Biology subjects

Cheng, H.

Publications and source records attributed to Cheng, H..

At least 37 records · Page 2Linked to original sources

bioRxiv2026

miRXplain: explainable isomiR-aware microRNA target prediction using CLIP-L experiments and hybrid attention transformers

MicroRNAs (miRNAs) are ~22 nt long noncoding RNAs that repress genes by base-pairing with complementary sequences in target mRNAs. Variants known as isomiRs arise from alternative hairpin processing and often shift the canonical seed region, thereby altering their target repertoire. Yet, the mRNA features that enable differential miRNA and isomiR target selection remain poorly understood, largely due to limited high-throughput assays that capture the exact miRNA bound to each target. Existing deep learning methods for miRNA-target prediction have neither leveraged such datasets nor investigated isomiR-specific interactions. To fill this gap, we developed miRXplain, an isomiR-aware transformer that predicts miRNA-mRNA interactions using miRNA and target sequences derived from CLIP-L data, directly linking precise miRNA variants to their target sites. These data however revealed a 5'-end nucleotide bias in target sites, which we corrected to generate high-quality, miRNA-target pairs for training, preserving isomiR-specific signal. miRXplain outperformed all benchmarked models, surpassing TEC-miTarget, in auROC and auPRC with 15x fewer parameters. Attention maps highlighted distinct sequence determinants for canonical versus isomiR interactions, and in silico saturation mutagenesis highlighted the importance of seed and 3' supplementary regions. miRXplain effectively prioritizes pathogenic single-nucleotide variants affecting mRNA-miRNA interactions and reveals isomiR targeting principles that deepen our understanding of miRNA biology.

bioinformatics↗

bioRxiv2026

ALK R1275Q mutation drives expansion of SCP-like cells during sympathoadrenal commitment and primes neuroblastoma initiation.

Neuroblastoma (NB) is a pediatric malignancy developing in the sympathoadrenal lineage of the neural crest, characterized by clinical heterogeneity ranging from spontaneous regression to poor outcomes. Activating mutations in the receptor tyrosine kinase anaplastic lymphoma kinase (ALK) are frequently observed in both sporadic and familial NB, yet the functional role of ALK in tumor initiation is not fully understood. Using a patient-derived human induced pluripotent stem cell (iPSC) model of sympathoadrenal development, we show that upon sympathoadrenal lineage commitment, ALK R1275Q, the most common hotspot mutation found in familial NB, sustain a proliferative, immature Schwann cell precursor (SCP)-like cell state with elevated ALK signaling and increased susceptibility to MYCN-driven transformation. While ALK-mutant cells alone did not form tumors in vivo, they cooperated with MYCN to accelerate tumor initiation, suggesting that ALK R1275Q creates a permissive but insufficient state for transformation. These findings define an ALK-driven cell progenitor-like state that facilitates the initiation of NB during embryonal development.

cancer biology↗

bioRxiv2026

Targeting PI3Kβ-dependent cancer with a novel small molecule inhibitor, GT220

PI3K{beta} is a critical oncogenic driver in cancers harboring PTEN loss or PIK3CB alterations, yet effective and selective PI3K{beta}-targeted therapies remain elusive. Here, we report the development and preclinical characterization of GT220, a highly selective and potent small-molecule PI3K{beta} inhibitor developed through integrated artificial-intelligence-driven design with medicinal chemistry and pharmacologic optimization. GT220 exhibits exceptional biochemical selectivity for PI3K{beta}, binding with sub-nanomolar affinity and with minimal activity against other class I PI3K isoforms or the broader protein kinome. In cellular models, GT220 potently suppresses AKT phosphorylation and selectively inhibits viability of PTEN-deficient cancer cells, while sparing PTEN/PIK3CB wild-type and PI3K-dependent cells. In vivo, GT220 achieves a favorable tumor exposure with sustained PI3K{beta} pathway inhibition and demonstrates robust antitumor efficacy and good tolerability in PTEN-deficient breast cancer xenograft models. In contrast, GT220 shows no antitumor activity or pathway inhibition in PTEN-wild-type or PI3K-dependent tumors, underscoring its context-dependent mechanism of action. Collectively, these findings establish GT220 as a promising next-generation PI3K{beta} inhibitor and provide a strong preclinical rationale for precision targeting of PI3K{beta}-dependent cancers.

cancer biology↗

bioRxiv2026

Microglia stabilize sleep homeostasis via adenosine A3 receptor signaling

Sleep homeostasis maintains the sleep-wake balance through sleep pressure, a process orchestrated by the accumulation of extracellular adenosine (eADO). Microglial Ca{superscript 2} activity has been implicated in sleep regulation, but the mechanism whereby microglia sense sleep pressure remains unclear. Here we show that microglia regulate sleep homeostasis through brain state-dependent Ca2+ activity driven by adenosine A3 receptor (A3R) signaling. Using miniaturized two-photon microscopy (mTPM) in freely behaving mice, we demonstrate that microglial Ca{superscript 2} activity is rapidly altered by brain-state transitions. Pharmacological experiments reveal that microglial Ca2+ dynamics are predominantly mediated by A3R in response to brain state-dependent eADO oscillations. Microglia-specific deletion of A3R attenuates these state-dependent Ca2+ dynamics, impairs microglial morphological plasticity across sleep-wake cycles, and leads to sleep fragmentation by increasing transitions between wakefulness and non-rapid eye movement (NREM) sleep. Together, these findings establish that microglia regulate sleep homeostasis by stabilizing both wakefulness and NREM sleep, a process critically dependent on eADO-A3R signaling.

neuroscience↗

bioRxiv2026

Auditory brainstem-cortical anatomy constrains the magnitude of frequency-following responses (FFRs) and event-related potentials (ERPs) coding speech-in-noise

Speech-evoked brain potentials provide a window into the neural encoding of speech, experience-dependent plasticity, and deficits in central auditory processing from communication disorders. Stronger and faster frequency-following responses (FFRs) and cortical event-related potentials (ERPs) have been interpreted as reflecting more robust and efficient auditory-sensory processing across brainstem and cortical levels. Importantly, these neural signatures relate to real-world listening skills like speech-in-noise (SIN) perception. Yet, how these speech-evoked FFRs/ERPs relate to underlying auditory anatomical structures is unknown. Using a multimodal imaging approach, we recorded FFRs and ERPs to clean and noise-degraded speech sounds to assess the strength of listeners neural encoding of speech at brainstem (FFR) and cortical (ERP) levels. MRI volumetrics of midbrain and transverse temporal gyrus (Heschls gyrus) quantified morphological variation in subcortical and cortical anatomy that underly these EEG potentials. The QuickSIN assessed behavioral SIN abilities. Results showed that larger and thicker right (but not left) Heschls gyrus was related to listeners SIN abilities as well as the size of their cortical ERPs. Structural and functional measures interacted at a subcortical level. For listeners with smaller midbrain volumes, larger speech FFRs were associated with better QuickSIN scores; whereas in individuals with larger midbrain volumes, larger FFRs were related to poorer QuickSIN. Our findings reveal common functional signatures of speech processing (FFRs, ERPs) are constrained by the anatomy of their underlying generators and suggest a complex interplay between auditory brain structure and function in accounting for perceptual SIN capacity.

neuroscience↗

bioRxiv2025

Complete genomes of a multi-generational pedigree to expand studies of genetic and epigenetic inheritance

Pedigree analysis remains the gold standard for rare disease diagnostics, yet whole genome sequencing studies typically omit critical regions like centromeres, telomeres, and acrocentric chromosome p-arms. Here, we present telomere-to-telomere (T2T) reference genomes for four self-identified African American individuals of admixed ancestry spanning three generations. Our parent-of-origin assigned, chromosome-level assemblies revealed precise meiotic recombination breakpoints in previously inaccessible regions, including recombination events across acrocentric and subtelomeric sequences. Centromeric regions were highly stable, with multi-megabase arrays inherited intact across three generations, while the position of kinetochore assembly sites remained consistent and predominantly associated with the p-arm proximal region. The relative lengths of telomeres on individual chromosomes were maintained across generations. Using a targeted rDNA assembly approach, we reconstructed a complete megabase-scale ribosomal DNA (rDNA) array corresponding to the paternal chromosome 14. This openly available pedigree provides a benchmark dataset for studying recombination and genetic and epigenetic variation across the complete genome.

genomics↗

bioRxiv2025

Predicting and Designing Red Fluorescent Protein Variants Using Sequence-to-Function Machine Learning Models

Fluorescent proteins (FPs) are widely used reporters for visualizing cellular structures and processes. Traditional wet-lab strategies for FP engineering (rational design and directed evolution) have enabled substantial improvements in photophysical performance but are limited by their requirement for deep expert knowledge or labor-intensive screening. AI-driven approaches have recently gained traction for engineering variants of green FPs, yet applications to red fluorescent proteins (RFPs) remain scarce. Here, we applied machine learning models to an RFP sequence-function dataset and trained these models to predict functional single-mutation variants of the state-of-the-art RFP mScarlet-I3. Guided by model predictions, we identified variants exhibiting red-shifted emission peaks, large Stokes shifts, or brightness comparable to the parental protein. Our findings show that even lightweight, data-efficient models can extract actionable design principles for improving RFPs. This work demonstrates the feasibility of AI-guided design for RFPs and provides a reliable benchmark for future development of more powerful AI-driven strategies for FP engineering.

molecular biology↗

bioRxiv2025

JADE: Joint Alignment and Deep Embedding for Multi-Slice Spatial Transcriptomics

As spatially resolved transcriptomics (SRT) datasets increasingly span multiple adjacent or replicated slices, effective joint analysis across slices is needed to reconstruct tissue structures and identify consistent spatial gene expression patterns. This requires resolving spatial correspondences between slices while capturing shared transcriptomic features, two tasks that are typically addressed in isolation. Multi-slice analysis remains challenging due to physical distortions, technical variability, and batch effects. To address these challenges, we introduce Joint Alignment and Deep Embedding for multi-slice SRT (JADE), a unified computational framework that simultaneously learns spatial location-wise alignments and shared low-dimensional embeddings across tissue slices. Unlike existing methods, JADE adopts a roundtrip framework in which each iteration alternates between alignment and embedding refinement. To infer alignment, we employ attention mechanisms that dynamically assess and weight the importance of different embedding dimensions, allowing the model to focus on the most alignment-relevant features while suppressing noise. To the best of our knowledge, JADE is the first method that jointly optimizes alignment and representation learning in a shared latent space, enabling robust multi-slice integration. We demonstrate that JADE outperforms existing alignment and embedding methods across multiple evaluation metrics in the 10x Visium human dorsolateral prefrontal cortex (DLPFC) and Stereo-seq axolotl brain datasets. By bridging spatial alignment and feature integration, JADE provides a scalable and accurate solution for cross-slice analysis of SRT data.

bioinformatics↗

bioRxiv2025

A BCG toxic effector induces mitochondrial DNA compaction to refrain protective immunity

Innate immune memory typically emphasizes nuclear chromatin remodeling that determines transcriptional memory in monocytes and macrophages. Here, we show that compaction of mitochondrial DNA (mtDNA), governed by mitochondrial transcription factor A (TFAM) abundance and mtATP-dependent remodeling, defines the transcriptional and functional state of trained macrophages. While typical pathogen-associated molecular pattern (PAMPs) triggered mtDNA accessibility, Bacillus Calmette-Guerin (BCG) effector PE18 directly interacted with and activated SLC25A5 to reduce mtATP levels, thereby inhibiting TFAM degradation by AFG3L2. Accumulated TFAM restricts mtDNA accessibility, suppressing the transcription of the mtNd1 and thereby reducing mtNd1-mediated mtROS production, which diminishes the trained capacity of macrophages against secondary infections. Notably, BCG{Delta}pe18 maintains macrophages cytokine transcriptional capacity independent of classical nuclear histone modifications. Moreover, BCG{Delta}pe18 provides robust and lifelong protective immunity against Mycobacterium tuberculosis (Mtb) infection. These findings reveal mitochondrial genome as a regulatory entity in trained immunity, and suggest a potential strategy for improving vaccine performance.

immunology↗

bioRxiv2025

Integrating Longitudinal Metabolite Profiles Improves Trait Prediction in Pigs in a Trait- and Timepoint-Dependent Manner

BackgroundAccurate prediction of genetic merit is essential for accelerating genetic improvement in pigs, particularly for traits that are costly or difficult to measure directly. This study investigated the potential of integrating individual-level blood serum metabolite profiles sampled at two developmental stages (10-week and 20-week) into genomic prediction models for five economically important traits: average daily feed intake (DFI), feed conversion ratio (FCR), backfat thickness (BF), test daily gain (TDG), and loin depth (LD). Using a BayesC modeling framework, we analyzed 1,637 pigs from a single purebred population with complete phenotype, genotype, and metabolite-profile data. We evaluated seven models, including a genotype-only model (G), metabolite-only models using metabolite-profile data from 10-week (M1), 20-week (M2), or both timepoints (M1+M2), and combined models integrating genotypes with metabolite-profile data (G+M1, G+M2, G+M1+M2). ResultsHeritability estimates were generally low to moderate, ranging from 0.07 to 0.30 for the 10-week metabolite profile and from 0.04 to 0.28 for the 20-week metabolite profile. Prediction accuracy of phenotype consistently improved when metabolite-profile data were integrated into genotype-based models, although the magnitude of improvement varied depending on the trait and the timepoint of metabolite sampling. Prediction accuracy increased from 0.31 (G) to 0.41 (G+M2; G+M1+M2) for DFI, and from 0.27 (G) to 0.33 (G+M2; G+M1+M2) for FCR. The latter two models also delivered the largest gains over G for BF (from 0.41 to 0.45) and TDG (from 0.28 to 0.32). However, LD benefited the most when both 10-week and 20-week metabolite profiles were combined (G+M1+M2: 0.45 compared to G: 0.42). ConclusionsAcross all traits, models combining genotype data with metabolite profiles from one or multiple timepoints achieved the highest or equally high prediction accuracies compared to the genotype-only model, reflecting complementary biological insights captured by metabolite profiles. These findings highlight the potential value of metabolite-profile data as an intermediate omics layer to enhance genomic prediction, particularly when integration strategies are tailored to trait-specific biology and sampling timepoints.

genetics↗

bioRxiv2025

Cryptococcus neoformans employs both PP1 and PP2A phosphatases to silence the spindle checkpoint

The spindle checkpoint preserves genomic integrity by delaying anaphase onset until all kinetochores achieve bi-polar microtubule attachment. While checkpoint activation is well-characterised across eukaryotes, the mechanisms governing its silencing remain less defined and are more variable in different systems. In this study, we reveal that the fungal pathogen Cryptococcus neoformans employs a dual phosphatase-mediated mechanism to silence the spindle checkpoint. We show that PP1 is recruited to kinetochores by Spc105KNL1, via conserved SILK and RVSF motifs, while PP2A-B56 is recruited to kinetochores through direct Bub1 binding via an LxxIxE motif. Disruption of either recruitment pathway leads to prolonged checkpoint signalling and mitotic defects, underscoring the critical roles of these phosphatases in checkpoint silencing and mitotic exit. These findings further establish C. neoformans as a powerful model for studying mitotic regulation in the basidiomycete lineage and identify phosphatase recruitment interfaces as potential targets for antifungal intervention.

cell biology↗

bioRxiv2025

End-to-end genomic prediction: direct prediction of images and text from genome-wide molecular markers

BackgroundMethods to predict the heritable component of phenotypes from genetic markers, collectively known as genomic prediction, have been widely applied in plant and animal breeding and in human genetics. Currently, genomic prediction is limited to numeric phenotypes. In some cases, though, plant and animal phenotypes are better understood through images and text rather than numbers. The current best practice for incorporating images and text in genomic prediction is to first extract scalar numeric phenotypes from images and text, and then to perform genomic prediction on the numeric phenotypes. While this approach is effective for some traits, it involves discarding most of the information in the image or text, including potentially useful information. Additionally, numeric phenotypes derived from images and text may not be as interpretable as the images and text themselves. ApproachWe present a novel approach for predicting images and text from SNP markers, which we refer to as end-to-end genomic prediction, and validate this approach using genotypes, text, and image phenotypes derived from a strawberry (Fragaria x ananassa) diversity panel. Our approach combines nonlinear latent space encoding with linear genomic prediction to generate accurate breeding values for a high-dimensional phenotype, for example, images or text. ResultsFor both genome-to-image and genome-to-text prediction, we found that predicting images and text and then extracting numeric traits from them was in some cases as accurate as directly predicting extracted numeric phenotypes, demonstrating for the first time that genome-to-image prediction accuracy can be comparable to conventional genomic prediction accuracy. ConclusionsBased on our proof-of-concept using the same core end-to-end method in both images and text, we believe end-to-end genomic prediction could be of use in a wide range of visual and multidimensional phenotypes in plants and animals, although further work to improve the accuracy of the embedding and genomic prediction steps is needed.

genomics↗

bioRxiv2025

Oncogenic and immunomodulatory functions of SUV420H1 in HPV-negative head and neck squamous cell carcinoma.

Despite the advent of immunotherapy, human-papilloma-virus (HPV)-negative head and neck squamous cell carcinoma (HNSCC) carries a high morbidity and mortality rate, thus novel therapies are urgently needed. Suppressor Of Variegation 4-20 Homolog 1 (SUV420H1) is a protein lysine methyltransferase that writes H4K20me3. Approximately 35% of HPV-negative HNSCC tumors carry gains/amplifications of SUV420H1. Gene Set Enrichment Analysis (GSEA) showed enrichment of proliferation-, epithelial-mesenchymal transition (EMT)- and immune-response pathways in SUV420H1-overexpressing HPV-negative HNSCC tumors. Depletion of SUV420H1 led to decreased proliferation, cell cycling and invasion in human HPV-negative HNSCC cell lines, while enzymatic inhibition decreased the invasive capacity but not the proliferation and cell cycling of HPV-negative HNSCC cell lines, supporting the presence of catalytically-independent and -dependent functions of SUV420H1. In a syngeneic mouse model of mouse oral carcinoma 1 tumors (MOC1), Suv420h1 knockout (KO) in MOC1 cancer cells halted tumor growth and synergized with anti-PD-1 therapy. In the tumor immune microenvironment (TIME) of Suv420h1 KO MOC1 tumors, a significant increase in the macrophage compartment with a concurrent decrease in the protumorigenic granulocytic myeloid derived suppressor cells (gMDSCs) was observed. Genome-wide mapping of SUV420H1-mediated H4K20me3 revealed enrichment of EMT-, IFN-response, and myeloid-attracting chemokines. This work provides rationale for the depletion/degradation of SUV420H1 as a novel therapeutic strategy to hinder proliferation and invasion, and to impart sensitization to anti-PD-1 immunotherapy for patients with HPV-negative HNSCC tumors with SUV420H1 gain/amplification.

cancer biology↗

bioRxiv2025

The complete genome of a songbird

Bird genomes are the smallest among amniotes, but remain challenging to assemble due to their structural complexity. This study presents the first fully phased, diploid, telomere-to-telomere (T2T) reference genome for the zebra finch (Taeniopygia guttata), a model organism for neuroscience and evolutionary genomics. Combining multiple sequencing strategies resulted in closing nearly all gaps, adding [~]90 Mbp of previously missing sequence (7.8%). This includes T2T assemblies for all microchromosomes, including dot chromosomes, and the previously almost entirely missing chr16. The T2T genome is comprehensively annotated for genes, repeats, structural variants, and long-read methylation calls. Complete centromeric structures were assembled and annotated along with kinetochore binding sites. Relative to the previous high-quality reference of the Vertebrate Genomes Project, 2,778 (8.51%) previously unassembled or unannotated genes were identified, of which 9% overlap with segmental duplications. This first complete genome of a songbird, now the new public reference, illuminates avian genome architecture and function.

genomics↗

bioRxiv2025

Network dynamics for sensory prioritization: Functional connectivity related to individual sensory weighting of vision versus proprioception during upper limb control

Precise control of the hand requires the dynamic integration of visual and proprioceptive (body position sense) sensory cues with internal models, task goals, and motor plans. Individual differences in how visual and proprioceptive cues are weighted have been related to neural substrates, such as posterior parietal regions, yet the underlying neural dynamics are unclear. This study investigated the relationship between visuo-proprioceptive perception during a bimanual pointing task and the whole-brain network dynamics using resting-state functional magnetic resonance imaging (rsfMRI) with both atlas-based and individually-localized ROI seeds. Our results confirm the existence of individual sensory biases, and find they have systematic influences on functional connectivity. Activity between sensorimotor regions and default mode network (DMN) nodes was related to sensory weighting (e.g. individual degree of reliance on vision versus proprioception) and may reflect updating of internal models. The ventral premotor cortex (PMv) emerged as an important node with functional connections suggesting its role for integrating motor plans, internal models, and sensory percepts. Connections from multisensory integration regions like the middle temporal gyrus (MTG) and the superior parietal lobule (SPL), and motor coordination regions in the cerebellum, were related to increased reliance on visual versus proprioceptive information. These findings suggest that individual sensory biases during sensorimotor behavior may be characterized by specific and specialized patterns of neural activity. This research establishes a foundational exploration of the neural systems underlying sensorimotor processing, supporting theories suggesting that learning in either the sensory or motor system may also cause plasticity in the other.

neuroscience↗

bioRxiv2025

NAD⁺ Replenishment Mitigates Cardiomyocyte Senescence and Corrects Heart Failure with Preserved Ejection Fraction in Aged Mice

Cardiomyocyte senescence, characterized by elevated cell cycle inhibitor expression, persistent DNA damage response, and mitochondrial dysfunction, contributes to myocardial stiffness and the progression of heart failure with preserved ejection fraction (HFpEF), the most common form of heart failure affecting individuals over 65. In this study, we investigated the role of NAD metabolism in cardiomyocyte senescence and cardiac function. Aged mice exhibited reduced cardiac NAD levels, impaired NAD biosynthesis and mobilization, and increased consumption, leading to suppressed SIRT1/6 activity and accumulation of senescent cardiomyocytes. This was accompanied by diastolic dysfunction consistent with HFpEF. In senescent AC16 cardiomyocytes, NAD depletion promoted senescence, which was reversed by the NAD precursors nicotinamide riboside (NR) and dihydronicotinamide riboside (NRH). In aged mice, two months of NR or NRH treatment improved diastolic function and reduced cardiomyocyte senescence. While NR primarily activated SIRT1 to suppress cell cycle arrest markers, NRH more robustly activated both SIRT1 and SIRT6, enhancing DNA damage repair. Acetylated H2AX, a SIRT6 substrate elevated in aged hearts and senescent cells, was selectively deacetylated by NRH. These findings identify NAD availability as a critical regulator of cardiac senescence and support NAD precursors, particularly NRH, as promising senescence-reducing therapies for treating aging-associated HFpEF.

pharmacology and toxicology↗

bioRxiv2025

An Intrinsic-hoc Framework for Heterogeneous Cellular Senescence Elucidation Using Deep Graph Representation Learning and Experimental Validation

Cellular senescence is a primordial driver of tissue and organ aging, and the accumulation of senescent cells (SnCs) has been implicated in numerous age-related diseases. A major barrier to studying senescence is the rarity and heterogeneity of SnCs, which are not a uniform population but instead comprise diverse senotypes shaped by cell-of-origin and microenvironmental context. Such heterogeneity exceeds what classical senescence hallmarks can resolve at single-cell resolution, motivating the need for computational frameworks that can capture senotype-level diversity intrinsically. Here, we introduce DeepSAS, a deep graph representation learning framework that robustly identifies cell-type-specific SnCs and their senescence-associated genes (SnGs). DeepSAS incorporates a heterogeneous graph that integrates intracellular transcriptional states with intercellular communication cues, enabling the joint inference of senescent cells and senescence-linked genes through attention-based contrastive learning. Applied to public healthy eye and lung atlases, DeepSAS identified SnCs whose proportions positively correlate with aging. From in-house idiopathic pulmonary fibrosis (IPF) patient scRNA-seq data, DeepSAS detected 1,678 SnCs (out of 24,125 cells) and 263 SnGs across 26 cell types, including 43 SnGs that are uniquely associated with a single cell type. We generated high-resolution Xenium spatial transcriptomics data to further validate SnGs in IPF, revealing NFE2L2 as a SnG specifically enriched in CTHRC1+ fibroblasts. Notably, the ex vivo bleomycin-induced senescence in human precision-cut lung slice (hPCLS) samples similarly identified NFE2L2 as an SnG in CTHRC1+ fibroblasts, albeit with stronger transcriptional signals, suggesting mechanistic differences in senescence cells associated with chronic and acute injury. Overall, DeepSAS uncovers distinct senescence programs and infers cell-type-specific SnGs that are difficult to resolve using existing marker-based approaches. We believe it offers a generalizable and translationally relevant strategy for advancing senescence biology and therapeutic development.

bioinformatics↗

bioRxiv2025

Prime assembly with linear DNA donors enables large genomic insertions

Targeted insertion of large DNA fragments has promising applications for genome engineering and gene therapy1,2. Twin prime editing (PE) guide RNAs (pegRNAs) have enabled relatively large insertions, but the efficiency remains low for insertions greater than 400 base pairs3-6. Here we describe a Prime Assembly (PA) approach for the insertion of large DNA donor fragments, whose ends are designed to overlap with the flaps generated by twinPE. We used PA to insert one, two, or three overlapping DNA fragments, with total insertion sizes ranging from 0.1 to 11 kilobase pairs. An inhibitor of non-homologous end joining (NHEJ) enhanced both the efficiency and precision of insertions. PA relies on DNA templates that are easily produced and does not require co-delivery of exogenous DNA-dependent DNA polymerases. Our study demonstrates that PA can initiate "Gibson-like" assembly in cells to generate gene insertions without double-stranded DNA breaks or recombinases.

biochemistry↗