Excluding false positives is critical for interpreting CRISPR screens. Here, we introduce a new Chronos module for estimating false discovery rates for identifying knockouts that cause loss of viability or have differential viability effects in different conditions. We introduce a rigorous benchmarking framework using real CRISPR data. We show with multiple real datasets that existing methods such as MAGeCK are miscalibrated and can generate uncontrolled numbers of false positives even after multiple hypothesis correction. Only Chronos correctly controls false discovery for all tested tasks. Additionally, Chronos’s estimates are well-calibrated, allowing users to accurately specify the acceptable false discovery rate.

Hereditary dystonia, particularly the isolated form of early onset, is often caused by a deletion of the GAG in the TOR1A gene, leading to a dysfunctional TorsinA protein and severe motor impairments. In this study, we investigate an AI-enhanced CRISPR prime editing framework designed for the precise correction of the TOR1A mutation. Our framework integrates the generation of candidate pegRNA with an empirical simulation of training data, enabling the tuning of a random forest model that predicts editing efficiency. This approach generates robust quantitative outputs (validation R 2 score of 0.701). SHapley Additive exPlanations (SHAP) revealed that GC content (mean SHAP = 0.22) and off-target risk (mean SHAP = 0.18) were the strongest drivers of predicted efficiency. The GC content showed strong positive correlations with melting temperature (r = 0.75) and guide efficiency (r = 0.82). By linking the simulated modeling work to the practical challenges of genetic neurosurgery in dystonia, a proof-of-concept framework is presented that is both ethically governed and clinically pertinent. This work looks toward a future of scalpel-less surgery by laying a foundation for future integration with empirical datasets and more advanced deep learning techniques.

Animals respond to environmental cues to time phenological events, but the intrinsic mechanism of circannual timing remains elusive. We used transcriptomic sequencing and frequent sampling, during three distinct phases (induction, maintenance and recovery) of circannual interval timing for Djungarian hamster energy balance, to investigate the molecular architecture of a neuroendocrine seasonal clock. Our study identified three distinct phases of transcript changes, with deiodinase type-3 (Dio3) expression activated during the early induction phase. Subsequent work demonstrated that targeted mutation of Dio3 resulted in a shorter period for circannual interval timing. Hamsters that exhibit naturally disrupted Dio3 expression do not show any change in body mass or pelage. Our work demonstrates that changes in Dio3 induction is essential for setting the period of circannual interval timing.

To gain access to the earliest stages of T cell development, we adapted a serum-free culture system that expands hematopoietic stem and progenitor-like cells. These expanded cells efficiently undergo normal T-cell differentiation in vivo and in vitro , verified by early gene expression trajectories from single-cell RNA sequencing, though their absolute differentiation speed is slower than that of fresh progenitors and can be modulated with cytokine priming. Leveraging this expansion system to observe the first T-lineage events, we revealed that initial Notch activation immediately induces chromatin opening and transcriptional activation of the TCR-Cβ locus. Additionally, acute CRISPR knockouts confirmed T-lineage entry requirements for Ikzf1 , Hes1 , Gabpa , and Myb while revealing that Lmo2 , Erg, Spi1, Hoxa9 , and Meis1 retard developmental progression with differing effects on proliferation. Endogenous expression of the stem, progenitor, and leukemia-associated factor Lmo2 markedly restrains initiation of the T cell program, with Lmo2 knockout greatly accelerating germline TCRβ locus transcription and expression of Tcf7, Gata3, Runx family, and E protein genes and their targets.

Treatment with immune checkpoint inhibitors induces remarkable clinical responses in several cancer types. However, most cancer patients fail to respond to immunotherapy, and patients who initially respond often exhibit acquired resistance. Understanding the universe of immune evasion strategies will enable design of more effective immunotherapies. Here, we identify genes that drive immune evasion using genome-scale in vivo CRISPR gain-of-function screens in tumors treated with anti-PD-1 antibodies and found that the transcription factor CREB5 drives immune checkpoint blockade resistance. Using transcriptional profiling and functional studies, we show that CREB5 promotes a mesenchymal-like phenotype in melanoma characterized by upregulation of extracellular matrix genes including collagen and collagen-stabilizing factors. Using engineered tumor models and knockout mice, we found that immunotherapy resistance is functionally mediated by tumor-intrinsic collagen deposition. Collagen is the major ligand for the inhibitory receptor LAIR1, broadly expressed on T cells, B cells, NK cells, and myeloid cells. Deletion of LAIR1 in mice or overexpression of the decoy receptor LAIR2 in tumors abrogated the resistance induced by CREB5 overexpression, demonstrating that collagen-LAIR1 inhibitory signaling drives resistance to immune checkpoint inhibitors. These observations define a transcriptional program that remodels the tumor microenvironment to promote immunotherapy resistance via extracellular matrix deposition and indicates that targeting this pathway may enhance immunotherapy efficacy. One-Sentence Summary: In vivo gain-of-function screening in immunotherapy-treated mice reveals a transcription factor, Creb5 , that drives the mesenchymal state in melanoma and facilitates immune escape by promoting tumor-intrinsic collagen matrix deposition.

ABSTRACT The cofactor Ldb1 (Chip) is linked to many processes in gene regulation, including enhancer-promoter communication, inter-chromosomal interactions and enhanceosome-cofactor-like activity. However, its functional requirement and molecular role during embryogenesis has not been assessed to date. Here, we used optogenetics (iLEXY), to rapidly deplete Drosophila Ldb1 (Chip) from the nucleus at precise time windows. Remarkably, this pinpointed the essential window of Chip’s function in just one-hour of embryogenesis, overlapping zygotic genome activation (ZGA). We show that Zelda, a pioneer factor essential for ZGA, recruits Chip to chromatin, and both factors regulate concordant changes in gene expression, suggesting that Chip is a cofactor of Zelda. Surprisingly, Chip is not required for chromatin architecture at these stages, instead it recruits CBP, and is essential for the placement of H3K27ac. Taken together, our results identify Chip (Ldb1) as a functional bridge between Zelda and the coactivator CBP to regulate gene expression in early embryogenesis.

Bacteriophages, viruses that infect bacteria, are pivotal in therapeutic, industrial, and bio-detection applications due to their unique ability to inject DNA into bacterial hosts and change the genetics and behavior of whole bacterial populations. Genetic engineering of phage genomes has expanded their potential applications with several established methods such as Golden Gate assembly, yeast cloning, λ Red recombineering, and CRISPR-Cas systems. Here, we present a novel, efficient method to design and make synthetic bacteriophages in vitro without using restriction enzymes to allow for modular insertion of DNA fragments into the phage genome. The ability to create synthetic bacteriophages in vitro without the use of restriction enzymes allows for simpler engineering without the hassle or cloning limitations encountered when building domesticated bacteriophage genomes.

ABSTRACT Gene drives offer revolutionary potential for the management of problematic plant populations, such as invasive weeds and herbicide-resistant species, by rapidly spreading desired genetic alterations. Two recent studies have provided the first experimental demonstrations of engineered CRISPR gene drive systems in plants (CAIN and ClvR). However, the successful application of such systems in the field will critically depend on an accurate understanding of plant-specific life-history traits, especially seed dormancy, a ubiquitous yet frequently overlooked eco-evolutionary force. In this study, we develop the first comprehensive modeling framework for gene drives in plant populations that incorporates a persistent soil seedbank. We show how the presence of a seedbank can significantly slow gene drive spread but also reduce the genetic load required to achieve population elimination. Furthermore, we show that seedbanks substantially increase the required introduction frequency of threshold-dependent gene drives, which could prevent establishment in some cases, yet also provide an intrinsic biosafety mechanism for confining a highly efficient drive to a target population. Our study highlights the need to incorporate seedbank dynamics into gene drive strategies to ensure realistic predictions and successful field applications.

Phenotypic variation arises from the interplay between genetic and environmental factors. However, disentangling these interactions for complex traits remains challenging in observational cohorts such as human biobanks. Instead, model organisms where genetic (G) and environmental (E) variation can be controlled offer a valuable complement to human studies in the analysis of higher-order genetic effects such as GxE interactions, dominance, and epistasis. Here, we utilized 76 medaka strains of the Medaka Inbred Kiyosu-Karlsruhe (MIKK) panel, to compare heart rate plasticity across temperatures. An F2 segregation analysis identified 16 quantitative trait loci (QTLs), with many exhibiting dominance, GxE, GxG, and GxGxE interactions. We experimentally validated four candidate genes using gene editing, revealing their temperature-sensitive impact on heart function. Finally, we devised simulations to assess how GWAS discovery power is influenced by the choice of statistical models, showing that the apparent additivity in human GWAS is to be expected given study design and sample sizes of current studies. This work demonstrates the value of controlled model organism studies for dissecting the genetics of complex traits and provides guidance on the design of genetic association studies.

Mutation rates can increase substantially under environmental stress, known as stress-induced mutagenesis. Specifically, heat stress has been shown to elevate mutation rates, thereby enhancing genetic variability and facilitating adaptation. However, the underlying mechanisms remain elusive in eukaryotes. Here, we investigated how heat stress affects DNA repair and induces mutations both locally and globally through CRISPR-Cas9 targeted DNA breaks and whole genome sequencing analyses in Arabidopsis thaliana . Heat stress was found to enhance CRISPR editing efficiency across all chromatin contexts, with particularly significant increases, up to 29.9-fold, in heterochromatic regions. Moreover, heat stress consistently shifts mutation outcomes toward 1-bp insertions regardless of chromatin states. We identified a heat-inducible, error-prone DNA polymerase, Polλ, as the key mediator of mutation profile changes. When extending our investigation from targeted mutations to genome-wide effects, we found that increases in global mutation rates under heat stress are also dependent on Polλ. Single-cell transcriptomic analysis further demonstrated that Polλ expression is tightly regulated and cell-type specific, with the highest expression levels in central zone meristematic cells. Together, these findings provide practical applications for improving editing efficiency in heterochromatic regions and fundamental insights into heat-induced mutagenesis, establishing Polλ as a crucial mediator of stress-induced genetic variation in plants.

Abstract Background Genome-wide profiling of DNA-protein interactions in cells can provide important information about mechanisms of gene regulation. Most current methods for genome-wide profiling of DNA-bound proteins such as ChIP-seq and CUT&Tag use conventional IgG antibodies to bind the target protein(s). This limits their applicability to targets with available high affinity and specificity antibodies and prevents their use for other targets. Here we describe NanoTag, an IgG-free method derived from CUT&Tag to profile DNA-protein interactions. NanoTag is based on a fusion between an anti-GFP nanobody and Tn5 transposase that can map GFP-tagged proteins associated with chromatin in a fast, cost-effective and animal-free manner. Results We used NanoTag to indirectly profile the histone mark H3K4me3 genome-wide via its binding partner TATA box-binding protein-associated factor 3 (TAF3) and the transcription factors Nanog and CTCF in mouse embryonic stem cells (mESCs). NanoTag results show high inter-replicate reproducibility, high signal-to-noise ratio and strong correlation with CUT&Tag datasets, validating its accuracy and reliability. Conclusions NanoTag provides a novel, flexible and cost-effective IgG-free method to generate high resolution DNA-binding profiles in cells and tissues.

ABSTRACT Epithelial fusion is a fundamental morphogenetic process critical for the closure and compartmentalisation of developing organs. While widely studied in systems such as neural tube and palatal closure, the cellular transitions that enable fusion remain poorly understood. Here, we investigate epithelial fusion during chick otic vesicle (OV) closure and identify a transient population of cells at the epithelial interface that mediate this process. These otic edge (OE) cells exhibit distinct morphology, reduced apicobasal polarity, and dynamic junctional remodelling, including altered distribution of ZO-1, CDH1, and RAC1. Notably, OE cells lack basal contact and display high sphericity, consistent with a partial epithelial-to-mesenchymal transition (EMT) phenotype. Transcriptomic profiling of microdissected tissues reveals that OE cells constitute a transcriptionally distinct population, enriched for EMT regulators, ECM remodelling genes, and WNT pathway components. Among these, the transcription factors Grhl2 and Sp8 were specifically expressed at the OE and exhibited opposing roles in epithelial identity. CRISPR-Cas9 mediated knockdown of either gene led to disrupted CDH1 localisation, loss of OE cell morphology, and failure in epithelial segregation. These results suggest that epithelial fusion requires a regulated, hybrid EMT state that balances junctional plasticity with tissue cohesion. Our findings demonstrate that fusion-competent epithelial cells are not merely passive participants but actively modulate their shape, polarity, adhesion, and genetic identity to enable morphogenesis.

CRISPR interference (CRISPRi) has emerged as a valuable tool for redirecting metabolic flux to enhance bioproduction. However, conventional approaches for identifying target genes for CRISPRi-mediated downregulation have largely relied on heuristic methods and trial and error, which are labor-intensive and time-consuming. Additionally, the ability to achieve multigene knockdowns is limited by the constraints of molecular cloning techniques required for building multiplexed CRISPRi systems. In this study, we describe two novel methodologies FluxRETAP (Flux-REaction TArget Prioritization, a Genome-Scale Modeling Technique) and VAMMPIRE (a Versatile Assembly Method for MultiPlexing CRISPRi-mediated downREgulation). FluxRETAP accurately identified gene targets whose knockdown led to substantial increase of isoprenol titers, outperforming traditional heuristic selection. The use of VAMMPIRE enabled accurate and position-independent assembly of CRISPRi constructs containing up to five sgRNA arrays. The integration of FluxRETAP and VAMMPIRE has the potential to advance metabolic engineering by rapidly identifying CRISPRi-mediated knockdowns and knockdown combinations that enhance bioproduction titers, with potential applicability to other microbial systems.

The RNA-binding protein Sex-lethal (Sxl) is classically known as a master regulator of sex determination and mRNA splicing in Drosophila melanogaster . However, this role is not conserved across species, and functions beyond this canonical pathway remain poorly understood. In this study, we uncover a splicing-independent role for Sxl at the chromatin level in the Drosophila brain. Using Targeted DamID (TaDa) profiling in neurons, we identify widespread recruitment of Sxl to promoter regions, independent of sex or RNA binding activity. Notably, Sxl chromatin occupancy exhibits near-complete overlap with Polr3E (RPC37), an RNA Polymerase III subunit, with Sxl binding abolished upon Polr3E knockdown. Depletion of Sxl in mature male neurons induces widespread transcriptional changes, particularly in metabolic genes, and improves negative geotaxis during ageing, phenotypes that closely mirror Polr3E knockdown. Conversely, overexpression of the brain-specific Sxl RAC transcript leads to enhanced tRNA synthesis and upregulated metabolic gene expression. Together, these findings reveal a previously unrecognised role for Sxl in regulating Pol III activity via Polr3E, regulating tRNA synthesis and supporting neuronal metabolism. Given the emerging tie between Pol III regulation and neuronal ageing, our study highlights Sxl as a novel modulator of neuronal homeostasis.

ALK+ Anaplastic Large Cell Lymphoma (ALCL) is an aggressive T-cell lymphoma that is characterized by expression of the Anaplastic Lymphoma Kinase (ALK), which is induced by the t(2;5) chromosomal rearrangement leading to the expression of the NPM-ALK fusion-oncogene. Most previous preclinical models of ALK+ ALCL were based on overexpression of the NPM-ALK cDNA from heterologous promoters. Due to the enforced expression, this approach is prone to artifacts arising from synthetic overexpression, promoter competition and insertional variation. To improve the existing ALCL models and more closely recapitulate the oncogenic events in ALK+ ALCL, we employed CRISPR/Cas-based chromosomal engineering to selectively introduce translocations between the Npm1 and Alk gene loci in murine cells. By inducing precise DNA cleavage at the syntenic loci on chromosome 11 and 17 in a murine IL-3-dependent Ba/F3 reporter cell line, we generated de novo Npm-Alk translocations in-vivo, leading to IL-3-independent cell growth. To verify efficient recombination, we analyzed the expression of the Npm-Alk fusion protein in the recombined cells and could also show the t(11;17) in the IL-3 independent Ba/F3 cells. Subsequent functional testing of these cells using an Alk-inhibitor showed exquisite responsiveness towards Crizotinib, demonstrating strong dependence on the newly generated Alk fusion oncoprotein. Furthermore, a comparison of the gene expression pattern between Ba/F3 cells overexpressing the Npm-Alk cDNA with Ba/F3 cells transformed by CRISPR-mediated Npm-Alk translocation indicated that, while broadly overlapping, a set of pathways including the unfolded protein response pathway was increased in the Npm-Alk overexpression model, suggesting increased reactive changes induced by exogenous overexpression of Npm-Alk. Furthermore, we observed clustered expression changes in genes located in chromosomal regions close to the breakpoint in the new CRISPR-based model, indicating positional effects on gene expression mediated by the translocation event, which are not part of the older models. Thus, CRISPR-mediated recombination provides a novel and more faithful approach to model oncogenic translocations, which may lead to an improved understanding of the molecular pathogenesis of ALCL and enable more accurate therapeutic models of malignancies driven by oncogenic fusion proteins.

SARS-CoV-2 replication remains a critical and main target for therapeutic interventions. The current review synthesizes existing knowledge to provide an in-depth analysis of molecular insights and both current and emerging therapy methods, moving past reviews centered on antivirals and general replication processes. We examined antiviral tactics aimed at these replication pathways, including direct-acting nucleoside analogs (remdesivir, molnupiravir), protease inhibitors (nirmatrelvir), and host-directed agents influencing viral entry and RNA synthesis. Emphasizing therapeutic constraints and evolutionary escape, this study also investigates synergistic drug combinations and resistance mechanisms. Discussed for their capacity to efficiently handle next coronavirus dangers are emerging methods—from CRISPR-based gene-silencing, nanoparticle-delivered siRNAs, and AI-driven drug discovery. Highlighted as new antiviral targets are host-pathogen interactions including adaptation via the TRiC complex and phosphatase pathways. This paper offers a road map for improving therapeutic tactics against SARS-CoV-2 and related developing viruses with molecular virology, pharmacology, and computational biology.

ABSTRACT Objective To investigate ESR1 and estrogen-driven transcription in human endometrial stromal cells. Design Telomerase-immortalized human endometrial stromal cells were engineered to activate ESR1 using the CRISPR activation system and treated with vehicle or E2. Primary endometrial stromal cells were treated with vehicle or decidualization cocktail. Subjects Biopsies from two healthy, reproductive-aged volunteers with regular menstrual cycles and no history of gynecological malignancies. Main Outcome Measures Bulk RNA-sequencing in ESR1-activated and E2-treated cells was compared to identify ligand-independent and -dependent ESR1 actions. Cut&Run was performed in ESR1-activated cells treated with E2 or vehicle to assess the ESR1 cistrome. H3K27ac HiChIP was conducted in primary endometrial stromal cells treated with vehicle or decidualization cocktail to identify hormonally induced chromatin interaction changes. Results Among six tested guide RNAs, the ESR1-3 gRNA induced robust ESR1 activation, which restored E2 responsiveness in THESC. Bulk RNA-seq revealed ESR1-mediated E2-dependent and E2-independent transcriptional programs, regulating pathways involved in inflammation, proliferation, extracellular matrix organization, and cancer. Notably, 72% of differentially expressed genes (DEGs) overlapped with genes active in human endometrial tissue during the proliferative E2 dominant phase, supporting their physiological relevance. Cut&Run-seq identified genome-wide ESR1 binding sites, with most located at distal regulatory elements. To associate distal ESR1 binding sites with genes, we integrated H3K27ac HiChIP chromatin loops in hESC to identify distal ESR1 binding sites that loop to gene promoters. We identified genes regulated by ESR1/E2 through long-range chromatin looping that are involved in stromal cell decidualization, including FOXO1 and IL6R. Additionally, we identified genes implicated in endometrial cancer, including ERRFI1, NRIP1, and EPAS1, suggesting a role for stromal ESR1 driven endometrial pathologies. Functional assays confirmed that ESR1 promotes cell viability and, in the presence of E2, enhances migration. Conclusions ESR1 activation through CRISPR restores E2 responsiveness in endometrial stromal cells. Changes to chromatin architecture support gene expression changes that drive decidualization. Integration of ESR1/E2 transcriptome and cistrome with HiChIP data identifies its role in regulating inflammation, proliferation, and decidualization, as well as its implications in endometrial cancer. This model serves as a powerful tool to study estrogen signaling in endometrial stromal biology and related pathologies.

SUMMARY Antibody responses are determined by activated B cells bifurcating into plasmablasts (PBs) and germinal center B cells (GCBCs). Gene regulatory networks (GRNs) underlying human B cell fate choice remain uncharted. Temporally resolved single-cell multi-omics, computational modeling and CRISPR-based perturbations were used to assemble, simulate and test high-resolution GRNs underlying PB and GC fates. The results converged with orthogonal predictions of transcription factor (TF) action at single-nucleotide resolution, revealing dominant and reciprocal actions of IRF4 and its binding partners at simple and composite IRF motifs. Single-cell perturbation analysis of these TFs demonstrated multiple reciprocal negative feedback loops controlling the bifurcation. Additionally, IRF4 and BLIMP1, co-repressed the cell cycle regulators MYC and CCND2 . G0/G1 lengthening accelerated the switching of cells to an IRF4 hi BLIMP1 hi regulatory state and enhanced the probability of PB specification, thereby uncovering a self-reinforcing regulatory module that couples cell cycle dynamics to B cell fate choice.

Abstract MDA5 is an innate immune RNA sensor that senses infection with a range of viruses and other pathogens. MDA5’s RNA agonists are not well defined. We used single-nucleotide resolution crosslinking and immunoprecipitation (iCLIP) to study its ligands. Surprisingly, upon infection with SARS-CoV-2 or encephalomyocarditis virus, MDA5 bound overwhelmingly to cellular RNAs. Many binding sites were intronic and proximal to Alu elements and to potentially base-paired structures. Concomitantly, cytoplasmic levels of intron-containing unspliced transcripts increased in infected cells and displayed enrichment of MDA5 iCLIP peaks. Moreover, overexpression of a splicing factor abrogated MDA5 activation. Finally, when depleted of viral sequences, RNA extracted from infected cells still stimulated MDA5. Taken together, MDA5 surveys RNA processing fidelity and detects infections by sensing perturbations of posttranscriptional events such as splicing, establishing a paradigm of innate immune ‘guarding’ for RNA sensors.

Human midbrain organoids (hMOs) show promise as a patient-derived model for the study of Parkinson's disease (PD). Yet, much remains unknown about how accurately hMOs recapitulate key features of PD in the human brain. In both PD patients and animal models, disease progression leads to characteristic changes in neural activity throughout the basal ganglia. Here we demonstrate that patient-derived induced pluripotent stem cell (iPSC) hMOs harboring a triplication in the SNCA gene, encoding α-synuclein, a key protein in PD pathogenesis, can recapitulate PD-associated changes in neural activity. Namely, we observe hyperactive network activity in SNCA triplication hMOs, but not in isogenic, CRISPR-corrected iPSC hMOs. These changes are characterized by an increase in the number of bursts and network-wide bursts. Moreover, SNCA triplication hMOs exhibit an increase in network synchrony and burst/network burst strength similar to observations in animal and human PD brains. Subsequently, we show that the observed changes in neuronal activity are attributed to dopamine D2 receptor hypoactivity due to dopamine depletion, which could be reversed by the D2 receptor agonist quinpirole. Thus, hMOs faithfully model network wide electrophysiological changes associated with PD progression and serve as a promising tool for PD research and personalized medicine.

Transcription initiation is a critical regulatory step in plant gene expression, yet its sequence determinants remain largely elusive. Here we introduce GenoRetriever , an interpretable deep learning model that deciphers the sequence basis of transcriptional initiation regulation across plant genomes. Trained on STRIPE-seq data from 16 soybean tissues and six other crop species, GenoRetriever identifies 27 core sequence motifs that govern transcription start site (TSS) selection and usage. The model predicts TSS locations and usage levels with high accuracy, as validated by in silico motif insertions, saturation mutagenesis, and CRISPR-Cas9 promoter editing. It further reveals that 31.85% of natural variation between wild and domesticated soybean drives shifts in promoter motif usage during domestication, and uncovers lineage-specific motif effects between monocots and dicots. This interpretable model and its user-friendly web server for promoter analysis and design make GenoRetriever both a methodological innovation and practical tool for plant functional genomics and crop improvement.

The identification of therapeutic protein targets is fundamental to the success of drug development and repurposing. Traditional approaches for target selection require extensive preclinical evaluation for toxicity and efficacy, making the process time-intensive and resource-heavy. Computational tools that efficiently prioritize and validate novel targets are needed to streamline drug discovery workflows. To address this gap, we developed TARRAGON: T herapeutic T arget A pplicability R anking and R etrieval- A ugmented G eneration O ver N etworks, a computational framework that integrates data mining and machine learning to identify, rank, and assess target-disease relationships to nominate new therapeutic targets. TARRAGON mines knowledge graphs to uncover meta-paths, or rules of graph traversal, linking potential therapeutic targets to diseases. It employs a classification model to rank target-disease hypotheses based on evidence patterns and utilizes a retrieval-augmented generation workflow to prompt a large language model for generating feasibility reports on prioritized targets. Using TARRAGON, we prioritized potential drug targets for non-muscle invasive urinary bladder cancer. Top-ranked candidates were validated using CRISPR gene effect and expression data from the Broad Institute DepMap portal. We further proposed chemical modulators for these targets to inform combination drug screening alongside approved bladder cancer therapeutics. TARRAGON introduces a novel, interpretable computational pipeline for therapeutic target discovery and pharmaceutical candidate nomination, offering the potential to accelerate drug development across diverse disease areas.

Recent approvals of PARP inhibitors (PARPi) for BRCA-mutant metastatic castration resistant prostate cancer (mCRPC) necessitate an understanding of the factors that shape sensitivity and resistance. Reversion mutations that restore homologous recombination (HR) repair are detected in ∼50-80% of BRCA-mutant patients who respond but subsequently relapse, but there is currently little insight into why only ∼50% of BRCA-mutant patients display upfront resistance. To address this question, we performed a genome-wide CRISPR screen to identify genomic determinants of PARPi resistance in murine Brca2 Δ / Δ prostate organoids genetically engineered in a manner that precludes the development of reversion mutations. Remarkably, we recovered multiple independent sgRNAs targeting three different members ( Cdt1, Cdc6, Dbf4 ) of the DNA pre-replication complex (pre-RC), each of which independently conferred resistance to olaparib and the next generation PARP-1 selective inhibitor AZD5305. Moreover, sensitivity to PARP inhibition was restored in Brca2 Δ / Δ , Cdc6-depleted prostate cells by knockdown of geminin, a negative regulator of Cdt1, further implicating the critical role of a functional pre-RC complex in PARPi sensitivity. Furthermore, ∼50% of CRPC tumors have copy number loss of pre-RC complex genes, particularly CDT1 . Mechanistically, prostate cells with impaired pre-RC activity displayed rapid resolution of olaparib-induced DNA damage as well as protection from replication fork degradation caused by Brca2 loss, providing insight into how Brca2-mutant cancer cells can escape cell death from replication stress induced by PARP inhibition in the absence of HR repair. Of note, a pharmacologic inhibitor that targets the CDT1/geminin complex (AF615) restored sensitivity to AZD5305, providing a potential translational avenue to enhance sensitivity to PARP inhibition in BRCA-mutant cancers. Significance Here, we address a major limitation in the effectiveness of PARP inhibitors in BRCA-mutant prostate cancer treatment: only ∼50% of patients respond despite clear genomic evidence of defective homologous recombination. Prior efforts to study PARP inhibitor resistance in prostate cancer have been plagued by the lack of suitable cell lines. We overcame this challenge using primary prostate organoids coupled with genome-wide CRISPR screening. The key finding is that loss of function mutations in the DNA pre-replication complex confer PARP inhibitor resistance. These genes map to chromosomal regions frequently lost in prostate cancer and could therefore serve as potential biomarkers of treatment response. Pharmacologic inhibition of geminin, a negative regulator of the pre-replication complex, can restore PARP inhibitor sensitivity.

ABSTRACT Colorectal cancer (CRC) is the second leading cause of cancer deaths in the United States, with a five-year survival rate of 65%. Oxaliplatin was the first platinum drug shown to improve CRC patient outcomes and is now a common adjuvant therapy for advanced disease, yet 90% of patients develop resistance. Oxaliplatin was developed as a third-generation derivative of cisplatin, but recent evidence points to divergent modes of action. Here, genome-wide CRISPR activation and knockout screens were conducted to identify genetic changes that confer resistance to oxaliplatin in two CRC cell lines with distinct molecular backgrounds (SW620 and RKO). Guide RNAs corresponding to the neutral amino acid transporter SLC43A1 (LAT3) were the most significantly enriched in knockout screens and depleted in activation screens, suggesting a potential role for LAT3 in modulating oxaliplatin resistance. In vitro CRISPR knockout and overexpression of LAT3 in SW620 and RKO cell lines confirm increased resistance or sensitivity to oxaliplatin, respectively. Further analysis demonstrates that increased LAT3 levels corrrelate with increased intracellular levels of oxaliplatin, increased levels of DNA-platinum adducts and DNA damage, demonstrating that enhanced LAT3-mediated uptake of oxaliplatin can exert its expected mechanism of action and induce cytotoxicity. These findings may lead to a better understanding of oxaliplatin’s mode of action in CRC and can provide new insights into the interplay between essential nutrient uptake and drug transport. STATEMENT OF SIGNIFICANCE Oxaliplatin resistance remains a major clinical challenge for CRC treatment. Our study identified a novel role for LAT3 as a modulator of oxaliplatin sensitivity, offering new insights into drug resistance mechanisms.

Abstract Selective RNA degradation during terminal erythropoiesis results in a globin-rich transcriptome in mature erythrocytes, but the specific RNA decay pathways remain unknown. We found that deficiency of the terminal uridylyl transferase enzyme Zcchc6 and the 3'-5' exoribonuclease Dis3l2 in mouse models led to fetal and perinatal reticulocytosis, an accumulation of RNA-rich precursors of terminal erythroid cells, suggesting their crucial roles in terminal red cell differentiation. Notably, knockout embryos exhibited persistent high-level expression of Hbb-bh1 globin, the ortholog of human fetal γ-globin. Perturbation of the Zcchc6-Dis3l2 pathway in mice engineered to express the human β-globin locus likewise increased γ-globin levels in fetal erythroid cells, suggesting that globin switching entails post-transcriptional mechanisms of mRNA destabilization in addition to transcriptional down-regulation. We cultured human hematopoietic stem and progenitor cells (HSPCs), performed CRISPR/Cas9-mediated knockout of ZCCHC6 and DIS3L2, and observed accumulation of RNA and elevated γ-globin levels in terminal erythroid cells. Our findings reveal a conserved role for the ZCCHC6/DIS3L2 RNA editors in terminal erythropoiesis and demonstrate a post-transcriptional mechanism for γ-globin gene switching, advancing research into in vitro erythrocyte generation and γ-globin stabilization to ameliorate hemoglobinopathies.