Biology subjects
Publications and source records attributed to Su, W..
The authors have withdrawn this manuscript because a major revision is ongoing, and the timeline for a revised version is pending. Therefore, the authors do not wish this work to be cited as a reference for the project. If you have any questions, please contact the corresponding authors.
The Eurasian avian{square}like (EA) H1N1 swine influenza virus (SIV), derived from avian influenza viruses (AIV), poses a serious threat to public health due to its capacity for cross{square}species transmission and pandemic emergence. The molecular determinants underlying its replication advantage over AIV remain poorly defined. Here, we identify RNA{square}binding motif protein 6 (RBM6) as a novel host factor that differentially regulates the replication of EA H1N1 SIV and AIV. Mechanistically, RBM6 binds to the critical M901 site of the viral M segment, thereby modulating RNA splicing. Substitution of M901C with M901T markedly reduced RBM6 binding, impaired M segment splicing, and attenuated viral replication both in vitro and in vivo. Conservation analysis revealed that M901T is common in avian strains, whereas M901C is predominantly maintained in swine strains, underscoring M901C as a determinant of swine adaptation. Complementation experiments further demonstrated that swine RBM6, but not avian RBM6, restored EA H1N1 SIV replication. Taken together, our findings uncover a previously unrecognized role of RBM6 in shaping influenza virus replication and highlight the RBM6-M901C axis may serve as potential targets for controlling influenza virus adaptation and interspecies transmission.
Immediate-early protein IE1 is an essential protein of baculoviruses that is involved in transcription and replication. In the present study, we provide several lines of evidence for how ancient IE1 evolved into its current version. Using progressive truncations and site-directed mutations coupled with fluorescence microscopy, surprisingly, we showed that the previously identified nuclear localization sequence (NLS), basic domain II, was essential for sequence non-specific DNA binding, but not required for IE1 nuclear import, and demonstrated that, BmNPV IE1 (BmIE1) possesses not only a unique bipartite NLS that contains a monopartite NLS but also a cryptic non-canonical NLS that is correlated with sequence non-specific DNA binding. The non-canonical NLS alone was sufficient to launch infection. We also found that N-terminally truncated IE1 can enter the nucleus through its sequence non-specific binding ability. Remarkably, the monopartite NLS, when fused to the N-terminus of EGFP, can form a novel NLS that is also functional in a mammalian cell line. Moreover, residues 58 to 151 of BmIE1 were shown to be dispensable, and residue 152 was found to be critical for launching a productive infection. To gain insight into how ancient IE1 acquired its multi-functionality, we reorganized the N-terminal 23 aa and 132 aa of BmIE1 with EGFP in a variety of manners and found that both fragments are separable and transferable. Notably, we observed that the nuclear levels of BmIE1 should reach certain thresholds to initiate infection. Additionally, we found that BmNPV could launch infection more efficiently in a BmN cell line over another BmN cell line by increasing transcription levels of immediate early genes. Collectively, these findings suggest a hypothesis where ancient IE1 might have evolved the two additional NLSs and acquired the N-terminal 132 aa through gene fusion so as to reach infection-initiating thresholds at a faster pace. Author SummaryBasic domain II was previously shown to be the NLS of AcMNPV IE1 protein. In the present study, we demonstrated that it is essential for sequence non-specific DNA binding, but not required for IE1 nuclear import, and found that BmNPV IE1 (BmIE1) possesses not only a unique bipartite NLS that contains a monopartite NLS but also a cryptic non-canonical NLS that is correlated with sequence non-specific DNA binding. We also found that N-terminally truncated BmIE1 can enter the nucleus through its sequence non-specific binding ability. Remarkably, the monopartite NLS, when fused to the N-terminus of EGFP, can form a novel NLS that is also functional in a mammalian cell line. Moreover, the domains of BmIE1 was shown to be separable and transferable. We also demonstrated that higher expression of functionally impaired BmIE1 achieved by higher MOIs or higher transfection efficiency can partially complement its compromised functions. Consistently, we found that BmNPV could launch infection more efficiently in a BmN cell line over another BmN cell line by increasing transcription levels of immediate early genes.
Bifidobacteriaceae and Lactobacillaceae are key probiotic families and widely used in food production, yet a comprehensive understanding of strain functions and their gut microbial interactions based on complete genomes remain understudied. Here we constructed a complete-genome dataset of 3,300 strains from these two families, including 1,151 newly isolated from China. Compared with draft assemblies, complete genomes substantially recovered a gene functional landscape encompassing stress tolerance, surface exopolysaccharide synthesis, nutrient utilization, and mobile genetic elements. Major species from both families exhibited a prevalence >60% in the Chinese population, far higher than that in US/Dutch cohorts. Notably, as a core probiotic species with remarkable genomic plasticity and gut-adaptive potential, Lactiplantibacillus plantarum stood out in our dataset for its enriched functional profile and was particularly abundant in the Chinese population. Moreover, compared with non-Chinese genomes, our isolates of key species displayed less metabolic complementarity and stronger competition with potentially pathogenic keystone species in the gut, thereby linking strain origin to enhanced probiotic potential and ecological fitness to benefit human gut health.
Currently circulating swine influenza viruses (SIVs) mainly include H1N1, H1N2, and H3N2 subtypes. In this study, two G4 genotype Eurasian avian-like (EA) H1N1 SIVs were isolated from 556 samples collected between 2023 and 2026. A systematic analysis was conducted on the two EA H1N1 isolates (FYD30 and YZF69) to assess their pandemic potential. The hemagglutinin (HA) proteins of both H1N1 viruses possessed residues 225E and 228S, indicating enhanced affinity for human-like -2,6-linked sialic acid receptors, which was confirmed by receptor-binding assays. Polymerase activity tests demonstrated that the two SIVs exhibited significantly higher activity in mammalian cells, relative to avian cells, which is consistent with the efficient replication in mammalian cells. Challenge experiments revealed that both H1N1 caused significant pathogenicity in mice and pigs, with YZF69 exhibited higher virulence than FYD30. The higher virulence of YZF69 may be attributed to its molecular features, including the NP Q357K mutation, and an additional glycosylation site in HA. In conclusion, currently circulating EA H1N1 SIVs have acquired key molecular signatures of mammalian adaptation, exhibit enhanced virulence in mammals, and continue to undergo extensive reassortment driven by international swine trade. These findings highlight the potential pandemic risk of SIVs and underscore the urgent need for strengthened surveillance.
Severe hemolysis is a life-threatening condition with limited therapeutic options. Although haptoglobin and hemopexin sequester hemoglobin and heme, these protective systems are rapidly saturated during acute hemolysis, leading to the accumulation of cytotoxic free heme. In this study, we identify scavenger receptor BI (SR-BI) as a critical mediator of free heme clearance. SR-BI binds heme and facilitates its hepatic uptake under pathological conditions. Mice lacking hepatic SR-BI exhibit impaired heme clearance and increased susceptibility to heme- and hemolysis-induced lethality. Pharmacological upregulation of hepatic SR-BI via imatinib or adenoviral delivery confers protection against heme toxicity. Using a humanized model of sickle cell disease (SCD), we further demonstrate that sickle hepatopathy significantly reduces hepatic SR-BI expression compared to non-SCD littermates, potentially increasing vulnerability to heme-induced injury. Notably, adenoviral-mediated SR-BI upregulation rescues SCD mice from heme toxicity. These findings reveal a previously unrecognized mechanism of heme detoxification via hepatic SR-BI and identify a promising therapeutic target for hemolytic disorders. One-Sentence SummaryIdentification of scavenger receptor BI as a targetable scavenger of heme in hemolysis
Probe design for fluorescence in situ hybridization (FISH) underpins spatial transcriptomics, three-dimensional genome studies, and clinical diagnostics, yet remains constrained by two challenges: dependence on expert knowledge for parameter selection and quality evaluation, and the inability of existing tools to accommodate the diverse probe architectures introduced by rapidly emerging methods. Here we present U-Probe, a universal and agentic probe design platform. U-Probe employs a declarative configuration system with a directed acyclic graph (DAG)-based assembly engine that supports arbitrary probe structures from established protocols such as MERFISH, seqFISH, and MiP-seq to entirely novel architectures without code modifications. Integrated LLM-based AI agents enable conversational design workflows, allowing users to specify experimental goals in natural language and receive synthesis-ready probe sequences. We validated U-Probe in three scenarios: agent-driven MiP-seq panel design for influenza-infected mouse lung tissue, genome-tiling DNA-FISH for herpesvirus detection, and a novel RCA-based ligation probe for single-nucleotide mutation discrimination. U-Probe is available as an open-source tool with CLI, Web, and agent interfaces.
Artemisinin-based combination therapies (ACTs) remain the cornerstone of malaria treatment, but emerging resistance threatens their efficacy. The potential for the development of drug resistance against plasmepsin X (PMX)-selective inhibitors and dual plasmepsin IX/X (PMIX/X) inhibitors was investigated in Plasmodium falciparum. A series of PMX-selective (WM4, WM76, WM92) and PMIX/X dual inhibitors (WM382, WM09, WM42) were characterised for potency against parasite growth and enzyme inhibition. In vitro selection experiments showed that all compounds had a high barrier to resistance, although parasites with reduced sensitivity to PMX-selective inhibitors could still be selected. Resistance mechanisms involved pmx gene amplification and point mutations (D245N, S315P, S359P, I363L) that alter inhibitor binding. Recombinant expression and Michaelis-Menten kinetics demonstrated that these mutations impair drug binding whilst preserving PMX catalytic function. Reverse genetics confirmed that introducing these mutations into the pmx gene resulted in decreased potency of the inhibitors. In this study, resistance to the PMIX/X dual inhibitors evaluated here could not be selected, despite prolonged selection pressure. Antimalarial Resistome Barcoding (AReBar) assays confirmed the absence of pre-existing resistance to either inhibitor class. Critically, PMIX/X dual inhibitors maintained efficacy against parasites with decreased sensitivity to PMX-selective compounds. These findings demonstrate that dual PMIX/X inhibitors present a substantially higher barrier to resistance than PMX-selective inhibitors, informing antimalarial drug development strategies and highlighting dual-target inhibition as a promising approach to mitigate resistance risks.
Essential genes are defined as those that are indispensable for an organisms survival. The loss of function of these genes results in cell death or an inability to complete the normal life cycle. Research on essential genes is pivotal in elucidating the origin and evolution of life, as well as in identifying potential therapeutic targets. Therefore, accurately predicting essential genes is of great scientific importance and has many applications in basic research and the biomedical field. In this study, we propose EssTFNet, a novel and interpretable deep learning framework that combines adaptive time-frequency analysis with a DNA language model to achieve accurate prediction of human essential genes while enabling mechanistic biological interpretation. Specifically, EssTFNet leverages the architecture of ATFNet, which innovatively maps DNA and protein sequences into equivalent time-series signals to extract periodic and non-stationary features, thereby enhancing the models capacity to capture complex sequence patterns. Through effective feature selection and architectural optimization, EssTFNet strikes a favorable balance among prediction accuracy, model interpretability, and cross-tissues generalization capability, significantly outperforming current mainstream sequence-based deep learning methods with an AUC of 0.8336 and an AUPR of 0.8212. Additionally, the DeepLIFT attribution method was employed to identify functional motifs closely associated with gene essentiality, offering valuable insights for experimental validation. For the convenience of researchers, we have developed an easy-to-use web server and made it along with the source code in a GitHub repository: https://github.com/QIANJINYDX/EssTFNet. Overall, this study presents a potentially useful methodological framework for human essential genes prediction, which could provide valuable insights for future research and applications in this field.
The long-standing phenotypic trade-off between Upland cotton (Gossypium hirsutum, high yield) and Sea Island cotton (G. barbadense, high quality) represents a major bottleneck in cotton breeding and industrialization. Despite rapid advances in genomics, the genomic structural variations driving the divergence of these two domestication strategies, and their subsequent impact on downstream transcriptional networks, remain unclear. This study integrates multi-tissue transcriptomic atlases spanning the full growth period with pan-genomic analyses of both Upland and Sea Island cotton, proposing a "Reciprocal Selective Retention (RSR)" strategy adopted by the two species during evolution. Transcriptomic analysis revealed that while the transcriptional chassis of the two species is highly conserved, a drastic "phase inversion" occurs during fiber development. Sea Island cotton appears to trade off for quality by specifically activating a "delayed elongation" module (at 20 DPA), whereas Upland cotton initiates a "precocious filling" program to pursue yield. Further Weighted Gene Co-expression Network Analysis (WGCNA), combined with genomic Presence/Absence Variation (PAV) analysis, identified a set of key transcription factors exhibiting a "3-vs-3" reciprocal loss pattern. Upland cotton specifically retained CRF10, WIND1, and MYB93, constructing a genetic foundation of "robust root system and strong regeneration" to support high yield. Conversely, Sea Island cotton specifically retained MYB111 and ERF105/017 to maintain long-staple characteristics and environmental buffering. Whole-genome structural variation and microsynteny analyses confirmed that these differences stem from asymmetric physical deletions (e.g., a 5 kb deletion in MYB111) or fragment sequence collapse (e.g., a 15.9 kb sequence divergence in MYB93) on homologous chromosomes. The RSR model proposed in this study not only offers a novel insight into the genetic roots of the negative correlation between yield and quality but also provides clear genomic targets for precision improvement using gene-editing technologies such as CRISPR.
BackgroundsNext-generation sequencing (NGS) and droplet digital PCR (ddPCR) are both established methods for detecting EGFR mutations in non-small cell lung cancer (NSCLC). However, comprehensive validation of their concordance in mutation detection and variant allele frequency (VAF) quantification across heterogeneous sample types remains limited. Inconsistent results from different sample types (e.g., cfDNA, FFPE) pose a significant challenge to clinical decisions. This is particularly critical for advanced NSCLC patients, who often rely on liquid biopsy, yet the concordance between liquid and tissue-based testing lacks validation in large-scale studies. Another issue in clinical practice is that many tumor samples are limited in quantity, making it difficult to meet testing requirements. Therefore, it is worth exploring whether pre-capture NGS libraries can serve as substitutes for original DNA. MethodsIn this study, we first developed three assays to detect EGFR L858R, exon 19 deletions (Ex19del) and T790M, respectively using ddPCR platform. Their Limit of Detection (LOD) could reach 0.01% at 100 ng of input DNA. Subsequently, we conducted a large retrospective clinical study to systematically compare the detection performance of ddPCR and NGS across three mutation types using approximately 1,000 EGFR-positive samples, including cell-free DNA (cfDNA), pre-capture NGS libraries of cfDNA (cfDNA-prePCR), FFPE-derived DNA (ffpeDNA), pre-capture NGS libraries of FFPE-derived DNA (ffpeDNA-prePCR), fresh tumor tissue DNA (ttDNA), pre-capture NGS libraries of ttDNA (ttDNA-prePCR), pleural effusion supernatants DNA (peDNA), and pre-capture NGS libraries of peDNA (peDNA-prePCR). They were analyzed for detection concordance and VAF correlation. Especially, we made comparisons of the tumor DNA, including ctDNA and tumor tissue DNA, with their paired pre-capture NGS library. ResultsKey findings demonstrated excellent overall agreement between NGS and ddPCR. The mutation detection concordance rates were 98.72% (overall), with subtype-specific rates of 98.93% (L858R), 99.23% (Ex19del), and 97.14% (T790M). VAF measurements between ddPCR and NGS showed exceptional correlation (Pearsons r = 0.975, P<0.001). Notably, pre-capture NGS libraries showed remarkable VAF concordance with their source materials (0.993 for cfDNA libraries vs cfDNA; 0.998 for tumor tissue libraries vs tumor DNA with EGFR L858R; 0.991 for tumor tissue libraries vs tumor DNA with EGFR Ex19del). ConclusionsNGS and ddPCR demonstrate high concordance in EGFR mutation detection and VAF quantification, supporting their complementary roles in clinical testing. Using pre-capture libraries as an alternative to source samples can avoid repeat biopsies and enables subsequent testing for patients with inadequate FFPE sample quantity. These findings establish an evidence base for integrated diagnostic paradigms leveraging NGSs multiplexing power and ddPCRs sensitivity.
AO_SCPLOWBSTRACTC_SCPLOWHuman inverse dynamics is an important technique for analyzing human motion. Previous studies have typically estimated joint torques from joint pose images, marker coordinates, or EMG signals, which severely limit their applicability in real-world scenarios. In this work, we aim to directly predict joint torques during human movements from real human images. To address this gap, we present the vision-based inverse dynamics dataset (VID), the first dataset tailored for the joint torque prediction from real human images. VID comprises 63,369 frames of synchronized monocular images, kinematic data, and dynamic data of real human subjects. All data are carefully synchronized, refined, and manually validated to ensure high quality. In addition, we introduce a comprehensive benchmark for the vision-based inverse dynamics of real human images, consisting of a new baseline method and a new evaluation criteria with three levels of difficulty: (i) overall joint torque estimation, (ii) joint-specific analysis, and (iii) action-specific prediction. We further compare the baseline result of our VID-Network with other representative approaches, our baseline method achieves the state-of-the-art performance on almost all the evaluation criteria. By releasing VID and the accompanying evaluation protocol, we aim to establish a foundation for advancing biomechanics from real human images and to facilitate the exploration of new approaches for human inverse dynamics in unconstrained environments. The project website is https://github.com/MedVisionKitchen/VID-Benchmark.
SummaryBioinformatics analyses often rely on shell commands and small shell scripts whose executable context is difficult to preserve, inspect and reuse. TAFFISH addresses this gap by packaging command-line tool calls and lightweight shell flows as installable, versioned and inspectable executable units. Through a curated public Hub, TAFFISH indexes command interfaces, execution backends, platform constraints, release metadata and smoke-test/validation records. Together, these components provide a command-level reproducibility layer that works directly in ordinary shells and can also be invoked from existing workflow systems. Availability and ImplementationTAFFISH is implemented in Common Lisp and released under Apache-2.0. TAFFISH version 0.10.1, together with the source code, documentation and public Hub resources, is available at https://github.com/taffish/taffish, https://taffish.com, https://taffish.github.io and https://github.com/taffish/. The submitted software release, frozen Hub snapshot, test data, command logs, checksums and reproducibility package are archived at https://doi.org/10.5281/zenodo.21054185. The authors will maintain the software and public resources for at least two years after publication. ContactHao Lyu: hao.lyu@uestc.edu.cn; Hao Lin: hlin@uestc.edu.cn; Kejun Deng: dengkj@uestc.edu.cn Supplementary informationSupplementary data are available online.
RNA modifications are critical for the regulation of gene expression. Comprehensive profiling of RNA modifications is a prerequisite for unravelling their identities and functions. The plant hormone auxin plays pivotal roles in regulating plant growth, development, and environmental responses. In contrast to the extensive studies in animals, there is a paucity of knowledge regarding the existence and functions of RNA modifications in plants. This is due to the low abundance of modified ribonucleosides and the lack of accurate analytical methods. This study employs a state-of-the-art hydrophilic interaction liquid chromatography-tandem-mass spectrometry (HILIC-MS/MS) platform to achieve the sensitive detection of 12 nucleosides at the pM to nM level. Utilizing this platform, we were able to identify N4,2-O-dimethylcytidine (m4Cm) in plants by LC-MS/MS for the first time, which provided direct evidence for the widespread of this modification in plants. Moreover, a comprehensive investigation was undertaken to examine the 12 types of RNA modifications and their response patterns to auxin in a range of prototype plants, spanning from bryophyte (Marchantia polymorpha) to flowering plants (Arabidopsis thaliana, Oryza sativa, and Zea mays). The evolutionary patterns of the 12 RNA modifications were depicted from bryophyte to flowering plants, and their potential functions in plant growth, development, and auxin responses were revealed.
Plasmepsin V (PMV), an essential aspartyl protease, plays a critical role during the asexual blood stage of infection of Plasmodium by enabling the export of parasite proteins into the host red blood cell. This export is vital for parasite survival and pathogenesis, making PMV an attractive target for antimalarial drug development. Peptidomimetic inhibitors designed to mimic the natural substrate of PMV have demonstrated potent parasite-killing activity by blocking protein export. While these compounds have been instrumental in validating PMV as a bona fide antimalarial target, inconsistencies between their biochemical potency and cellular activity have raised questions regarding their precise mechanism of action. In this study, we employed chemoproteomic approaches, including solvent induced protein precipitation (SIP) and intact-cell thermal PISA profiling, to demonstrate PMV target engagement by the peptidomimetics. To further support these findings, we generated parasite lines exhibiting reduced sensitivity to peptidomimetics. Through whole-genome sequencing of these parasite lines, a single nucleotide variant (SNV) within the pmv gene was revealed. This mutation was later validated using reverse genetics, confirming its role in mediating resistance. Together, these data provide strong evidence that the peptidomimetics exert their antimalarial activity by directly targeting PMV. These findings further support the potential of PMV as a validated and promising target for future antimalarial drug development.
Molecular representation learning is a critical task in AI-driven drug development. While graph neural networks (GNNs) have demonstrated strong performance and gained widespread adoption in this field, efficiently extracting and explicitly analyzing functional groups remains a challenge. To address this issue, we propose MolUNet++, a novel model that employs Molecular Edge Shrinkage Pooling (MESPool) for hierarchical substructure extraction, utilizes a Nested UNet framework for multi-granularity feature integration, and incorporates a substructure masking explainer for quantitative fragment analysis. We evaluated MolUNet++ on tasks including molecular property prediction, drug-drug interaction (DDI) prediction, and drug-target interaction (DTI) prediction. Experimental results demonstrate that MolUNet++ not only outperforms traditional GNN models in predictive performance but also exhibits explicit, intuitive, and chemically logical interpretability. This capability provides valuable insights and tools for researchers in drug design and optimization.
Serine incorporator 5 (SERINC5) is a host restriction factor that targets certain enveloped viruses, including human immunodeficiency virus type 1 (HIV-1) and murine leukemia virus (MLV). It integrates into the viral envelope from the cell surface, inhibiting viral entry. SERINC5 is transported to the cell surface via polyubiquitination, while a single K130R mutation retains it in the cytoplasm. Both HIV-1 Nef and MLV glycoGag proteins antagonize SERINC5 by reducing its expression in producer cells. Here, we report that MLV glycoGag employs selective autophagy to downregulate SERINC5, demonstrating a more potent mechanism for decreasing its cell surface expression. Although glycoGag is a type II integral membrane protein, it primarily localizes to the cytoplasm and undergoes rapid proteasomal degradation. Employing the K130R mutant, we show that Nef, primarily associated with the plasma membrane, downregulates SERINC5 only after it has trafficked to the cell surface, whereas glycoGag can reduce its expression before reaching the plasma membrane while still in the cytoplasm. Nonetheless, an interaction with SERINC5 stabilizes and recruits glycoGag to the plasma membrane, enabling it to downregulate SERINC5 from the cell surface. Through affinity-purified mass spectrometry analysis combined with CRISPR/Cas9 knockouts, we find that glycoGags activity depends on reticulophagy regulator 1 (RETREG1), an ER-phagy receptor. Further knockout experiments of critical autophagy genes demonstrate that glycoGag downregulates cytoplasmic SERINC5 via micro-ER-phagy. These findings provide crucial new insights into the ongoing arms race between retroviruses and SERINC5 during infection. AUTHOR SUMMARYHIV-1 Nef and MLV glycoGag are unrelated viral proteins, yet both counteract the same host restriction factor, SERINC5, to facilitate productive infection. In this study, we report a novel pathway through which glycoGag downregulates SERINC5. We demonstrate that while Nef downregulates SERINC5 only after it has trafficked to the cell surface, glycoGag can directly downregulate SERINC5 in the cytoplasm before it reaches the plasma membrane. Furthermore, we show that this pathway is mediated by the ER-phagy receptor RETREG1, which targets SERINC5 for degradation via micro-ER-phagy. This mechanism provides a more effective means of blocking SERINC5 antiviral activity. These findings reveal that retroviruses have evolved different strategies to antagonize SERINC5, highlighting the critical role of SERINC5 in restricting retroviral infections.
Extrachromosomal circular DNA (ecDNA) are commonly produced within the nucleus to drive genome dynamics and heterogeneity, enabling cancer cell evolution and adaptation. However, the mechanisms underlying ecDNA biogenesis remain poorly understood. Here using genome-wide CRISPR screening in human cells, we identified the BRCA1-A and the LIG4 complexes mediate ecDNA production. Following DNA fragmentation, the upstream BRCA1-A complex protects DNA ends from excessive resection, promoting end-joining for circularization. Conversely, the MRN complex, which mediates end resection and thus antagonizes the BRCA1-A complex, suppresses ecDNA formation. Downstream, LIG4 conservatively catalyzes ecDNA production in Drosophila and mammals, with patient tumor ecDNA harboring junctions marked by LIG4 activity. Notably, disrupting LIG4 or BRCA1-A in cancer cells impairs ecDNA-mediated adaptation, hindering resistance to both chemotherapy and targeted therapies. Together, our study reveals the roles of the LIG4 and BRCA1-A complexes in ecDNA biogenesis, and uncovers new therapeutic targets to block ecDNA-mediated adaptation for cancer treatment.