J Sci Food Agric. 2024 Apr 9. doi: 10.1002/jsfa.13521. Online ahead of print.

ABSTRACT

BACKGROUND: The gluten composition is an important quality parameter of wheat flour. RP-HPLC is state-of-the-art for its analysis. As this is a very labour-intensive and time-consuming procedure, alternative faster methods are desirable. ELISA is a high-throughput method often used for the analysis of gluten traces in gluten-free products. In this proof-of-principle study, we introduce an experimental triple ELISA for the relative quantitation of gliadins, high-molecular-weight glutenin subunits (HMW-GS) and low-molecular-weight glutenin subunits (LMW-GS) out of one wheat flour extract.

RESULTS: The results of 80 common wheat flour samples obtained from the triple ELISA and RP-HPLC were correlated. The results for gliadins (r = 0.69) and HMW-GS (r = 0.81) showed a medium and high correlation, respectively. Only a very weak correlation of ELISA and RP-HPLC results was observed for LMW-GS (r = 0.49). Results for glutenins (r = 0.69) and gluten (r = 0.72) had a medium correlation. The gliadin/glutenin ratio (r = 0.47) and LMW-GS/HMW-GS ratio (r = 0.40) showed a weak or no correlation. The gliadin, LMW-GS and gluten contents were lower and the HMW-GS content was higher in the ELISA measurement compared to RP-HPLC.

CONCLUSION: The quantitation of gliadins and HMW-GS by the experimental triple ELISA showed comparable results to RP-HPLC, whereas no strong correlation between the results from the two methods was found for LMW-GS. Overall, the experimental triple ELISA is suitable for relative gluten quantitation, especially for the analysis of large sample sets. Further work will focus on improving the experimental procedure of the ELISA. This article is protected by copyright. All rights reserved.

PMID:38591632 | DOI:10.1002/jsfa.13521

PNAS Nexus. 2024 Mar 21;3(4):pgae121. doi: 10.1093/pnasnexus/pgae121. eCollection 2024 Apr.

ABSTRACT

Little is known about the contribution of 3D surface geometry to the development of multilayered tissues containing fibrous extracellular matrix components, such as those found in bone. In this study, we elucidate the role of curvature in the formation of chiral, twisted-plywood-like structures. Tissues consisting of murine preosteoblast cells (MC3T3-E1) were grown on 3D scaffolds with constant-mean curvature and negative Gaussian curvature for up to 32 days. Using 3D fluorescence microscopy, the influence of surface curvature on actin stress-fiber alignment and chirality was investigated. To gain mechanistic insights, we did experiments with MC3T3-E1 cells deficient in nuclear A-type lamins or treated with drugs targeting cytoskeleton proteins. We find that wild-type cells form a thick tissue with fibers predominantly aligned along directions of negative curvature, but exhibiting a twist in orientation with respect to older tissues. Fiber orientation is conserved below the tissue surface, thus creating a twisted-plywood-like material. We further show that this alignment pattern strongly depends on the structural components of the cells (A-type lamins, actin, and myosin), showing a role of mechanosensing on tissue organization. Our data indicate the importance of substrate curvature in the formation of 3D tissues and provide insights into the emergence of chirality.

PMID:38590971 | PMC:PMC10999733 | DOI:10.1093/pnasnexus/pgae121

Front Toxicol. 2024 Mar 25;6:1394396. doi: 10.3389/ftox.2024.1394396. eCollection 2024.

ABSTRACT

[This corrects the article DOI: 10.3389/ftox.2023.1212509.].

PMID:38590784 | PMC:PMC11000267 | DOI:10.3389/ftox.2024.1394396

RSC Adv. 2024 Apr 8;14(16):11258-11265. doi: 10.1039/d4ra01956c. eCollection 2024 Apr 3.

ABSTRACT

Synthetic polymer nanoparticles (NPs) with biomimetic properties are ideally suited for different biomedical applications such as drug delivery and direct therapy. However, bulk synthetic approaches can suffer from poor reproducibility and scalability when precise size control or multi-step procedures are required. Herein, we report an integrated microfluidic chip for the synthesis of polymer NPs. The chip could sequentially perform homopolymer synthesis and subsequent crosslinking into NPs without intermediate purification. This was made possible by fabrication of the chip with a fluorinated elastomer and incorporation of two microfluidic mixers. The first was a long channel with passive mixing features for the aqueous RAFT synthesis of stimuli-responsive polymers in ambient conditions. The polymers were then directly fed into a hydrodynamic flow focusing (HFF) junction that rapidly mixed them with a crosslinker solution to produce NPs. Compared to microfluidic systems made of PDMS or glass, our chip had better compatibility and facile fabrication. The polymers were synthesized with high monomer conversion and the NP size was found to be influenced by the flow rate ratio between the crosslinker solution and polymer solution. This allowed for the size to be predictably controlled by careful adjustment of the fluid flow rates. The size of the NPs and their stimuli-responses were studied using DLS and SEM imaging. This microfluidic chip design can potentially streamline and provide some automation for the bottom-up synthesis of polymer NPs while offering on-demand size control.

PMID:38590347 | PMC:PMC11000227 | DOI:10.1039/d4ra01956c

Genesis. 2024 Apr;62(2):e23597. doi: 10.1002/dvg.23597.

ABSTRACT

Sensory signals detected by olfactory sensory organs are critical regulators of animal behavior. An accessory olfactory organ, the vomeronasal organ, detects cues from other animals and plays a pivotal role in intra- and inter-species interactions in mice. However, how ethologically relevant cues control mouse behavior through approximately 350 vomeronasal sensory receptor proteins largely remains elusive. The type 2 vomeronasal receptor-A4 (V2R-A4) subfamily members have been repeatedly detected from vomeronasal sensory neurons responsive to predator cues, suggesting a potential role of this receptor subfamily as a sensor for predators. This review focuses on this intriguing subfamily, delving into its receptor functions and genetic characteristics.

PMID:38590121 | DOI:10.1002/dvg.23597

Plant Cell Physiol. 2024 Apr 8:pcae040. doi: 10.1093/pcp/pcae040. Online ahead of print.

ABSTRACT

Cellular responses to internal and external stimuli are orchestrated by intricate intracellular signaling pathways. To ensure an efficient and specific information flow, cells employ scaffold proteins as critical signaling organizers. With the ability to bind multiple signaling molecules, scaffold proteins can sequester signaling components within specific subcellular domains or modulate the efficiency of signal transduction. Scaffolds can also tune the output of signaling pathways by serving as regulatory targets. This review focuses on scaffold proteins associated with the plant GLYCOGEN SYNTHASE KINASE3-like kinase, BRASSINOSTEROID-INSENSITIVE2 (BIN2) that serve as a key negative regulator of brassinosteroid (BR) signaling. Here we summarize the current understanding of how scaffold proteins actively shape BR signaling outputs and crosstalk in plant cells via interactions with BIN2.

PMID:38590034 | DOI:10.1093/pcp/pcae040

Sci Total Environ. 2024 Apr 6:172297. doi: 10.1016/j.scitotenv.2024.172297. Online ahead of print.

ABSTRACT

Soil pollution by As and Hg is a pressing environmental issue given their persistence. The intricate removal processes and subsequent accumulation of these elements in soil adversely impact plant growth and pose risks to other organisms in the food chain and to underground aquifers. Here we assessed the effectiveness of non-toxic industrial byproducts, namely coal fly ash and steelmaking slag, as soil amendments, both independently and in conjunction with an organic fertilizer. This approach was coupled with a phytoremediation technique involving Betula pubescens to tackle soil highly contaminated. Greenhouse experiments were conducted to evaluate amendments' impact on the growth, physiology, and biochemistry of the plant. Additionally, a permeable barrier made of byproducts was placed beneath the soil to treat leachates. The application of the byproducts reduced pollutant availability, the production of contaminated leachates, and pollutant accumulation in plants, thereby promoting plant development and survival. Conversely, the addition of the fertilizer alone led to an increase in As accumulation in plants and induced the production of antioxidant compounds such as carotenoids and free proline. Notably, all amendments led to increased thiolic compound production without affecting chlorophyll synthesis. While fertilizer application significantly decreased parameters associated with oxidative stress, such as hydrogen peroxide and malondialdehyde, no substantial reduction was observed after byproduct application. Thermal desorption analysis of the byproducts revealed Hg immobilization mechanisms, thereby indicating retention of this metalloid in the form of Hg chloride. In summary, the revalorization of industrial byproducts in the context of the circular economy holds promise for effectively immobilizing metal(loid)s in heavily polluted soils. Additionally, this approach can be enhanced through synergies with phytoremediation.

PMID:38588736 | DOI:10.1016/j.scitotenv.2024.172297

Nature. 2024 Apr 8. doi: 10.1038/s41586-024-07379-z. Online ahead of print.

ABSTRACT

Broad-spectrum RAS inhibition holds the potential to benefit roughly a quarter of human cancer patients whose tumors are driven by RAS mutations1,2. RMC-7977 is a highly selective inhibitor of the active GTP-bound forms of KRAS, HRAS, and NRAS, with affinity for both mutant and wild type (WT) variants (RAS(ON) multi-selective)3. As >90% of human pancreatic ductal adenocarcinoma (PDAC) cases are driven by activating mutations in KRAS4, we assessed the therapeutic potential of the RAS(ON) multi-selective inhibitor RMC-7977 in a comprehensive range of PDAC models. We observed broad and pronounced anti-tumor activity across models following direct RAS inhibition at exposures that were well-tolerated in vivo. Pharmacological analyses revealed divergent responses to RMC-7977 in tumor versus normal tissues. Treated tumors exhibited waves of apoptosis along with sustained proliferative arrest whereas normal tissues underwent only transient decreases in proliferation, with no evidence of apoptosis. In the autochthonous KPC model, RMC-7977 treatment resulted in a profound extension of survival followed by on-treatment relapse. Analysis of relapsed tumors identified Myc copy number gain as a prevalent candidate resistance mechanism, which could be overcome by combinatorial TEAD inhibition in vitro. Together, these data establish a strong preclinical rationale for the use of broad-spectrum RAS-GTP inhibition in the setting of PDAC and identify a promising candidate combination therapeutic regimen to overcome monotherapy resistance.

PMID:38588697 | DOI:10.1038/s41586-024-07379-z

Genome Med. 2024 Apr 8;16(1):54. doi: 10.1186/s13073-024-01317-4.

ABSTRACT

BACKGROUND: Lung cancer is the leading cause of cancer-related death in the world. In contrast to many other cancers, a direct connection to modifiable lifestyle risk in the form of tobacco smoke has long been established. More than 50% of all smoking-related lung cancers occur in former smokers, 40% of which occur more than 15 years after smoking cessation. Despite extensive research, the molecular processes for persistent lung cancer risk remain unclear. We thus set out to examine whether risk stratification in the clinic and in the general population can be improved upon by the addition of genetic data and to explore the mechanisms of the persisting risk in former smokers.

METHODS: We analysed transcriptomic data from accessible airway tissues of 487 subjects, including healthy volunteers and clinic patients of different smoking statuses. We developed a computational model to assess smoking-associated gene expression changes and their reversibility after smoking is stopped, comparing healthy subjects to clinic patients with and without lung cancer.

RESULTS: We find persistent smoking-associated immune alterations to be a hallmark of the clinic patients. Integrating previous GWAS data using a transcriptional network approach, we demonstrate that the same immune- and interferon-related pathways are strongly enriched for genes linked to known genetic risk factors, demonstrating a causal relationship between immune alteration and lung cancer risk. Finally, we used accessible airway transcriptomic data to derive a non-invasive lung cancer risk classifier.

CONCLUSIONS: Our results provide initial evidence for germline-mediated personalized smoke injury response and risk in the general population, with potential implications for managing long-term lung cancer incidence and mortality.

PMID:38589970 | DOI:10.1186/s13073-024-01317-4

Genome Biol. 2024 Apr 8;25(1):88. doi: 10.1186/s13059-024-03226-6.

ABSTRACT

Inferring gene regulatory networks (GRNs) from single-cell data is challenging due to heuristic limitations. Existing methods also lack estimates of uncertainty. Here we present Probabilistic Matrix Factorization for Gene Regulatory Network Inference (PMF-GRN). Using single-cell expression data, PMF-GRN infers latent factors capturing transcription factor activity and regulatory relationships. Using variational inference allows hyperparameter search for principled model selection and direct comparison to other generative models. We extensively test and benchmark our method using real single-cell datasets and synthetic data. We show that PMF-GRN infers GRNs more accurately than current state-of-the-art single-cell GRN inference methods, offering well-calibrated uncertainty estimates.

PMID:38589899 | DOI:10.1186/s13059-024-03226-6

Nat Rev Cancer. 2024 Apr 8. doi: 10.1038/s41568-024-00686-7. Online ahead of print.

NO ABSTRACT

PMID:38589556 | DOI:10.1038/s41568-024-00686-7

Nat Plants. 2024 Apr 8. doi: 10.1038/s41477-024-01676-1. Online ahead of print.

NO ABSTRACT

PMID:38589486 | DOI:10.1038/s41477-024-01676-1

Nat Plants. 2024 Apr 8. doi: 10.1038/s41477-024-01666-3. Online ahead of print.

ABSTRACT

The symbiotic interaction of plants with arbuscular mycorrhizal (AM) fungi is ancient and widespread. Plants provide AM fungi with carbon in exchange for nutrients and water, making this interaction a prime target for crop improvement. However, plant-fungal interactions are restricted to a small subset of root cells, precluding the application of most conventional functional genomic techniques to study the molecular bases of these interactions. Here we used single-nucleus and spatial RNA sequencing to explore both Medicago truncatula and Rhizophagus irregularis transcriptomes in AM symbiosis at cellular and spatial resolution. Integrated, spatially registered single-cell maps revealed infected and uninfected plant root cell types. We observed that cortex cells exhibit distinct transcriptome profiles during different stages of colonization by AM fungi, indicating dynamic interplay between both organisms during establishment of the cellular interface enabling successful symbiosis. Our study provides insight into a symbiotic relationship of major agricultural and environmental importance and demonstrates a paradigm combining single-cell and spatial transcriptomics for the analysis of complex organismal interactions.

PMID:38589485 | DOI:10.1038/s41477-024-01666-3

Nat Commun. 2024 Apr 8;15(1):3018. doi: 10.1038/s41467-024-47124-8.

ABSTRACT

Ionizing radiation induces cell death in the gastrointestinal (GI) epithelium by activating p53. However, p53 also prevents animal lethality caused by radiation-induced acute GI syndrome. Through single-cell RNA-sequencing of the irradiated mouse small intestine, we find that p53 target genes are specifically enriched in regenerating epithelial cells that undergo fetal-like reversion, including revival stem cells (revSCs) that promote animal survival after severe damage of the GI tract. Accordingly, in mice with p53 deleted specifically in the GI epithelium, ionizing radiation fails to induce fetal-like revSCs. Using intestinal organoids, we show that transient p53 expression is required for the induction of revival stem cells and is controlled by an Mdm2-mediated negative feedback loop. Together, our findings reveal that p53 suppresses severe radiation-induced GI injury by promoting fetal-like reprogramming of irradiated intestinal epithelial cells.

PMID:38589357 | DOI:10.1038/s41467-024-47124-8

Bioinformatics. 2024 Apr 8:btae190. doi: 10.1093/bioinformatics/btae190. Online ahead of print.

ABSTRACT

MOTIVATION: Supervised deep learning is used to model the complex relationship between genomic sequence and regulatory function. Understanding how these models make predictions can provide biological insight into regulatory functions. Given the complexity of the sequence to regulatory function mapping (the cis-regulatory code), it has been suggested that the genome contains insufficient sequence variation to train models with suitable complexity. Data augmentation is a widely used approach to increase the data variation available for model training, however current data augmentation methods for genomic sequence data are limited.

RESULTS: Inspired by the success of comparative genomics, we show that augmenting genomic sequences with evolutionarily related sequences from other species, which we term phylogenetic augmentation, improves the performance of deep learning models trained on regulatory genomic sequences to predict high-throughput functional assay measurements. Additionally, we show that phylogenetic augmentation can rescue model performance when the training set is down-sampled and permits deep learning on a real-world small dataset, demonstrating that this approach improves data efficiency. Overall, this data augmentation method represents a solution for improving model performance that is applicable to many supervised deep learning problems in genomics.

AVAILABILITY: The open-source GitHub repository agduncan94/phylogenetic_augmentation_paper includes the code for rerunning the analyses here and recreating the figures.

SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

PMID:38588559 | DOI:10.1093/bioinformatics/btae190

Int J Syst Evol Microbiol. 2024 Apr;74(4). doi: 10.1099/ijsem.0.006293.

ABSTRACT

By investigating wet and dry age-related ripening of beef, Pseudomonas strains V3/3/4/13T and V3/K/3/5T were isolated. Strain V3/3/4/13T exhibited more than 99 % 16S rRNA gene-based similarity to Pseudomonas fragi and other members of this group, while isolate V3/K/3/5T was very close to Pseudomonas veronii and a number of relatives within the Pseudomonas fluorescens group. Additional comparisons of complete rpoB sequences and draft genomes allowed us to place isolate V3/3/4/13T close to Pseudomonas deceptionensis DSM 26521T. In the case of V3/K/3/5T the closest relative was P. veronii DSM 11331T. Average nucleotide identity (ANIb) and digital DNA-DNA hybridization (dDDH) values calculated from the draft genomes of V3/3/4/13T and P. deceptionensis DSM 26521T were 88.5 and 39.8 %, respectively. For V3/K/3/5T and its closest relative P. veronii DSM 11331T, the ANIb value was 95.1 % and the dDDH value was 60.7 %. The DNA G+C contents of V3/3/4/13T and V3/K/3/5T were 57.4 and 60.8 mol%, respectively. Predominant fatty acids were C16 : 0, C18 : 1 ω7c, C17 : 0 cyclo and summed feature C16 : 1 ω7ct/C15 : 0 iso 2OH. The main respiratory quinones were Q9, with minor proportions of Q8 and, in the case of V3/K/3/5T, additional Q10. The main polar lipids were diphosphatidylglycerol, phosphatidylglycerol, phosphatidylethanolamine and, in the case of V3/K/3/5T, additional phosphatidylcholine. Based on the combined data, isolates V3/3/4/13T and V3/K/3/5T should be considered as representatives of two novel Pseudomonas species. The type strain of the newly proposed Pseudomonas kulmbachensis sp. nov. is V3/3/4/13T (=DSM 113654T=LMG 32520T), a second strain belonging to the same species is FLM 004-28 (=DSM 113604=LMG 32521); the type strain for the newly proposed Pseudomonas paraveronii sp. nov. is V3/K/3/5T (=DSM 113573T=LMG 32518T) with a second isolate FLM 11 (=DSM 113572=LMG 32519).

PMID:38587505 | DOI:10.1099/ijsem.0.006293

J Cell Biol. 2024 Jun 3;223(6):e202309140. doi: 10.1083/jcb.202309140. Epub 2024 Apr 8.

ABSTRACT

β-Coronaviruses remodel host endomembranes to form double-membrane vesicles (DMVs) as replication organelles (ROs) that provide a shielded microenvironment for viral RNA synthesis in infected cells. DMVs are clustered, but the molecular underpinnings and pathophysiological functions remain unknown. Here, we reveal that host fragile X-related (FXR) family proteins (FXR1/FXR2/FMR1) are required for DMV clustering induced by expression of viral non-structural proteins (Nsps) Nsp3 and Nsp4. Depleting FXRs results in DMV dispersion in the cytoplasm. FXR1/2 and FMR1 are recruited to DMV sites via specific interaction with Nsp3. FXRs form condensates driven by liquid-liquid phase separation, which is required for DMV clustering. FXR1 liquid droplets concentrate Nsp3 and Nsp3-decorated liposomes in vitro. FXR droplets facilitate recruitment of translation machinery for efficient translation surrounding DMVs. In cells depleted of FXRs, SARS-CoV-2 replication is significantly attenuated. Thus, SARS-CoV-2 exploits host FXR proteins to cluster viral DMVs via phase separation for efficient viral replication.

PMID:38587486 | DOI:10.1083/jcb.202309140

Nucleic Acids Res. 2024 Apr 8:gkae227. doi: 10.1093/nar/gkae227. Online ahead of print.

ABSTRACT

Microbiology and synthetic biology depend on reverse genetic approaches to manipulate bacterial genomes; however, existing methods require molecular biology to generate genomic homology, suffer from low efficiency, and are not easily scaled to high throughput. To overcome these limitations, we developed a system for creating kilobase-scale genomic modifications that uses DNA oligonucleotides to direct the integration of a non-replicating plasmid. This method, Oligonucleotide Recombineering followed by Bxb-1 Integrase Targeting (ORBIT) was pioneered in Mycobacteria, and here we adapt and expand it for Escherichia coli. Our redesigned plasmid toolkit for oligonucleotide recombineering achieved significantly higher efficiency than λ Red double-stranded DNA recombineering and enabled precise, stable knockouts (≤134 kb) and integrations (≤11 kb) of various sizes. Additionally, we constructed multi-mutants in a single transformation, using orthogonal attachment sites. At high throughput, we used pools of targeting oligonucleotides to knock out nearly all known transcription factor and small RNA genes, yielding accurate, genome-wide, single mutant libraries. By counting genomic barcodes, we also show ORBIT libraries can scale to thousands of unique members (>30k). This work demonstrates that ORBIT for E. coli is a flexible reverse genetic system that facilitates rapid construction of complex strains and readily scales to create sophisticated mutant libraries.

PMID:38587185 | DOI:10.1093/nar/gkae227

EClinicalMedicine. 2024 Feb 1;68:102435. doi: 10.1016/j.eclinm.2024.102435. eCollection 2024 Feb.

ABSTRACT

BACKGROUND: Immune-mediated inflammatory diseases (IMIDs) are likely to complicate maternal health. However, literature on patients with IMIDs undergoing pregnancy is scarce and often overlooks the presence of comorbidities. We aimed to evaluate the impact of IMIDs on adverse pregnancy outcomes after assessing and addressing any discrepancies in the distribution of covariates associated with adverse pregnancy outcomes between patients with and without IMIDs.

METHODS: We conducted a retrospective cohort study using data from an integrated U.S. community healthcare system that provides care across Alaska, California, Montana, Oregon, New Mexico, Texas, and Washington. We used a database containing all structured data from electronic health record (EHRs) and analyzed the cohort of pregnant people who had live births from January 1, 2013, through December 31, 2022. We investigated 12 selected IMIDs: psoriasis, inflammatory bowel disease, rheumatoid arthritis, spondyloarthritis, multiple sclerosis, systemic lupus erythematosus, psoriatic arthritis, antiphospholipid syndrome, Sjögren's syndrome, vasculitides, sarcoidosis, and systemic sclerosis. We characterized patients with IMIDs prior to pregnancy (IMIDs group) based on pregnancy/maternal characteristics, comorbidities, and pre-pregnancy/prenatal immunomodulatory medications (IMMs) prescription patterns. We 1:1 propensity score matched the IMIDs cohort with people who had no IMID diagnoses prior to pregnancy (non-IMIDs cohort). Outcome measures were preterm birth (PTB), low birth weight (LBW), small for gestational age (SGA), and caesarean section.

FINDINGS: Our analytic cohort had 365,075 people, of which 5784 were in the IMIDs group and 359,291 were in the non-IMIDs group. The prevalence rate of pregnancy of at least 20 weeks duration in people with a previous IMID diagnosis has doubled in the past ten years. 17% of the IMIDs group had at least one prenatal IMM prescription. Depending on the type of IMM, 48%-70% of the patients taking IMMs before pregnancy continued them throughout pregnancy. Overall, patients with one or more of these 12 IMIDs had increased risk of PTB (Relative risk (RR) = 1.1 [1.0, 1.3]; p = 0.08), LBW (RR = 1.2 [1.0, 1.4]; p = 0.02), SGA (RR = 1.1 [1.0, 1.2]; p = 0.03), and caesarean section (RR = 1.1 [1.1, 1.2], p < 0.0001) compared to a matched cohort of people without IMIDs. When adjusted for comorbidities, patients with rheumatoid arthritis (PTB RR = 1.2, p = 0.5; LBW RR = 1.1, p = 0.6) and/or inflammatory bowel disease (PTB RR = 1.2, p = 0.3; LBW RR = 1.0, p = 0.8) did not have significantly increased risk for PTB and LBW.

INTERPRETATION: For patients who have been pregnant for 20 weeks or greater, the association between IMIDs and adverse pregnancy outcomes depends on both the nature of the IMID and the presence of comorbidities. Because this study was limited to pregnancies resulting in live births, results must be interpreted together with other studies on early pregnancy loss and stillbirth in patient with IMIDs.

FUNDING: National Institutes of Health.

PMID:38586478 | PMC:PMC10994966 | DOI:10.1016/j.eclinm.2024.102435

Comput Struct Biotechnol J. 2024 Mar 27;23:1339-1347. doi: 10.1016/j.csbj.2024.03.021. eCollection 2024 Dec.

ABSTRACT

Over the past decade, information for precision disease medicine has accumulated in the form of textual data. To effectively utilize this expanding medical text, we proposed a multi-task learning-based framework based on hard parameter sharing for knowledge graph construction (MKG), and then used it to automatically extract gastric cancer (GC)-related biomedical knowledge from the literature and identify GC drug candidates. In MKG, we designed three separate modules, MT-BGIPN, MT-SGTF and MT-ScBERT, for entity recognition, entity normalization, and relation classification, respectively. To address the challenges posed by the long and irregular naming of medical entities, the MT-BGIPN utilized bidirectional gated recurrent unit and interactive pointer network techniques, significantly improving entity recognition accuracy to an average F1 value of 84.5% across datasets. In MT-SGTF, we employed the term frequency-inverse document frequency and the gated attention unit. These combine both semantic and characteristic features of entities, resulting in an average Hits@ 1 score of 94.5% across five datasets. The MT-ScBERT integrated cross-text, entity, and context features, yielding an average F1 value of 86.9% across 11 relation classification datasets. Based on the MKG, we then developed a specific knowledge graph for GC (MKG-GC), which encompasses a total of 9129 entities and 88,482 triplets. Lastly, the MKG-GC was used to predict potential GC drugs using a pre-trained language model called BioKGE-BERT and a drug-disease discriminant model based on CNN-BiLSTM. Remarkably, nine out of the top ten predicted drugs have been previously reported as effective for gastric cancer treatment. Finally, an online platform was created for exploration and visualization of MKG-GC at https://www.yanglab-mi.org.cn/MKG-GC/.

PMID:38585647 | PMC:PMC10995799 | DOI:10.1016/j.csbj.2024.03.021