Chemical-protein Interactions Research Articles

Investigation of two-pore K+ (TPK) channels in Triticum aestivum L. suggests their role in stress response

Two-pore K+ (TPK) channels are voltage-independent and involved in stress response in plants. Herein, we identified 12 TaTPK genes located on nine chromosomes in the Triticum aestivum genome. The majority of TaTPK genes comprised two exons. Each TaTPK channel comprised four transmembrane (TM) helices, N- and C-terminal ion-channel domains, two EF-hand domains and one 14-3-3 binding site. Additionally, highly conserved ‘GYGD’ motif responsible for K+ ion specificity, was found in between the TMs in both the ion-channel domains. Nine TaTPK channels were predicted to be localised at the plasma membrane, while three were vacuolar. The protein-protein and protein-chemical interactions indicated the coordinated functioning of the TaTPK channels with the other K+ transporters and their possible interaction with the Ca2+-signaling pathway. Expression studies suggested their importance in both vegetative and reproductive tissues development. Significantly modulated expression of various TaTPK genes during heat, drought, combined heat and drought and salt stresses, and after fungal infestation, depicted their function in stress responses. The miRNAs and transcription factors interaction analyses suggested their role in the hormone, light, growth and development-related, and stress-responsive signaling cascades. The current study suggested vital functions of various TaTPK genes, especially in stress response, and would provide an opportunity for their detailed characterization in future studies.

Identification and validation of pyroptosis-related genes as potential biomarkers for hypertrophic cardiomyopathy: A comprehensive bioinformatics analysis.

Pyroptosis plays a key role in the death of cells including cardiomyocytes, and it is associated with a variety of cardiovascular diseases. However, the role of pyroptosis-related genes (PRGs) in hypertrophic cardiomyopathy (HCM) is not well characterized. This study aimed to identify key biomarkers and explore the molecular mechanisms underlying the functions of the PRGs in HCM. The differentially expressed genes were identified by GEO2R, and the differentially expressed pyroptosis-related genes (DEPRGs) of HCM were identified by combining with PRGs. Enrichment analysis was performed using the "clusterProfiler" package of the R software. Protein-protein interactions (PPI) network analysis was performed using the STRING database, and hub genes were screened using cytoHubba. TF-miRNA coregulatory networks and protein-chemical interactions were analyzed using NetworkAnalyst. RT-PCR/WB was used for expression validation of HCM diagnostic markers. Quantitative reverse transcription-polymerase chain reaction (qRT-PCR) and Western Blot (WB) were used to measure and compare the expression of the identified genes in the cardiac hypertrophy model and the control group. A total of 20 DEPRGs were identified, which primarily showed enrichment for the positive regulation of cytokine production, regulation of response to biotic stimulus, tumor necrosis factor production, and other biological processes. These processes primarily involved pathways related to Renin-angiotensin system, Adipocytokine signaling pathway and NF-kappa B signaling pathway. Then, a PPI network was constructed, and 8 hub genes were identified. After verification analysis, the finally identified HCM-related diagnostic markers were upregulated gene protein tyrosine phosphatase non-receptor type 11 (PTPN11), downregulated genes interleukin-1 receptor-associated kinase 3 (IRAK3), and annexin A2 (ANXA2). Further GSEA analysis revealed these 3 biomarkers primarily related to cardiac muscle contraction, hypertrophic cardiomyopathy, fatty acid degradation and ECM - receptor interaction. Moreover, we also elucidated the interaction network of these biomarkers with the miRNA network and known compounds, respectively. RT-PCR/WB results indicated that PTPN11 expression was significantly increased, and IRAK3 and ANXA2 expressions were significantly decreased in HCM. This study identified PTPN11, IRAK3, and ANXA2 as pyroptosis-associated biomarkers of HCM, with the potential to reveal the development and pathogenesis of HCM and could be potential therapeutic targets.

A novel cholesterol metabolism-related ferroptosis pathway in hepatocellular carcinoma

BackgroundEmerging studies have reported the contribution of cholesterol to hepatocellular carcinoma (HCC) progression. However, the specific role and mechanism of cholesterol metabolism on spontaneous and progressive HCC development from the point of view of ferroptosis are still worth exploring. The present study aimed to reveal a novel mechanism of cholesterol metabolism-related ferroptosis in hepatocellular carcinoma cells.MethodsTwo microarray datasets (GSE25097, GSE22058) related to HCC were downloaded from Gene Expression Omnibus (GEO) datasets. Metabolomics analysis was performed by ultra performance liquid chromatography - tandem mass spectrometer (UPLC-MS/MS). The cholesterol-related proteins were downloaded from HMBD. Ferroptosis-related genes were extracted from FerrDb database. Data sets were separated into two groups. GSE25097 was used to identify ferroptosis-related genes, and GSE22058 was used to verify results. During these processes, chemical–protein interaction (CPI), protein–protein interaction (PPI), the Gene Ontology (GO), and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were conducted. Multivariate logistic regression analysis was used to test the associated pathway.ResultsWe identified 8 differentially expressed ferroptosis-related genes (HAMP, PTGS2, IL1B, ALOX15B, CDKN2A, RRM2, NQO1 and KIF20A) and 4 differentially expressed cholesterol-related genes (LCAT, CH25H, CEL and CYP7A1). Furthermore, based on the predicted results with STITCH, we identified indomethacin and IL1B as the essential node for cholesterol-mediated ferroptosis in hepatocellular carcinoma cell. Multivariate logistic regression analysis showed the activities of plasma IL1B in liver cancer patients enrolled have been significantly affected by the level of plasma cholesterol (P < 0.001) and the test result of IL1B is a predictor variable causing the changes of serum Fe levels (P < 0.001).ConclusionsOur findings shed new light on the association between cholesterol metabolism and ferroptosis in HCC, and suggest that IL1B is the necessary node for cholesterol to lead to ferroptosis process in HCC. Also, we identified the potential role of indomethacin in adjuvant therapy of HCC with complications of abnormal cholesterol metabolism.

DeepCompoundNet: enhancing compound–protein interaction prediction with multimodal convolutional neural networks

Virtual screening has emerged as a valuable computational tool for predicting compound–protein interactions, offering a cost-effective and rapid approach to identifying potential candidate drug molecules. Current machine learning-based methods rely on molecular structures and their relationship in the network. The former utilizes information such as amino acid sequences and chemical structures, while the latter leverages interaction network data, such as protein–protein interactions, drug–disease interactions, and protein–disease interactions. However, there has been limited exploration of integrating molecular information with interaction networks. This study presents DeepCompoundNet, a deep learning-based model that integrates protein features, drug properties, and diverse interaction data to predict chemical–protein interactions. DeepCompoundNet outperforms state-of-the-art methods for compound–protein interaction prediction, as demonstrated through performance evaluations. Our findings highlight the complementary nature of multiple interaction data, extending beyond amino acid sequence homology and chemical structure similarity. Moreover, our model’s analysis confirms that DeepCompoundNet gets higher performance in predicting interactions between proteins and chemicals not observed in the training samples. Communicated by Ramaswamy H. Sarma

Combination of computational new approach methodologies for enhancing evidence of biological pathway conservation across species

The ability to predict which chemicals are of concern for environmental safety is dependent, in part, on the ability to extrapolate chemical effects across many species. This work investigated the complementary use of two computational new approach methodologies to support cross-species predictions of chemical susceptibility: the US Environmental Protection Agency Sequence Alignment to Predict Across Species Susceptibility (SeqAPASS) tool and Unilever's recently developed Genes to Pathways – Species Conservation Analysis (G2P-SCAN) tool. These stand-alone tools rely on existing biological knowledge to help understand chemical susceptibility and biological pathway conservation across species. The utility and challenges of these combined computational approaches were demonstrated using case examples focused on chemical interactions with peroxisome proliferator activated receptor alpha (PPARα), estrogen receptor 1 (ESR1), and gamma-aminobutyric acid type A receptor subunit alpha (GABRA1). Overall, the biological pathway information enhanced the weight of evidence to support cross-species susceptibility predictions. Through comparisons of relevant molecular and functional data gleaned from adverse outcome pathways (AOPs) to mapped biological pathways, it was possible to gain a toxicological context for various chemical-protein interactions. The information gained through this computational approach could ultimately inform chemical safety assessments by enhancing cross-species predictions of chemical susceptibility. It could also help fulfill a core objective of the AOP framework by potentially expanding the biologically plausible taxonomic domain of applicability of relevant AOPs.

A hierarchical convolutional model for biomedical relation extraction

Biomedical relation extraction aims to extract the interactive relations between biomedical entities in a sentence. In biomedical literature, the sentence lengths are longer compared to general domain. Because all entity pairs in a sentence share the same contextual features and each of them should be evaluated, the biomedical relation extract task suffers from more serious semantic overlapping and data imbalance problems. In this study, a hierarchical convolutional model is proposed to address these problems. This model organizes the convolution operation in a hierarchical structure. First, after a sentence is divided into several channels by using the semantic structure of a relation instance, a token-level convolution is used to encodes local contextual features of different channels. Then, a channel-level convolution is designed to encode global semantic dependencies of a sentence. The hierarchical convolution is effective to learn contextual features and semantic dependencies relevant to ordered named entities, which enhance the discriminability of a neural network for biomedical relation extraction. The proposed method was evaluated on seven public datasets for three tasks: including protein-protein, drug-drug, and chemical-protein interactions. It outperformed state-of-the-art performance about 3.7% in the F1 score.

A large-scale evaluation of NLP-derived chemical-gene/protein relationships from the scientific literature: Implications for knowledge graph construction.

One area of active research is the use of natural language processing (NLP) to mine biomedical texts for sets of triples (subject-predicate-object) for knowledge graph (KG) construction. While statistical methods to mine co-occurrences of entities within sentences are relatively robust, accurate relationship extraction is more challenging. Herein, we evaluate the Global Network of Biomedical Relationships (GNBR), a dataset that uses distributional semantics to model relationships between biomedical entities. The focus of our paper is an evaluation of a subset of the GNBR data; the relationships between chemicals and genes/proteins. We use Evotec's structured 'Nexus' database of >2.76M chemical-protein interactions as a ground truth to compare with GNBRs relationships and find a micro-averaged precision-recall area under the curve (AUC) of 0.50 and a micro-averaged receiver operating characteristic (ROC) curve AUC of 0.71 across the relationship classes 'inhibits', 'binding', 'agonism' and 'antagonism', when a comparison is made on a sentence-by-sentence basis. We conclude that, even though these micro-average scores are modest, using a high threshold on certain relationship classes like 'inhibits' could yield high fidelity triples that are not reported in structured datasets. We discuss how different methods of processing GNBR data, and the factuality of triples could affect the accuracy of NLP data incorporated into knowledge graphs. We provide a GNBR-Nexus(ChEMBL-subset) merged datafile that contains over 20,000 sentences where a protein/gene-chemical co-occur and includes both the GNBR relationship scores as well as the ChEMBL (manually curated) relationships (e.g., 'agonist', 'inhibitor') -this can be accessed at https://doi.org/10.5281/zenodo.8136752. We envisage this being used to aid curation efforts by the drug discovery community.

Systems biology approach discovers comorbidity interaction of Parkinson's disease with psychiatric disorders utilizing brain transcriptome.

Several studies found that most patients with Parkinson's disorder (PD) appear to have psychiatric symptoms such as depression, anxiety, hallucination, delusion, and cognitive dysfunction. Therefore, recognizing these psychiatrically symptoms of PD patients is crucial for both symptomatic therapy and better knowledge of the pathophysiology of PD. In order to address this issue, we created a bioinformatics framework to determine the effects of PD mRNA expression on understanding its relationship with psychiatric symptoms in PD patients. We have discovered a significant overlap between the sets of differentially expressed genes from PD exposed tissue and psychiatric disordered tissues using RNA-seq datasets. We have chosen Bipolar disorder and Schizophrenia as psychiatric disorders in our study. A number of significant correlations between PD and the occurrence of psychiatric diseases were also found by gene set enrichment analysis, investigations of the protein-protein interaction network, gene regulatory network, and protein-chemical agent interaction network. We anticipate that the results of this pathogenetic study will provide crucial information for understanding the intricate relationship between PD and psychiatric diseases.

SPBERE: Boosting span-based pipeline biomedical entity and relation extraction via entity information

Triplet extraction is one of the fundamental tasks in biomedical text mining. Compared with traditional pipeline approaches, joint methods can alleviate the error propagation problem from entity recognition to relation classification. However, existing methods face challenges in detecting overlapping entities and overlapping relations, which are ubiquitous in biomedical texts. In this work, we propose a novel pipeline method of end-to-end biomedical triplet extraction. In particular, a span-based detection strategy is used to detect the overlapping triplets by enumerating possible candidate spans and entity pairs. The strategy is further used to capture different contextualized representations via an entity model and a relation model, respectively. Furthermore, to enhance interrelation between spans, entity information from the output of the entity model is used to construct the input for the relation model without utilizing any external knowledge. Our approach is evaluated on the drug–drug interaction (DDI) and chemical-protein interaction (CHEMPROT) datasets, exhibiting improvement of the absolute F1-score in relation extraction by 3.5%–3.7% compared prior work. The experimental results highlight the importance of overlapping triplet detection using the span-based approach, acquisition of various contextualized representations via different in-domain pre-trained language models, and early fusion of entity information in the relation model.

End-to-End Models for Chemical-Protein Interaction Extraction: Better Tokenization and Span-Based Pipeline Strategies.

End-to-end relation extraction (E2ERE) is an important task in information extraction, more so for biomedicine as scientific literature continues to grow exponentially. E2ERE typically involves identifying entities (or named entity recognition (NER)) and associated relations, while most RE tasks simply assume that the entities are provided upfront and end up performing relation classification. E2ERE is inherently more difficult than RE alone given the potential snowball effect of errors from NER leading to more errors in RE. A complex dataset in biomedical E2ERE is the ChemProt dataset (BioCreative VI, 2017) that identifies relations between chemical compounds and genes/proteins in scientific literature. ChemProt is included in all recent biomedical natural language processing benchmarks including BLUE, BLURB, and BigBio. However, its treatment in these benchmarks and in other separate efforts is typically not end-to-end, with few exceptions. In this effort, we employ a span-based pipeline approach to produce a new state-of-the-art E2ERE performance on the ChemProt dataset, resulting in > 4% improvement in F1-score over the prior best effort. Our results indicate that a straightforward fine-grained tokenization scheme helps span-based approaches excel in E2ERE, especially with regards to handling complex named entities. Our error analysis also identifies a few key failure modes in E2ERE for ChemProt.

Construction of Protein–Protein Interaction Networks on Capsicum Against Sepsis by Regulating of Molecular Mechanisms: Bioinformatics Approaches

We aimed to create the construction of protein–protein interaction networks for target proteins and to analyze the molecular characteristics of capsicum, screen and identify the potential targets of capsicum against sepsis using bioinformatics approaches. The potential anti-sepsis targets of capsicum were screened by chemical protein interaction analysis and matched with the published anti-sepsis related disease targets in the Online Mendelian Inheritance in Man database (OMIM) and the Genecards Database to screen out the anti-sepsis targets of capsicum. The potential targets of capsicum against sepsis were further identified by molecular docking server. Capsicum exhibited good medicinal properties. Overall, 14 potential targets were screened and 3453 disease targets were obtained from OMIM and Genecards databases. Subsequently, we found that there were 14 proteins that were both molecular targets and disease targets. The molecular docking server identified that the active ingredient beta-carotene had the ability to bind to 11 targeted proteins, which was a potential target of capsicum against sepsis. Capsicum can exert its therapeutic effects by interacting with key amino acid residues of the targeted proteins. Capsicum can inhibit tissue and organ damage of sepsis by regulating CAV1 and CTNNB1, and reduce inflammatory response of sepsis by regulating MYC, VEGFA, NR3C1 and DUSP1d, and ultimately prevent and treat sepsis.

Biomedical Relation Extraction Using Dependency Graph and Decoder-Enhanced Transformer Model.

The identification of drug-drug and chemical-protein interactions is essential for understanding unpredictable changes in the pharmacological effects of drugs and mechanisms of diseases and developing therapeutic drugs. In this study, we extract drug-related interactions from the DDI (Drug-Drug Interaction) Extraction-2013 Shared Task dataset and the BioCreative ChemProt (Chemical-Protein) dataset using various transfer transformers. We propose BERTGAT that uses a graph attention network (GAT) to take into account the local structure of sentences and embedding features of nodes under the self-attention scheme and investigate whether incorporating syntactic structure can help relation extraction. In addition, we suggest T5slim_dec, which adapts the autoregressive generation task of the T5 (text-to-text transfer transformer) to the relation classification problem by removing the self-attention layer in the decoder block. Furthermore, we evaluated the potential of biomedical relation extraction of GPT-3 (Generative Pre-trained Transformer) using GPT-3 variant models. As a result, T5slim_dec, which is a model with a tailored decoder designed for classification problems within the T5 architecture, demonstrated very promising performances for both tasks. We achieved an accuracy of 91.15% in the DDI dataset and an accuracy of 94.29% for the CPR (Chemical-Protein Relation) class group in ChemProt dataset. However, BERTGAT did not show a significant performance improvement in the aspect of relation extraction. We demonstrated that transformer-based approaches focused only on relationships between words are implicitly eligible to understand language well without additional knowledge such as structural information.

Interpretable Ensembles of Classifiers for Uncertain Data With Bioinformatics Applications.

Data uncertainty remains a challenging issue in many applications, but few classification algorithms can effectively cope with it. An ensemble approach for uncertain categorical features has recently been proposed, achieving promising results. It consists in biasing the sampling of features for each model in an ensemble so that less uncertain features are more likely to be sampled. Here we extend this idea of biased sampling and propose two new approaches: one for selecting training instances for each model in an ensemble and another for sampling features to be considered when splitting a node in a Random Forest training. We applied these approaches to classify ageing-related genes and predict drugs' side effects based on uncertain features representing protein-protein and protein-chemical interactions. We show that ensembles based on our proposed approaches achieve better predictive performance. In particular, our proposed approaches improved the performance of a Random Forest based on the most sophisticated approach for handling uncertain data in ensembles of this kind. Furthermore, we propose two new approaches for interpreting an ensemble of Naive Bayes classifiers and analyse their results on our datasets of ageing-related genes and drug's side effects.

Structural Characterization and Functional Interaction Analysis of Radiation Induced Protein Flt3 in Human and Mouse

Increasing risk of radiation exposure can be due to various reasons such as terrorist attacks, detonation of nuclear reactors, nuclear weapons, as well as acute radiation exposure on large scale poses high risk to humankind. Therefore, there is a critical need of identification and authentication of fast and early-response radiation induced harm biomarkers to distinguish between exposed and non-exposed individuals on a large-scale basis. This study deals with structural characterization of radiation induced proteins that are probable biomarkers such as Flt3L using database tools such as prot-param for physio-chemical properties, GOR IV to predict the secondary structure, swiss modelling for 3D model of protein, stitch and string analysis for protein-chemical interaction and protein-protein interaction respectively. In-silico analysis showcasing, protein Flt3L physical parameters such as theoretical PI 8.72, hydrophobicity 0.100, aliphatic index 111.81. Secondary structure, alpha helix 41.38%, extended strand 11.64%. Homology modelling showing sequence identity 74.26%, coverage 0.50 and range 1-101. In stich, Flt3 activating KITLG and showing binding affinity to Zn molecule. Identification of protein-protein interacting functional partner of Flt3L using string shows interactions with polyubiquitin-C (UBC), KITLG, Nras proto onco genes, etc, it showed experimental evidence of Flt3L and KITLG protein. Hence, by using in-silico analysis we can perform a preliminary study to predict the structures and chemical interaction of potential radiation induced probable biomarkers that can be utilized for rapid detection of radiation exposure and drugdiscovery.

Building the Environmental Chemical-Protein Interaction Network (eCPIN): An Exposome-Wide Strategy for Bioactive Chemical Contaminant Identification.

Although advancements in nontargeted analysis have made it possible to detect hundreds of chemical contaminants in a single run, the current environmental toxicology approaches lag behind, precluding the transition from analytical chemistry efforts to health risk assessment. We herein highlighted a recently developed "top-down" bioanalytical method, protein Affinity Purification with Nontargeted Analysis (APNA), to screen for bioactive chemical contaminants at the "exposome-wide" level. To achieve this, a tagged functional protein is employed as a "bait" to directly isolate bioactive chemical contaminants from environmental mixtures, which are further identified by nontargeted analysis. Advantages of this protein-guided approach, including the discovery of new bioactive ligands as well as new protein targets for known chemical contaminants, were highlighted by several case studies. Encouraged by these successful applications, we further proposed a framework, i.e., the environmental Chemical-Protein Interaction Network (eCPIN), to construct a complete map of the 7 billion binary interactions between all chemical contaminants (>350,000) and human proteins (∼20,000) via APNA. The eCPIN could be established in three stages through strategically prioritizing the ∼20,000 human proteins, such as focusing on the 48 nuclear receptors (e.g., thyroid hormone receptors) in the first stage. The eCPIN will provide an unprecedented throughput for screening bioactive chemical contaminants at the exposome-wide level and facilitate the identification of molecular initiating events at the proteome-wide level.

End-to-end sequence-structure-function meta-learning predicts genome-wide chemical-protein interactions for dark proteins.

Systematically discovering protein-ligand interactions across the entire human and pathogen genomes is critical in chemical genomics, protein function prediction, drug discovery, and many other areas. However, more than 90% of gene families remain "dark"-i.e., their small-molecule ligands are undiscovered due to experimental limitations or human/historical biases. Existing computational approaches typically fail when the dark protein differs from those with known ligands. To address this challenge, we have developed a deep learning framework, called PortalCG, which consists of four novel components: (i) a 3-dimensional ligand binding site enhanced sequence pre-training strategy to encode the evolutionary links between ligand-binding sites across gene families; (ii) an end-to-end pretraining-fine-tuning strategy to reduce the impact of inaccuracy of predicted structures on function predictions by recognizing the sequence-structure-function paradigm; (iii) a new out-of-cluster meta-learning algorithm that extracts and accumulates information learned from predicting ligands of distinct gene families (meta-data) and applies the meta-data to a dark gene family; and (iv) a stress model selection step, using different gene families in the test data from those in the training and development data sets to facilitate model deployment in a real-world scenario. In extensive and rigorous benchmark experiments, PortalCG considerably outperformed state-of-the-art techniques of machine learning and protein-ligand docking when applied to dark gene families, and demonstrated its generalization power for target identifications and compound screenings under out-of-distribution (OOD) scenarios. Furthermore, in an external validation for the multi-target compound screening, the performance of PortalCG surpassed the rational design from medicinal chemists. Our results also suggest that a differentiable sequence-structure-function deep learning framework, where protein structural information serves as an intermediate layer, could be superior to conventional methodology where predicted protein structures were used for the compound screening. We applied PortalCG to two case studies to exemplify its potential in drug discovery: designing selective dual-antagonists of dopamine receptors for the treatment of opioid use disorder (OUD), and illuminating the understudied human genome for target diseases that do not yet have effective and safe therapeutics. Our results suggested that PortalCG is a viable solution to the OOD problem in exploring understudied regions of protein functional space.

Mr.Vc v2: An updated version of database with increased data of transcriptome and experimental validated interactions.

Mr.Vc is a database of curated Vibrio cholerae transcriptome data and annotated information. The main objective is to facilitate the accessibility and reusability of the rapidly growing Vibrio cholerae omics data and relevant annotation. To achieve these goals, we performed manual curation on the transcriptome data and organized the datasets in an experiment-centric manner. We collected unknown operons annotated through text-mining analysis that would provide more clues about how Vibrio cholerae modulates gene regulation. Meanwhile, to understand the relationship between genes or experiments, we performed gene co-expression analysis and experiment-experiment correlation analysis. In additional, functional module named "Interactions" which dedicates to collecting experimentally validated interactions about Vibrio cholerae from public databases, MEDLINE documents and literature in life science journals. To date, Mr.Vc v2, which is significantly increased from the previous version, contains 107 microarray experiments, 106 RNA-seq experiments, and 3 Tn-seq projects, covering 56,839 entries of DEGs (Differentially Expressed Genes) from transcriptomes and 7,463 related genes from Tn-seq, respectively. and a total of 270,129 gene co-expression entries and 11,990 entries of experiment-experiment correlation was obtained, in total 1,316 entries of interactions were collected, including 496 protein-chemical signaling molecule interactions, 472 protein-protein interactions, 306 TF (Transcription Factor)-gene interactions and 42 Vibrio cholerae-virus interactions, most of which obtained from 402 literature through text-mining analysis. To make the information easier to access, Mr.Vc v2 is equipped with a search widget, enabling users to query what they are interested in. Mr.Vc v2 is freely available at http://mrvcv2.biownmc.info.

Chemical-protein relation extraction with ensembles of carefully tuned pretrained language models.

The identification of chemical-protein interactions described in the literature is an important task with applications in drug design, precision medicine and biotechnology. Manual extraction of such relationships from the biomedical literature is costly and often prohibitively time-consuming. The BioCreative VII DrugProt shared task provides a benchmark for methods for the automated extraction of chemical-protein relations from scientific text. Here we describe our contribution to the shared task and report on the achieved results. We define the task as a relation classification problem, which we approach with pretrained transformer language models. Upon this basic architecture, we experiment with utilizing textual and embedded side information from knowledge bases as well as additional training data to improve extraction performance. We perform a comprehensive evaluation of the proposed model and the individual extensions including an extensive hyperparameter search leading to 2647 different runs. We find that ensembling and choosing the right pretrained language model are crucial for optimal performance, whereas adding additional data and embedded side information did not improve results. Our best model is based on an ensemble of 10 pretrained transformers and additional textual descriptions of chemicals taken from the Comparative Toxicogenomics Database. The model reaches an F1 score of 79.73% on the hidden DrugProt test set and achieves the first rank out of 107 submitted runs in the official evaluation. Database URL: https://github.com/leonweber/drugprot.

SIMEON: Prediction of Chemical-protein Interaction via Stacked Bi-GRU-normalization Network and External Biomedical Knowledge

Background: Chemical compounds and proteins/genes are an important class of entities in biomedical research, and their interactions play a key role in precision medicine, drug discovery, basic clinical research, and building knowledge bases. Many computational methods have been proposed to identify chemical–protein interactions. However, the majority of these proposed models cannot model long-distance dependencies between chemical and protein, and the neural networks used to suffer from gradient descent, with little taking into account the characteristics of the chemical structure characteristics of the compound. Methods: To address the above limitations, we propose a novel model, SIMEON, to identify chemical– protein interactions. First, an input sequence is represented with pre-trained language model and an attention mechanism is used to uncover contribution degree of different words to entity relations and potential semantic information. Secondly, key features are extracted by a multi-layer stacked Bidirectional Gated Recurrent Units (Bi-GRU)-normalization residual network module to resolve higherorder dependencies while overcoming network degradation. Finally, the representation is introduced to be enhanced by external knowledge regarding the chemical structure characteristics of the compound external knowledge Results: Excellent experimental results show that our stacked integration model combines the advantages of Bi-GRU, normalization methods, and external knowledge to improve the performance of the model by complementing each other Conclusion: Our proposed model shows good performance in chemical-protein interaction extraction, and it can be used as a useful complement to biological experiments to identify chemical-protein interactions.

Quantitative Framework for Bench-to-Bedside Cancer Research.

Bioscience is an interdisciplinary venture. Driven by a quantum shift in the volume of high throughput data and in ready availability of data-intensive technologies, mathematical and quantitative approaches have become increasingly common in bioscience. For instance, a recent shift towards a quantitative description of cells and phenotypes, which is supplanting conventional qualitative descriptions, has generated immense promise and opportunities in the field of bench-to-bedside cancer OMICS, chemical biology and pharmacology. Nevertheless, like any burgeoning field, there remains a lack of shared and standardized framework for quantitative cancer research. Here, in the context of cancer, we present a basic framework and guidelines for bench-to-bedside quantitative research and therapy. We outline some of the basic concepts and their parallel use cases for chemical-protein interactions. Along with several recommendations for assay setup and conditions, we also catalog applications of these quantitative techniques in some of the most widespread discovery pipeline and analytical methods in the field. We believe adherence to these guidelines will improve experimental design, reduce variabilities and standardize quantitative datasets.