and versatile Cas-based genome editing tools.

Purpose: Regenerative therapies for retinal diseases include cell and gene therapy modalities that are targeted to the subretinal space. Several recent clinical trials have shown that the morbidity of surgical access is the major limitation of safe subretinal space delivery. We aimed to develop an image-guided procedure for minimally invasive subretinal access (MISA) as a platform to deliver therapeutic agents for the treatment of degenerative retinal diseases.

Methods: We engineered prototypes of a novel common-path swept source optical coherence tomography (CP-SSOCT)-enabled needle, coaxial guide (COG), and subretinal access cannula (SAC). We pilot tested the MISA procedure in ex vivo bovine eyes and in vivo porcine ocular surgery.

Results: A- and M-mode scan recordings of ex vivo and in vivo animal eye models demonstrated that CP-SSOCT imaging from the scleral side ( ab externo ) was capable of identifying the retinal laminae and the sub-retinal space. We show results from in vivo porcine MISA surgeries (N=4) using the novel CP-SSOCT-enabled sub-retinal injection needle, COG, and SAC through the transscleral approach. The MISA approach enabled subretinal device placement in the posterior pole, however, cases of retinal incarceration and retinal perforation were encountered.

Conclusions: We describe a novel CP-SSOCT-guided subretinal access approach that, with further optimization, may be useful in regenerative retinal surgery.

Learning and memory are fundamental for an individual to be able to respond to changing stimuli in their environment. Between individuals we see variation in their ability to perform learning and memory tasks, however, it is still largely unknown what genetic factors may impact this variability. To gain better insight to the genetic components impacting variation in learning and memory, we use recombinant inbred lines (RILs) from the Drosophila synthetic population resource (DSPR), a multiparent mapping population exhibiting natural variation in many traits. Using a reward based associative learning and memory assay, we trained flies to associate an odor with a sucrose reward under starvation condition and measured olfactory learning and memory ability in y-mazes for 50 DSPR RILs. While we do not find significant QTLs for olfactory learning or memory, we found suggestive regions that may be contributing to variability in performance when trained to different odors. We provide evidence that performance with specific odors should be considered different phenotypes and introduce new methods for analysis for olfactory y-maze assays with multiple decision points. Additionally, we compare our data to previously collected place learning and memory data to show there is limited correlation in performance outcomes.

Gene clusters, groups of physically adjacent genes that work collectively, are pivotal to bacterial fitness and valuable in biotechnology and medicine. While various genome mining tools can identify and characterize gene clusters, they often overlook their evolutionary diversity, a crucial factor in revealing novel cluster functions and applications. To address this gap, we developed GATOR-GC, a targeted genome mining tool that enables comprehensive and flexible exploration of gene clusters in a single execution. We show that GATOR-GC identified a diversity of over 4 million gene clusters similar to experimentally validated biosynthetic gene clusters (BGCs) that other tools fail to detect. To highlight the utility of GATOR-GC, we identified previously uncharacterized co-occurring conserved genes potentially involved in mycosporine-like amino acid biosynthesis and mapped the taxonomic and evolutionary patterns of genomic islands that modify DNA with 7-deazapurines. Additionally, with its proximity-weighted similarity scoring, GATOR-GC successfully differentiated BGCs of the FK-family of metabolites (e.g., rapamycin, FK506/520) according to their chemistries. We anticipate GATOR-GC will be a valuable tool to assess gene cluster diversity for targeted, exploratory, and flexible genome mining. GATOR-GC is available at https://github.com/chevrettelab/gator-gc .

Graphical abstract:

For gene expression traits, cis -genetic heritability can quantify the strength of genetic regulation in particular cell types, elucidating the cell-type-specificity of disease variants and genes. To estimate gene expression heritability, standard models require a single gene expression value per individual, forcing data from single cell RNA-sequencing (scRNA-seq) experiments to be "pseudobulked". Here, we show that applying standard heritability models to pseudobulk data overestimates gene expression heritability and produces inflated false positive rates for detecting cis -heritable genes. Therefore, we introduce a new method called scGeneHE ( s ingle c ell Gene expression H eritability E stimation), a Poisson mixed-effects model that quantifies the cis -genetic component of gene expression using individual cellular profiles. In simulations, scGeneHE has a consistently well-calibrated false positive rate for eGene detection and unbiasedly estimates cis -heritability at many parameter settings. We applied scGeneHE to scRNA-seq data from 969 individuals, 11 immune cell types, and 822,552 cells from the OneK1K cohort to infer cell-type-specificity of genetic regulation at risk genes for immune-mediated diseases and trace the fluctuation of cis -heritability across cellular populations of varying resolution. In summary, we developed a new statistical method that resolves the analytical challenge of estimating gene expression cis -heritability from native scRNA-seq data.

The inference of nonlinear dynamics and parameters in biological data modeling is challenging. Conventional methodologies, based on hypothetical underlying mechanisms, complicate inference because standard parameter optimization methods are difficult to constrain to biological ranges. Here, we propose a novel method to evaluate and improve putative models using neural networks to simultaneously address biological modeling, parametrization, and parameter inference. As an example, utilizing data from clinical frequently sampled intravenous glucose tolerance testing, we introduce two physiological lipolysis models (with parameters) of the dynamics of glucose, insulin, and free fatty acids (FFA). Parameter values are obtained via optimization from the limited clinical data. We then generate large quantities of simulated data from the model by sampling parameters within physiological ranges. A convolutional neural network is trained to take the simulated data time courses of glucose, insulin, and FFA as input and output the model parameters. The performance of the trained neural network is evaluated for both parameter inference and reconstruction of trajectories over a testing dataset and from optimized model-fitting curves. We show that our methodology enables accurate parameter inference and trajectory reconstruction over the testing dataset and optimized model-fitting curves. The trained neural network produces consistently high R ² values and low p -values across different feature engineering strategies and training dataset sizes. We assess the impact of feature engineering choices and training dataset size on inference performance, demonstrating that appropriately designed feature transformations and certain activation function improve accuracy. Our results establish a deep learning framework for parameter inference in mathematical models, which can be adapted to various physiological systems.

The use of cannabis during pregnancy and nursing is a growing public health concern, and the multigenerational impacts of perinatal cannabis exposure remain largely unknown. To address this knowledge gap, we sought to examine the long-term consequences of perinatal cannabis use on reproductive function and how it might impact subsequent generations. Pregnant female mice were exposed to control vehicle or cannabis extract [25, 100, or 200 mg/ml Δ9-tetrahydrocannabinol (THC) in the cannabis extract] from gestational day 1 to postnatal day 21 (twice/day), encompassing the duration of pregnancy through weaning. Based on plasma THC concentrations in F0 females, we chose 100 and 200 mg/ml THC in the cannabis extract for subsequent studies. The selected doses and exposure conditions did not disrupt pregnancy or nursing in F0 females. Pregnancy and neonatal outcomes, including gestational length, litter size, and sexual ratio, were not affected by cannabis exposure. However, cannabis-exposed neonatal F1 pups were smaller. Cannabis exposure delayed vaginal opening as a sign of puberty onset and disrupted estrous cyclicity in F1 females. However, its effects were minor in F2 and F3 females. F1-F3 females showed no abnormal ovarian and uterine histology or plasma estradiol-17β levels and could produce normal offspring without pregnancy issues. These results suggest that the hypothalamus and pituitary are likely perturbed by perinatal cannabis exposure, and the early hypothalamus-pituitary-ovarian axis is disrupted in F1 females. However, they are not sufficient to compromise adult reproductive function. The present results indicate limited transgenerational effects of perinatal cannabis exposure on female reproductive parameters.

Objective: Translation of biological insights from preclinical studies to human disease is a pressing challenge in biomedical research, including in osteoarthritis. Translatable Components Regression (TransComp-R) is a computational framework that has previously been used to synthesize preclinical and human OA data to identify biological pathways predictive of human disease conditions. We aimed to evaluate the translatability of two common murine models of post-traumatic osteoarthritis - surgical destabilization of the medial meniscus (DMM) and noninvasive anterior cruciate ligament rupture (ACLR) - to transcriptomics cartilage data from human OA outcomes.

Design: Transcriptomics cartilage data of DMM and ACLR mouse and human data was acquired from Gene Expression Omnibus. TransComp-R was used to project human OA data into a mouse model (DMM or ACLR) principal component analysis space. The principal components (PCs) were regressed against human OA conditions using increasing complexity of linear regression models incorporating human demographic covariates of OA, sex, and age. Biological pathways of the mouse PCs that significantly stratified human OA and control groups were then interpreted using Gene Set Enrichment Analysis.

Results: From the TransComp-R model, we identified different enriched biological pathways across DMM and ACLR models. While PCs among the DMM models revealed pathways associated with cell signaling and metabolism, ACLR PCs represented immune function and cellular pathways associated with OA condition. The immune pathways presented in the ACLR further highlighted the potential relevance of the OA pathways observed in human conditions.

Conclusions: The ACLR mouse model more successfully predicted human OA conditions, particularly with the human control groups without a history of joint injury or disease. Cross-species translational approaches support the selection of preclinical models intended for therapeutic discovery and pathway analysis in humans.

Myc hyperactivation coordinately regulates numerous metabolic processes to drive lymphomagenesis. Here, we elucidate the temporal and functional relationships between the medley of pathways, factors, and mechanisms that cooperate to control redox homeostasis in Myc-overexpressing B cell lymphomas. We find that Myc overexpression rapidly stimulates the oxidative pentose phosphate pathway (oxPPP), nucleotide synthesis, and mitochondrial respiration, which collectively steers cellular equilibrium to a more oxidative state. We identify Myc-dependent hyperactivation of the phosphoribosyl pyrophosphate synthetase (PRPS) enzyme as a primary regulator of redox status in lymphoma cells. Mechanistically, we show that genetic inactivation of the PRPS2 isozyme, but not PRPS1, in MYC-driven lymphoma cells leads to elevated NADPH levels and reductive stress-mediated death. Employing a pharmacological screen, we demonstrate how targeting PRPS1 or PRPS2 elicits opposing sensitivity or resistance, respectively, to chemotherapeutic agents affecting the thioredoxin and glutathione network, thus providing a therapeutic blueprint for treating MYC-driven lymphomas.

Metabolism and post-translational modifications (PTMs) are intrinsically linked and the number of identified metabolites that can covalently modify proteins continues to increase. This metabolism/PTM crosstalk is especially true for lactate, the product of anaerobic metabolism following glycolysis. Lactate forms an amide bond with the ε-amino group of lysine, a modification known as lysine lactylation, or Kla. Multiple independent mechanisms have been proposed in the formation of Kla, including p300/CBP-dependent transfer from lactyl-CoA, via a high-energy intermediate lactoylglutathione species that non-enzymatically lactylates proteins, and several enzymes are reported to have lactyl transferase capability. We recently discovered that class I histone deacetylases (HDACs) 1, 2, and 3 can all reverse their canonical chemical reaction to catalyze lysine β-hydroxybutyrylation. Here we tested the hypothesis that HDACs can also catalyze Kla formation. Using biochemical, pharmacological, and genetic approaches, we found that HDAC-catalyzed lysine lactylation accounts for the majority of Kla formation in cells. Dialysis experiments confirm this is a reversible reaction that depends on lactate concentration. We also directly quantified intracellular lactyl-CoA and found that Kla abundance can be uncoupled from lactyl-CoA levels. Therefore, we propose a model in which the majority of Kla is formed through enzymatic addition of lactate by HDACs 1, 2, and 3.