Biology subjects

Zeng, Y.

Publications and source records attributed to Zeng, Y..

At least 91 records · Page 5Linked to original sources

bioRxiv2024

Pervasive findings of directional selection realize the promise of ancient DNA to elucidate human adaptation

We present a method for detecting evidence of natural selection in ancient DNA time-series data that leverages an opportunity not utilized in previous scans: testing for a consistent trend in allele frequency change over time. By applying this to 8433 West Eurasians who lived over the past 14000 years and 6510 contemporary people, we find an order of magnitude more genome-wide significant signals than previous studies: 347 independent loci with >99% probability of selection. Previous work showed that classic hard sweeps driving advantageous mutations to fixation have been rare over the broad span of human evolution, but in the last ten millennia, many hundreds of alleles have been affected by strong directional selection. Discoveries include an increase from [~]0% to [~]20% in 4000 years for the major risk factor for celiac disease at HLA-DQB1; a rise from [~]0% to [~]8% in 6000 years of blood type B; and fluctuating selection at the TYK2 tuberculosis risk allele rising from [~]2% to [~]9% from [~]5500 to [~]3000 years ago before dropping to [~]3%. We identify instances of coordinated selection on alleles affecting the same trait, with the polygenic score today predictive of body fat percentage decreasing by around a standard deviation over ten millennia, consistent with the "Thrifty Gene" hypothesis that a genetic predisposition to store energy during food scarcity became disadvantageous after farming. We also identify selection for combinations of alleles that are today associated with lighter skin color, lower risk for schizophrenia and bipolar disease, slower health decline, and increased measures related to cognitive performance (scores on intelligence tests, household income, and years of schooling). These traits are measured in modern industrialized societies, so what phenotypes were adaptive in the past is unclear. We estimate selection coefficients at 9.9 million variants, enabling study of how Darwinian forces couple to allelic effects and shape the genetic architecture of complex traits.

evolutionary biology↗

bioRxiv2024

Human and bats genome robustness under COSMIC mutational signatures

Carcinogenesis is an evolutionary process, and mutations can fix the selected phenotypes in selective microenvironments. Both normal and neoplastic cells are robust to the mutational stressors in the microenvironment to the extent that secure their fitness. To test the robustness of genes under a range of mutagens, we developed a sequential mutation simulator, Sinabro, to simulate single base substitution under a given mutational process. Then, we developed a pipeline to measure the robustness of genes and cells under those mutagenesis processes. We discovered significant human genome robustness to the APOBEC mutational signature SBS2, which is associated with viral defense mechanisms and is implicated in cancer. Robustness evaluations across over 70,000 sequences against 41 signatures showed higher resilience under signatures predominantly causing C-to-T (G-to-A) mutations. Principal component analysis indicates the GC content at the codons wobble position significantly influences robustness, with increased resilience noted under transition mutations compared to transversions. Then, we tested our results in bats at extremes of the lifespan-to-mass relationship and found the long-lived bat is more robust to APOBEC than the short-lived one. By revealing robustness to APOBEC ranked highest in human (and bats with much more than number of APOBEC) genome, this work bolsters the key potential role of APOBECs in aging and cancer, as well as evolved countermeasures to this innate mutagenic process. It also provides the baseline of the human and bat genome robustness under mutational processes associated with aging and cancer. HighlightsO_LISinabro, the sequential mutation simulator, facilitates measuring the robustness of human protein-coding sequences under all COSMIC mutational signatures. C_LIO_LIRobustness under APOBEC mutational signatures showed the largest mean and standard deviation in the human genome. C_LIO_LIRobustness to mutational signatures analysis reveals the role of APOBECs is complementary to cancer in the evolvability of cancer cells in later stages. C_LIO_LIPrincipal component analysis indicates that the GC content at the codons wobble position significantly influences robustness. C_LIO_LIA long-lived bat (Myotis myotis) has higher robustness to APOBECs than a short-lived one (Molossus molossus) than humans. C_LI

genomics↗

bioRxiv2024

Reusing excess staple oligonucleotides for economical production of DNA origami

DNA origami has enabled the development of responsive drug-delivery vehicles with precision features that were previously not attainable in bionanotechnology. To reduce the costs of creating therapeutics-scale amounts of DNA origami that need to bear costly modifications with high occupancy, we reused the excess staple oligonucleotides that are leftover from the folding process to fold additional origami. We determined that a DNA origami can be successfully folded with up to 80% cost savings by cyclic recovery and reuse of excess staple strands. We found evidence that higher quality staple strands are preferentially incorporated into origami, consistent with past reports, and therefore are preferentially depleted from the free-strand pool. The folding of DNA origami with staple strands that were reused up to eleven times was indistinguishable by our panel of assays versus a control folded with new strands, so long as the reused oligonucleotides were replenished each cycle with a small excess of fresh strands. We also observed a high degree of incorporation of guests on the DNA origami. By recovering, reusing, and replenishing excess staple oligonucleotides, it is possible to significantly lessen production costs to create well-formed origami, which is useful to allow more therapeutic designs to be tested.

synthetic biology↗

bioRxiv2024

Cargo quantification of functionalized DNA origami for therapeutic application

In recent years, notable advances in nanotechnology-based drug delivery have emerged. A particularly promising platform in this field is DNA origami-based nanoparticles, which offer highly programmable surfaces, providing precise control over the nanoscale spacing and stoichiometry of various cargo. These versatile particles are finding diverse applications ranging from basic molecular biology to diagnostics and therapeutics. This growing interest creates the need for effective methods to quantify cargo on DNA origami nanoparticles. Our study consolidates several previously validated methods focusing on gel-based and fluorescence-based techniques, including multiplexed quantification of protein, peptide, and nucleic acid cargo on these nanoparticles. This work may serve as a valuable resource for groups researchers keen on utilizing DNA origami-based nanoparticles in therapeutic applications. Graphical abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=135 SRC="FIGDIR/small/609963v1_ufig1.gif" ALT="Figure 1"> View larger version (42K): org.highwire.dtl.DTLVardef@1703793org.highwire.dtl.DTLVardef@50b368org.highwire.dtl.DTLVardef@327a3corg.highwire.dtl.DTLVardef@1f732b0_HPS_FORMAT_FIGEXP M_FIG C_FIG

synthetic biology↗

bioRxiv2024

Your Blood is Out for Delivery: Considerations of Shipping Time and Temperature on Degradation of RNA from Stabilized Whole Blood

Remote research studies are an invaluable tool for reaching populations in geographical regions with limited access to large medical centers or universities. To expand the remote study toolkit, we have previously developed homeRNA, which allows for at-home self-collection and stabilization of blood and demonstrated the feasibility of using homeRNA in high temperature climates. Here, we expand upon this work through a systematic study exploring the effects of high temperature on RNA integrity through in-lab and field experiments. Compared to the frozen controls (overall mean RIN of 8.2, n = 8), samples kept at 37{degrees}C for 2, 4, and 8 days had mean RINs of 7.6, 5.9, and 5.2 (n = 3), respectively, indicating that typical shipping conditions ([~]2 days) yield samples suitable for downstream RNA sequencing. Shorter time intervals (6 hours) resulted in minimal RNA degradation (median RIN of 6.4, n = 3) even at higher temperatures (50{degrees}C) compared to the frozen control (mean RIN of 7.8, n = 3). Additionally, we shipped homeRNA-stabilized blood from a single donor to 14 different states and back during the summer with continuous temperature probes (7.1 median RIN, n = 42). Samples from all locations were analyzed with 3 mRNA-seq to assess differences in gene counts, with the transcriptomic data suggesting that there was no preferential degradation of transcripts as a result of different shipping times, temperatures, and regions. Overall, our data support that homeRNA can be used in elevated temperature conditions, enabling decentralized sample collection for telemedicine, global health, and clinical research.

bioengineering↗

bioRxiv2024

CellFM: a large-scale foundation model pre-trained on transcriptomics of 100 million human cells

The rapid evolution of single-cell sequencing technologies has facilitated precise transcriptomics profiling at the single-cell level, shedding light on the intricate heterogeneity within cellular populations. Despite these advances, the inherent diversity of cells and data challenges such as noise, batch effects, and sparsity, underscores the pressing need for a unified model to learn and represent cellular states effectively. Single-cell Large Language Models (LLMs) have been crafted to bridge this gap yet exhibit limited performance on human cells. This short-fall may stem from the confounding effects of training data from diverse species, partly because of limited cells for the single species. Here, we have compiled a dataset of approximately 100 million human cells sequenced by multiple technolo-gies from human single-cell datasets with various file types deposited in public databases and websites. Leveraging these extensive data cohorts, we developed CellFM, a robust single-cell foundation model with an impressive 800 million parameters, marking an eight-fold increase over the current largest single-species model. To ensure the training of CellFM on the MindSpore AI framework from Huawei, we have integrated RetNet, a Transformer architecture variant with lin-ear complexity for a balance between efficiency and performance, serving as the backbone of our model. Our comprehensive experiments have shown that CellFM outperforms existing models across diverse applications, such as cell annotation, perturbation prediction, and gene function prediction.

bioinformatics↗

bioRxiv2024

Sacrificing Adaptability for Functionality: The Ivory Tower of Macular Muller Cells

The predilection of many retinal diseases for the macula suggests it may be less resistant to stress than the peripheral retina. Profiling of single-cell level transcriptional changes found that the peripheral retina exhibited more transcriptional changes than the macula in response to stress. One pronounced change was in a subgroup of Muller cells (MCs) that were dominant in the peripheral retina. Genes more abundantly expressed in peripheral MCs were mainly associated with stress responses and were more influenced by light stress. In contrast, genes highly expressed in MCs that predominated in the macula had roles in cellular function and were less influenced by light stress. Metallothionein 1, A Kinase Anchor Protein 12 and MAF BZIP Transcription Factor F were more abundantly expressed in peripheral MCs than in macular MCs. Knockdown of these genes in primary MCs reduced their viability under stress. Our findings indicate that macular MCs are more directed toward maintaining retinal function rather than mounting a stress response when exposed to stress, which may contribute to the maculas vulnerability to degenerative diseases.

neuroscience↗

bioRxiv2024

Engineering PEG10 assembled endogenous virus-like particles with genetically encoded neoantigen peptides for cancer vaccination

Tumor neoantigen peptide vaccines hold potential for boosting cancer immunotherapy, yet efficiently co-delivering peptides and adjuvants to antigen-presenting cells in vivo remains challenging. Virus-like particle (VLP), which is a kind of multiprotein structure organized as virus, can deliver therapeutic substances into cells and stimulate immune response. However, the weak targeted delivery of VLP in vivo and its susceptibility to neutralization by antibodies hinder their clinical applications. Here, we firstly designed a novel protein carrier using the mammalian-derived capsid protein PEG10, which can self-assemble into endogenous VLP (eVLP) with high protein loading and transfection efficiency. Then, an engineered tumor vaccine, named ePAC, was developed by packaging genetically encoded neoantigen into eVLP with further modification of CpG-ODN on its surface to serve as an adjuvant and targeting unit to dendritic cells (DCs). Significantly, ePAC can efficiently target and transport neoantigens to DCs, and promote DCs maturation to induce neoantigen-specific T cells. Moreover, in mouse orthotopic liver cancer and humanized mouse tumor models, ePAC combined with anti-TIM-3 exhibited remarkable antitumor efficacy. Overall, these results support that ePAC could be safely utilized as cancer vaccines for antitumor therapy, showing significant potential for clinical translation.

cancer biology↗

bioRxiv2024

Generative Modelling of Oncogene-carrying Extrachromosomal Circular DNA Biogenesis and Dynamics in Cells

Extrachromosomal circular DNAs (ecDNA) are focal gene amplifications frequently associated with cancer development and often indicating a poor prognosis. To understand the early dynamics of oncogene-carrying ecDNAs, we previously developed CRISPR-C, a tool for precise ecDNA generation by deleting specific chromosomal regions. Here, we adapted CRISPR-C to recreate tumor ecDNAs. This method also allowed us to enhance ecDNA generation efficiency by directly delivering Cas9 protein and sgRNAs as a ribonucleoprotein complex. By using the modified CRISPR-C, we successfully generated ecDNAs carrying oncogenes (EGFR, CDK4, MDM2, MYC, MYCN, FGFR2, ABCB1, and DHFR) in various human cell types. Furthermore, we demonstrated that our method could generate chimeric ecDNAs composed of target sequences from distant intra or inter-chromosomal regions. Using these generative ecDNA cell models, we studied the oncogene ecDNA expression and stability. The MDM2 expression was increased after CRISPR-C, while CDK4 was decreased indicating genomic-context dependent effect. The copy number of CRISPR-C generated CDK4 was ecDNA increased in cells after a long period of treatment with the CDK4 inhibitor palbociclib. Unlike CDK4, the CRISPR-C generated ABCB1 ecDNA was unstable in cells under normal growth conditions, but is stably retained when the cells were treated with colcemid, a recognized substrate for ABCB1. We thus provide valuable tools and an attractive platform for studying ecDNA biogenesisy and in vitro drug screening on ecDNA stability.

cancer biology↗

bioRxiv2024

Exploration of the link between COVID-19 and gastric cancer from the perspective of bioinformatics and systems biology

BackgroundCoronavirus disease 2019 (COVID-19), an infectious disease caused by severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2), has caused a global pandemic. Gastric cancer (GC) poses a great threat to peoples health, which is a high- risk factor for COVID-19. Previous studies have found some associations between GC and COVID-19, whereas the underlying molecular mechanisms are not well understood. MethodsWe used a bioinformatics and systems biology approach to investigate the relationship between GC and COVID-19. The gene expression profiles of COVID-19 (GSE196822) and GC (GSE179252) were downloaded from the Gene Expression Omnibus (GEO) database. After identifying the shared differentially expressed genes (DEGs) for GC and COVID-19, functional annotation, protein-protein interaction (PPI) network, hub genes, transcriptional regulatory networks and candidate drugs were analyzed. ResultsA total of 209 shared DEGs were identified to explore the linkages between COVID-19 and GC. Functional analyses showed that Immune-related pathway collectively participated in the development and progression of COVID-19 and GC. In addition, there are selected 10 hub genes including CDK1, KIF20A, TPX2, UBE2C, HJURP, CENPA, PLK1, MKI67, IFI6, and IFIT2. The transcription factor/gene and miRNA/gene interaction networks identified 38 transcription factors (TFs) and 234 miRNAs. More importantly, we identified ten potential therapeutic agents, including ciclopirox, resveratrol, etoposide, methotrexate, trifluridine, enterolactone, troglitazone, calcitriol, dasatinib and deferoxamine, some of which have been reported to improve and treat GC and COVID-19. This study also provides insight into the diseases most associated with mutual DEGs, which may provide new ideas for research on the treatment of COVID-19. ConclusionsThis research has the possibility to be contributed to effective therapeutic in COVID-19 and GC.

bioinformatics↗

bioRxiv2024

Alpha Traveling Waves during Working Memory: Disentangling Bottom-up Gating and Top-down Gain Control

While previous works established the inhibitory role of alpha oscillations during working memory maintenance, it remains an open question whether such an inhibitory control is a top-down process. Here, we attempted to disentangle this issue by considering the spatio-temporal component of waves in the alpha band, i.e., alpha traveling waves. We reanalyzed two pre-existing and open-access EEG datasets where participants performed lateralized delayed match-to-sample working memory tasks. In the first dataset, the distractor load was manipulated (2, 4, or 6), whereas in the second dataset, the memory span varied between 1, 3, and 6 items. In both datasets, we focused on the propagation of alpha waves on the anterior-posterior axis during the retention period. Our results reveal an increase in alpha-band forward waves as the distractor load increased, but also an increase in forward waves and a decrease in backward waves as the memory set size increased. Notably, our results also showed a lateralization effect: alpha forward waves exhibited a more pronounced increase in the hemisphere contralateral to the distractors, whereas the reduction in backward waves was stronger in the hemisphere contralateral to the targets. In short, the forward waves were regulated by distractors, whereas targets inversely modulated backward waves. Such a dissociation of goal-related and goal-irrelevant physiological signals suggests the co-existence of bottom-up and top-down inhibitory processes: alpha forward waves might convey a gating effect driven by distractor load, while backward waves may represent direct top-down gain control of downstream visual areas. Significance StatementWhen exploring the functional role of alpha band neural oscillations during working memory, significant attention has been directed towards its amplitude modulation, with relatively limited exploration of spatial-temporal dynamics of this rather global brain oscillatory signature. The present study seeks to address this gap by examining the directionality of alpha wave propagation during working memory retention. Our findings offer novel insights into the well-established inhibitory role of alpha waves, demonstrating that this function is manifested differently according to their propagation directions: forward waves seem to facilitate bottom-up gating, while backward waves might mediate top-down gain control.

neuroscience↗

bioRxiv2024

Developmental Regulation of Alternative Polyadenylation in an Adult Stem Cell Lineage

Alternative Cleavage and Polyadenylation (APA) often results in production of mRNA isoforms with either longer or shorter 3UTRs from the same genetic locus, potentially impacting mRNA translation, localization and stability. Developmentally regulated APA can thus make major contributions to cell-type-specific gene expression programs as cells differentiate. During Drosophila spermatogenesis, approximately 500 genes undergo APA when proliferating spermatogonia differentiate into spermatocytes, producing transcripts with shortened 3 UTRs, leading to profound stage-specific changes in the proteins expressed. The molecular mechanisms that specify usage of upstream polyadenylation sites in spermatocytes are thus key to understanding the changes in cell state. Here, we show that upregulation of PCF11 and Cbc, the two components of Cleavage Factor II (CFII), orchestrates APA during Drosophila spermatogenesis. Knock down of PCF11 or cbc in spermatocytes caused dysregulation of APA, with many transcripts normally cleaved at a proximal site in spermatocytes now cleaved at their distal site, as in spermatogonia. Forced overexpression of CFII components in spermatogonia switched cleavage of some transcripts to the proximal site normally used in spermatocytes. Our findings reveal a developmental mechanism where changes in expression of specific cleavage factors can direct cell-type-specific APA at selected genes.

developmental biology↗

bioRxiv2024

Glioma-associated fibroblasts promote glioblastoma resistance to temozolomide through CCL2-CCR2 paracrine signaling

Complicated tumor microenvironment contributes mostly to chemoresistance in glioblastoma. Glioma-associated fibroblasts (GAFs) were recently identified as non-tumor stromal cells in the glioblastoma microenvironment, whereas their function in glioblastoma chemoresistance is unclear. Herein, we interrogated the correlation between GAFs and chemoresistance of glioblastoma by examining a series of patient-derived GAFs and glioblastoma organoids (GBOs), revealing that GAFs could promote temozolomide resistance in glioblastoma. Mechanistically, C-C motif chemokine ligand 2 (CCL2) secreted by GAFs selectively activated the ERK1/2 signaling in glioblastoma cells to potentiate temozolomide resistance. Pharmacologically disrupting the CCL2-CCR2 axis or MEK1/2-ERK1/2 pathway effectively improved the therapeutic efficacy of temozolomide in GBM cells and patient-derived GBOs, and both decreased phospho-ERK1/2 expression. Collectively, our results identified that targeting the GAF-dominated CCL2-ERK1/2 pathway may be an alternative strategy to alleviate the GAF-mediated chemoresistance of glioblastoma. SignificanceComprehensive interpretation of the mutual support between tumor microenvironment and cancer cells is demanded for glioma with poor response rates to chemotherapy. This study demonstrates that GAFs promote the temozolomide resistance of glioblastoma by secreting cytokine CCL2 to activate ERK1/2 pathway, which may serve as a potential druggable candidate. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=140 SRC="FIGDIR/small/581575v1_ufig1.gif" ALT="Figure 1"> View larger version (50K): org.highwire.dtl.DTLVardef@167b955org.highwire.dtl.DTLVardef@b892bcorg.highwire.dtl.DTLVardef@1a9b082org.highwire.dtl.DTLVardef@18327a8_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOGraphic abstract.C_FLOATNO Schematic illustration for GAFs mediated chemoresistance of GBMs and underlying mechanisms. We demonstrate that Glioma-associated Fibroblasts (GAFs) grow in gliomas by isolating and identifying a panel of patient-derived GAFs. CCL2 secreted by GAFs stimulates CCR2 in GBM cells, which promotes activation of the ERK1/2 expression to potentiate GBM chemoresistance. C_FIG

cancer biology↗

bioRxiv2024

Brain preparedness: The cortisol awakening response proacts dynamic organization of large-scale brain networks across emotional and executive functions

Emotion and cognition involve an intricate crosstalk of neural and endocrine systems that support allostatic processes for maintenance of dynamic equilibrium and rapid adaptation for upcoming challenges. As a hallmark of human endocrine activity, the cortisol awakening response (CAR) is recognized to play a critical role in modulating emotional and executive functions. Yet, the underlying mechanisms of such effects remain elusive. By leveraging pharmacological neuroimaging technique and Hidden Markov Modeling of brain state dynamics, we show that the CAR proactively modulates rapid reconfigurations (state) of large-scale brain networks across multi-task demands. Behaviorally, suppression of CAR proactively and selectively impaired accuracy for emotional discrimination task but not for working memory (WM). In parallel, suppressed CAR led to a decrease in the occurrence rate of brain state dominant to emotional processing, but an increase in brain state linking to executive control under high WM demand. Further energy-based analyses revealed an increase in transition frequency and sequence complexity along with an increased entropy during emotional tasks when suppressed CAR, suggesting a decreased energy supply. Moreover, an increased transition frequency was observed when shifting from neutral to emotional conditions, but an opposite pattern during WM task, with n decreased transition frequency shifts from low to high-executive demands. Our findings establish a causal link between CAR and dynamic allocation of neural resources for emotional and executive functions, suggesting a cognitive neuroendocrine account for CAR-mediated proactive effects and human allostasis.

neuroscience↗

bioRxiv2024

Potent pollen gene regulation by DNA glycosylases in maize

Although DNA methylation primarily represses TEs, it also represses select genes that are methylated in plant body tissues but demethylated by DNA glycosylases (DNGs) in endosperm or pollen. Activity of either one of two DNGs, MDR1 or DNG102, is essential for pollen viability in maize. Using single-pollen mRNA sequencing on pollen segregating mutations in both genes, we identified 58 candidate DNG target genes that account for 11.1% of the wild-type transcriptome but are silent or barely detectable in the plant body (sporophyte). They are unusual in their tendency to lack introns but even more so in their having TE-like methylation in their CDS. The majority have predicted functions in cell wall modification, and they likely support the rapid tip growth characteristic of pollen tubes. These results suggest a critical role for DNA methylation and demethylation in regulating maize genes with potential for extremely high expression in pollen but constitutive silencing elsewhere.

genetics↗

bioRxiv2024

Human Amniotic Epithelial Cells Promote Chx10-/Pax6+ Muller Glia Subpopulation Reprogramming into Photoreceptor-like Cells

Reprogramming Muller glia to regenerate neurons is a promising strategy for treating retinal degeneration, but whether Muller glia contain subpopulations with different regenerative fates remains unclear. Here, using single-cell RNA-seq analysis and Muller glia lineage-tracing mice with retinal degeneration, we reveal that Muller glia were heterogeneous and identify a specific Muller glial subpopulation (Chx10-/Pax6+) in healthy retinas that is activated and migrate to the outer nuclear layer (ONL) during photoreceptor degeneration. Transplantation of human amniotic epithelial cells (hAECs) facilitates the activation and extensive migration of the Chx10-/Pax6+ Muller glial subpopulation to the ONL, where they are reprogrammed into photoreceptor-like cells. Mechanistically, hAECs degrade the inhibitory extracellular matrix through regulating matrix metalloproteinases, which probably induces remodeling of the microenvironment of Muller glia and contributes to cell reprogramming. Consequently, hAEC transplantation improves visual function in rd10 mice. Our findings uncover a distinctive Muller glial subpopulation with the potential for reprogramming into photoreceptors.

cell biology↗

bioRxiv2024

reguloGPT: Harnessing GPT for Knowledge Graph Construction of Molecular Regulatory Pathways

MotivationMolecular Regulatory Pathways (MRPs) are crucial for understanding biological functions. Knowledge Graphs (KGs) have become vital in organizing and analyzing MRPs, providing structured representations of complex biological interactions. Current tools for mining KGs from biomedical literature are inadequate in capturing complex, hierarchical relationships and contextual information about MRPs. Large Language Models (LLMs) like GPT-4 offer a promising solution, with advanced capabilities to decipher the intricate nuances of language. However, their potential for end-to-end KG construction, particularly for MRPs, remains largely unexplored. ResultsWe present reguloGPT, a novel GPT-4 based in-context learning prompt, designed for the end-to-end joint name entity recognition, N-ary relationship extraction, and context predictions from a sentence that describes regulatory interactions with MRPs. Our reguloGPT approach introduces a context-aware relational graph that effectively embodies the hierarchical structure of MRPs and resolves semantic inconsistencies by embedding context directly within relational edges. We created a benchmark dataset including 400 annotated PubMed titles on N6-methyladenosine (m6A) regulations. Rigorous evaluation of reguloGPT on the benchmark dataset demonstrated marked improvement over existing algorithms. We further developed a novel G-Eval scheme, leveraging GPT-4 for annotation-free performance evaluation and demonstrated its agreement with traditional annotation-based evaluations. Utilizing reguloGPT predictions on m6A-related titles, we constructed the m6A-KG and demonstrated its utility in elucidating m6As regulatory mechanisms in cancer phenotypes across various cancers. These results underscore reguloGPTs transformative potential for extracting biological knowledge from the literature. Availability and implementationThe source code of reguloGPT, the m6A title and benchmark datasets, and m6A-KG are available at: https://github.com/Huang-AI4Medicine-Lab/reguloGPT.

bioinformatics↗

bioRxiv2024

TDP-43 nuclear loss in FTD/ALS causes widespread alternative polyadenylation changes

In frontotemporal dementia and amyotrophic lateral sclerosis, the RNA-binding protein TDP-43 is depleted from the nucleus. TDP-43 loss leads to cryptic exon inclusion but a role in other RNA processing events remains unresolved. Here, we show that loss of TDP-43 causes widespread changes in alternative polyadenylation, impacting expression of disease-relevant genes (e.g., ELP1, NEFL, and TMEM106B) and providing evidence that alternative polyadenylation is a new facet of TDP-43 pathology.

neuroscience↗