Predicting metabolite-disease associations based on auto-encoder and non-negative matrix factorization.

Metabolism refers to a series of orderly chemical reactions used to maintain life activities in organisms. In healthy individuals, metabolism remains within a normal range. However, specific diseases can lead to abnormalities in the levels of certain metabolites, causing them to either increase or decrease. Detecting these deviations in metabolite levels can aid in diagnosing a disease. Traditional biological experiments often rely on a lot of manpower to do repeated experiments, which is time consuming and labor intensive. To address this issue, we develop a deep learning model based on the auto-encoder and non-negative matrix factorization named as MDA-AENMF to predict the potential associations between metabolites and diseases. We integrate a variety of similarity networks and then acquire the characteristics of both metabolites and diseases through three specific modules. First, we get the disease characteristics from the five-layer auto-encoder module. Later, in the non-negative matrix factorization module, we extract both the metabolite and disease characteristics. Furthermore, the graph attention auto-encoder module helps us obtain metabolite characteristics. After obtaining the features from three modules, these characteristics are merged into a single, comprehensive feature vector for each metabolite-disease pair. Finally, we send the corresponding feature vector and label to the multi-layer perceptron for training. The experiment demonstrates our area under the receiver operating characteristic curve of 0.975 and area under the precision-recall curve of 0.973 in 5-fold cross-validation, which are superior to those of existing state-of-the-art predictive methods. Through case studies, most of the new associations obtained by MDA-AENMF have been verified, further highlighting the reliability of MDA-AENMF in predicting the potential relationships between metabolites and diseases.

Predicting potential interactions between lncRNAs and proteins via combined graph auto-encoder methods.

Long noncoding RNA (lncRNA) is a kind of noncoding RNA with a length of more than 200 nucleotide units. Numerous research studies have proven that although lncRNAs cannot be directly translated into proteins, lncRNAs still play an important role in human growth processes by interacting with proteins. Since traditional biological experiments often require a lot of time and material costs to explore potential lncRNA-protein interactions (LPI), several computational models have been proposed for this task. In this study, we introduce a novel deep learning method known as combined graph auto-encoders (LPICGAE) to predict potential human LPIs. First, we apply a variational graph auto-encoder to learn the low dimensional representations from the high-dimensional features of lncRNAs and proteins. Then the graph auto-encoder is used to reconstruct the adjacency matrix for inferring potential interactions between lncRNAs and proteins. Finally, we minimize the loss of the two processes alternately to gain the final predicted interaction matrix. The result in 5-fold cross-validation experiments illustrates that our method achieves an average area under receiver operating characteristic curve of 0.974 and an average accuracy of 0.985, which is better than those of existing six state-of-the-art computational methods. We believe that LPICGAE can help researchers to gain more potential relationships between lncRNAs and proteins effectively.

Predicting drug-induced liver injury using graph attention mechanism and molecular fingerprints.

Drug-induced liver injury (DILI) is a significant issue in drug development and clinical treatment due to its potential to cause liver dysfunction or damage, which, in severe cases, can lead to liver failure or even fatality. DILI has numerous pathogenic factors, many of which remain incompletely understood. Consequently, it is imperative to devise methodologies and tools for anticipatory assessment of DILI risk in the initial phases of drug development. In this study, we present DMFPGA, a novel deep learning predictive model designed to predict DILI. To provide a comprehensive description of molecular properties, we employ a multi-head graph attention mechanism to extract features from the molecular graphs, representing characteristics at the level of compound nodes. Additionally, we combine multiple fingerprints of molecules to capture features at the molecular level of compounds. The fusion of molecular fingerprints and graph features can more fully express the properties of compounds. Subsequently, we employ a fully connected neural network to classify compounds as either DILI-positive or DILI-negative. To rigorously evaluate DMFPGA's performance, we conduct a 5-fold cross-validation experiment. The obtained results demonstrate the superiority of our method over four existing state-of-the-art computational approaches, exhibiting an average AUC of 0.935 and an average ACC of 0.934. We believe that DMFPGA is helpful for early-stage DILI prediction and assessment in drug development.

A deep learning method for predicting metabolite-disease associations via graph neural network.

Metabolism is the process by which an organism continuously replaces old substances with new substances. It plays an important role in maintaining human life, body growth and reproduction. More and more researchers have shown that the concentrations of some metabolites in patients are different from those in healthy people. Traditional biological experiments can test some hypotheses and verify their relationships but usually take a considerable amount of time and money. Therefore, it is urgent to develop a new computational method to identify the relationships between metabolites and diseases. In this work, we present a new deep learning algorithm named as graph convolutional network with graph attention network (GCNAT) to predict the potential associations of disease-related metabolites. First, we construct a heterogeneous network based on known metabolite-disease associations, metabolite-metabolite similarities and disease-disease similarities. Metabolite and disease features are encoded and learned through the graph convolutional neural network. Then, a graph attention layer is used to combine the embeddings of multiple convolutional layers, and the corresponding attention coefficients are calculated to assign different weights to the embeddings of each layer. Further, the prediction result is obtained by decoding and scoring the final synthetic embeddings. Finally, GCNAT achieves a reliable area under the receiver operating characteristic curve of 0.95 and the precision-recall curve of 0.405, which are better than the results of existing five state-of-the-art predictive methods in 5-fold cross-validation, and the case studies show that the metabolite-disease correlations predicted by our method can be successfully demonstrated by relevant experiments. We hope that GCNAT could be a useful biomedical research tool for predicting potential metabolite-disease associations in the future.

Predicting the potential human lncRNA-miRNA interactions based on graph convolution network with conditional random field.

Long non-coding RNA (lncRNA) and microRNA (miRNA) are two typical types of non-coding RNAs (ncRNAs), their interaction plays an important regulatory role in many biological processes. Exploring the interactions between unknown lncRNA and miRNA can help us better understand the functional expression between lncRNA and miRNA. At present, the interactions between lncRNA and miRNA are mainly obtained through biological experiments, but such experiments are often time-consuming and labor-intensive, it is necessary to design a computational method that can predict the interactions between lncRNA and miRNA. In this paper, we propose a method based on graph convolutional neural (GCN) network and conditional random field (CRF) for predicting human lncRNA-miRNA interactions, named GCNCRF. First, we construct a heterogeneous network using the known interactions of lncRNA and miRNA in the LncRNASNP2 database, the lncRNA/miRNA integration similarity network, and the lncRNA/miRNA feature matrix. Second, the initial embedding of nodes is obtained using a GCN network. A CRF set in the GCN hidden layer can update the obtained preliminary embeddings so that similar nodes have similar embeddings. At the same time, an attention mechanism is added to the CRF layer to reassign weights to nodes to better grasp the feature information of important nodes and ignore some nodes with less influence. Finally, the final embedding is decoded and scored through the decoding layer. Through a 5-fold cross-validation experiment, GCNCRF has an area under the receiver operating characteristic curve value of 0.947 on the main dataset, which has higher prediction accuracy than the other six state-of-the-art methods.

Designing high thermally stable deep red phosphors based on low thermal expansion coefficients for optical applications.

Mn4+-activated oxide phosphors with low cost and unique luminescent properties have been considered as a promising candidate for various optical applications, while the search for high thermal stable red-emitting phosphors is still a huge challenge. In our work, we find and unveil the relationship between luminescence thermal quenching behavior and thermal expansion coefficients (α/10-6 K-1) based on double-perovskite niobate phosphors Ca2LnNbO6:Mn4+ (Ln3+ = Y3+, Gd3+, La3+, or Lu3+). It can be concluded that the phosphors with low thermal expansion coefficients contribute to high thermal stability. Subsequently, Ca2LuNbO6:Mn4+ accomplishes accurate temperature testing and high-CRI white light-emitting diodes. Thus, a thermal expansion coefficient strategy is a new guide to select the appropriate substrate with high thermal stability for an Mn4+-activated emitter.

MPFFPSDC: A multi-pooling feature fusion model for predicting synergistic drug combinations.

Drug combination therapies are common practice in the treatment of cancer, but not all combinations result in synergy. As traditional screening approaches are restricted in their ability to uncover synergistic drug combinations, computer-aided medicine is becoming a increasingly prevalent in this field. In this work, a predictive model of potential interactions between drugs named MPFFPSDC is presented, which can maintain the symmetry of drug inputs and eliminate inconsistencies in predictive results caused by different drug inputting sequences or positions. The experimental results show that MPFFPSDC outperforms comparative models in major performance indicators and exhibits better generalization for independent data. Furthermore, the case study demonstrates that our model can capture molecular substructures that contribute to the synergistic effect of two drugs. These results indicate that MPFFPSDC not only offers strong predictive performance, but also has good model interpretability that may provide new insights for the study of drug interaction mechanisms and the development of new drugs.

DCAMCP: A deep learning model based on capsule network and attention mechanism for molecular carcinogenicity prediction.

The carcinogenicity of drugs can have a serious impact on human health, so carcinogenicity testing of new compounds is very necessary before being put on the market. Currently, many methods have been used to predict the carcinogenicity of compounds. However, most methods have limited predictive power and there is still much room for improvement. In this study, we construct a deep learning model based on capsule network and attention mechanism named DCAMCP to discriminate between carcinogenic and non-carcinogenic compounds. We train the DCAMCP on a dataset containing 1564 different compounds through their molecular fingerprints and molecular graph features. The trained model is validated by fivefold cross-validation and external validation. DCAMCP achieves an average accuracy (ACC) of 0.718 ± 0.009, sensitivity (SE) of 0.721 ± 0.006, specificity (SP) of 0.715 ± 0.014 and area under the receiver-operating characteristic curve (AUC) of 0.793 ± 0.012. Meanwhile, comparable results can be achieved on an external validation dataset containing 100 compounds, with an ACC of 0.750, SE of 0.778, SP of 0.727 and AUC of 0.811, which demonstrate the reliability of DCAMCP. The results indicate that our model has made progress in cancer risk assessment and could be used as an efficient tool in drug design.

Toxicokinetics and Particle Number-Based Trophic Transfer of a Metallic Nanoparticle Mixture in a Terrestrial Food Chain.

Herein, we investigated to which extent metallic nanoparticles (MNPs) affect the trophic transfer of other coexisting MNPs from lettuce to terrestrial snails and the associated tissue-specific distribution using toxicokinetic (TK) modeling and single-particle inductively coupled plasma mass spectrometry. During a period of 22 days, snails were fed with lettuce leaves that were root exposed to AgNO3 (0.05 mg/L), AgNPs (0.75 mg/L), TiO2NPs (200 mg/L), and a mixture of AgNPs and TiO2NPs (equivalent doses as for single NPs). The uptake rate constants (ku) were 0.08 and 0.11 kg leaves/kg snail/d for Ag and 1.63 and 1.79 kg leaves/kg snail/d for Ti in snails fed with NPs single- and mixture-exposed lettuce, respectively. The elimination rate constants (ke) of Ag in snails exposed to single AgNPs and mixed AgNPs were comparable to the corresponding ku, while the ke for Ti were lower than the corresponding ku. As a result, single TiO2NP treatments as well as exposure to mixtures containing TiO2NPs induced significant biomagnification from lettuce to snails with kinetic trophic transfer factors (TTFk) of 7.99 and 6.46. The TTFk of Ag in the single AgNPs treatment (1.15 kg leaves/kg snail) was significantly greater than the TTFk in the mixture treatment (0.85 kg leaves/kg snail), while the fraction of Ag remaining in the body of snails after AgNPs exposure (36%) was lower than the Ag fraction remaining after mixture exposure (50%). These results indicated that the presence of TiO2NPs inhibited the trophic transfer of AgNPs from lettuce to snails but enhanced the retention of AgNPs in snails. Biomagnification of AgNPs from lettuce to snails was observed in an AgNPs single treatment using AgNPs number as the dose metric, which was reflected by the particle number-based TTFs of AgNPs in snails (1.67, i.e., higher than 1). The size distribution of AgNPs was shifted across the lettuce-snail food chain. By making use of particle-specific measurements and fitting TK processes, this research provides important implications for potential risks associated with the trophic transfer of MNP mixtures.

Recent advances on the machine learning methods in predicting ncRNA-protein interactions.

Recent transcriptomics and bioinformatics studies have shown that ncRNAs can affect chromosome structure and gene transcription, participate in the epigenetic regulation, and take part in diseases such as tumorigenesis. Biologists have found that most ncRNAs usually work by interacting with the corresponding RNA-binding proteins. Therefore, ncRNA-protein interaction is a very popular study in both the biological and medical fields. However, due to the limitations of manual experiments in the laboratory, machine-learning methods for predicting ncRNA-protein interactions are increasingly favored by the researchers. In this review, we summarize several machine learning predictive models of ncRNA-protein interactions over the past few years, and briefly describe the characteristics of these machine learning models. In order to optimize the performance of machine learning models to better predict ncRNA-protein interactions, we give some promising future computational directions at the end.

Fine-scale empirical data on niche divergence and homeolog expression patterns in an allopolyploid and its diploid progenitor species.

Polyploidization is pervasive in plants, but little is known about the niche divergence of wild allopolyploids (species that harbor polyploid genomes originating from different diploid species) relative to their diploid progenitor species and the gene expression patterns that may underlie such ecological divergence. We conducted a fine-scale empirical study on habitat and gene expression of an allopolyploid and its diploid progenitors. We quantified soil properties and light availability of habitats of an allotetraploid Cardamine flexuosa and its diploid progenitors Cardamine amara and Cardamine hirsuta in two seasons. We analyzed expression patterns of genes and homeologs (homeologous gene copies in allopolyploids) using RNA sequencing. We detected niche divergence between the allopolyploid and its diploid progenitors along water availability gradient at a fine scale: the diploids in opposite extremes and the allopolyploid in a broader range between diploids, with limited overlap with diploids at both ends. Most of the genes whose homeolog expression ratio changed among habitats in C. flexuosa varied spatially and temporally. These findings provide empirical evidence for niche divergence between an allopolyploid and its diploid progenitor species at a fine scale and suggest that divergent expression patterns of homeologs in an allopolyploid may underlie its persistence in diverse habitats.

Frequency Domain Analysis of Partial-Tensor Rotating Accelerometer Gravity Gradiometer.

The output model of a rotating accelerometer gravity gradiometer (RAGG) established by the inertial dynamics method cannot reflect the change of signal frequency, and calibration sensitivity and self-gradient compensation effect for the RAGG is a very important stage in the development process that cannot be omitted. In this study, a model based on the outputs of accelerometers on the disc of RGAA is established to calculate the gravity gradient corresponding to the distance, through the study of the RAGG output influenced by a surrounding mass in the frequency domain. Taking particle, sphere, and cuboid as examples, the input-output models of gravity gradiometer are established based on the center gradient and four accelerometers, respectively. Simulation results show that, if the scale factors of the four accelerometers on the disk are the same, the output signal of the RAGG only contains (4k+2)ω (ω is the spin frequency of disc for RAGG) harmonic components, and its amplitude is related to the orientation of the surrounding mass. Based on the results of numerical simulation of the three models, if the surrounding mass is close to the RAGG, the input-output models of gravity gradiometer are more accurate based on the four accelerometers. Finally, some advantages and disadvantages of cuboid and sphere are compared and some suggestions related to calibration and self-gradient compensation are given.

Enhanced emissions of brominated flame retardants from indoor sources by direct contact with dust.

The emissions of brominated flame retardants (BFRs) from consumer products have been considered the major to the ubiquitous occurrence of contaminants in indoor environments. Direct contact with dust covering the surface of source materials in a real environment could introduce significant uncertainty. This study investigated the effects of dust coverage on the emissions of four BFRs, including 1, 2, 5, 6, 9, and 10-hexabromocyclododecane (HBCD), bis(2-ethyl-1-hexyl) tetrabromophthalate (BEHTBP), tetrabromobisphenol A (TBBPA), and hexabromobenzene (HBBZ), from decorative laminate, cotton sound insulation, PVC floor, and carpet. Direct contact with dust was confirmed to increase the total emissions by 30.8-98.1% compared with the emissions in the non-dust group. The emissions of HBCD, TBBPA, and HBBZ from cotton sound insulation were obviously enhanced by dust with smaller particles but did not linearly increase along with the dust amounts. Thus, these findings have practical implications in that the frequent removal of dust could be important to minimize the exposure risk from indoor emissions of BFRs.

Preparation of Fe-Co based MOF-74 and its effective adsorption of arsenic from aqueous solution.

To obtain a cost-effective adsorbent for the removal of arsenic in water, a novel nanostructured Fe-Co based metal organic framework (MOF-74) adsorbent was successfully prepared via a simple solvothermal method. The adsorption experiments showed that the optimal molar ratio of Fe/Co in the adsorbent was 2:1. The Fe2Co1 MOF-74 was characterized by various techniques and the results showed that the nanoparticle diameter ranged from 60 to 80 nm and the specific surface area was 147.82 m2/g. The isotherm and kinetic parameters of arsenic removal on Fe2Co1 MOF-74 were well-fitted by the Langmuir and pseudo-second-order models. The maximum adsorption capacities toward As(III) and As(V) were 266.52 and 292.29 mg/g, respectively. The presence of sulfate, carbonate and humic acid had no obvious effect on arsenic adsorption. However, coexisting phosphate significantly hindered the removal of arsenic, especially at high concentrations (10 mmol/L). Electrostatic interaction and hydroxyl and metal-oxygen groups played important roles in the adsorption of arsenic. Furthermore, the prepared adsorbent had stable adsorption ability after regeneration and when used in a real-water matrix. The excellent adsorption performance of Fe2Co1 MOF-74 material makes it a potentially promising adsorbent for the removal of arsenic.

Silhouette Scores for Arbitrary Defined Groups in Gene Expression Data and Insights into Differential Expression Results.

BACKGROUND: Hierarchical Sample clustering (HSC) is widely performed to examine associations within expression data obtained from microarrays and RNA sequencing (RNA-seq). Researchers have investigated the HSC results with several possible criteria for grouping (e.g., sex, age, and disease types). However, the evaluation of arbitrary defined groups still counts in subjective visual inspection. RESULTS: To objectively evaluate the degree of separation between groups of interest in the HSC dendrogram, we propose to use Silhouette scores. Silhouettes was originally developed as a graphical aid for the validation of data clusters. It provides a measure of how well a sample is classified when it was assigned to a cluster by according to both the tightness of the clusters and the separation between them. It ranges from 1.0 to - 1.0, and a larger value for the average silhouette (AS) over all samples to be analyzed indicates a higher degree of cluster separation. The basic idea to use an AS is to replace the term cluster by group when calculating the scores. We investigated the validity of this score using simulated and real data designed for differential expression (DE) analysis. We found that larger (or smaller) AS values agreed well with both higher (or lower) degrees of separation between different groups and higher percentages of differentially expressed genes (PDEG). We also found that the AS values were generally independent on the number of replicates (Nrep). Although the PDEG values depended on Nrep, we confirmed that both AS and PDEG values were close to zero when samples in the data showed an intermingled nature between the groups in the HSC dendrogram. CONCLUSION: Silhouettes is useful for exploring data with predefined group labels. It would help provide both an objective evaluation of HSC dendrograms and insights into the DE results with regard to the compared groups.

Evaluation of methods for differential expression analysis on multi-group RNA-seq count data.

BACKGROUND: RNA-seq is a powerful tool for measuring transcriptomes, especially for identifying differentially expressed genes or transcripts (DEGs) between sample groups. A number of methods have been developed for this task, and several evaluation studies have also been reported. However, those evaluations so far have been restricted to two-group comparisons. Accumulations of comparative studies for multi-group data are also desired. METHODS: We compare 12 pipelines available in nine R packages for detecting differential expressions (DE) from multi-group RNA-seq count data, focusing on three-group data with or without replicates. We evaluate those pipelines on the basis of both simulation data and real count data. RESULTS: As a result, the pipelines in the TCC package performed comparably to or better than other pipelines under various simulation scenarios. TCC implements a multi-step normalization strategy (called DEGES) that internally uses functions provided by other representative packages (edgeR, DESeq2, and so on). We found considerably different numbers of identified DEGs (18.5 ~ 45.7% of all genes) among the pipelines for the same real dataset but similar distributions of the classified expression patterns. We also found that DE results can roughly be estimated by the hierarchical dendrogram of sample clustering for the raw count data. CONCLUSION: We confirmed the DEGES-based pipelines implemented in TCC performed well in a three-group comparison as well as a two-group comparison. We recommend using the DEGES-based pipeline that internally uses edgeR (here called the EEE-E pipeline) for count data with replicates (especially for small sample sizes). For data without replicates, the DEGES-based pipeline with DESeq2 (called SSS-S) can be recommended.

Uptake and Metabolism of Phthalate Esters by Edible Plants.

Phthalate esters (PAEs) are large-volume chemicals and are found ubiquitously in soil as a result of widespread plasticulture and waste disposal. Food plants such as vegetables may take up and accumulate PAEs from soil, potentially imposing human health risks through dietary intake. In this study, we carried out a cultivation study using lettuce, strawberry, and carrot plants to determine the potential of plant uptake, translocation, and metabolism of di-n-butyl phthalate (DnBP) and di(2-ethylhexyl) phthalate (DEHP) and their primary metabolites mono-n-butyl phthalate (MnBP) and mono(2-ethylhexyl) phthalate (MEHP). All four compounds were detected in the plant tissues, with the bioconcentration factors (BCFs) ranging from 0.16 ± 0.01 to 4.78 ± 0.59. However, the test compounds were poorly translocated from roots to leaves, with a translocation factor below 1. Further, PAEs were readily transformed to their monoesters following uptake. Incubation of PAEs and monoalkyl phthalate esters (MPEs) in carrot cell culture showed that DnBP was hydrolyzed more rapidly than DEHP, while the monoesters were transformed more quickly than their parent precursors. Given the extensive metabolism of PAEs to monoesters in both whole plants and plant cells, metabolism intermediates such as MPEs should be considered when assessing human exposure via dietary intake of food produced from PAE-contaminated soils.

Enhancement of tunability of MAPK cascade due to coexistence of processive and distributive phosphorylation mechanisms.

The processive phosphorylation mechanism becomes important when there is macromolecular crowding in the cytoplasm. Integrating the processive phosphorylation mechanism with the traditional distributive one, we propose a mixed dual-site phosphorylation (MDP) mechanism in a single-layer phosphorylation cycle. Further, we build a degree model by applying the MDP mechanism to a three-layer mitogen-activated protein kinase (MAPK) cascade. By bifurcation analysis, our study suggests that the crowded-environment-induced pseudoprocessive mechanism can qualitatively change the response of this biological network. By adjusting the degree of processivity in our model, we find that the MAPK cascade is able to switch between the ultrasensitivity, bistability, and oscillatory dynamical states. Sensitivity analysis shows that the theoretical results remain unchanged within a reasonably chosen variation of parameter perturbation. By scaling the reaction rates and also introducing new connections into the kinetic scheme, we further construct a proportion model of the MAPK cascade to validate our findings. Finally, it is illustrated that the spatial propagation of the activated MAPK signal can be improved (or attenuated) by increasing the degree of processivity of kinase (or phosphatase). Our research implies that the MDP mechanism makes the MAPK cascade become a flexible signal module, and the coexistence of processive and distributive phosphorylation mechanisms enhances the tunability of the MAPK cascade.

Visualizing an ultra-weak protein-protein interaction in phosphorylation signaling.

Proteins interact with each other to fulfill their functions. The importance of weak protein-protein interactions has been increasingly recognized. However, owing to technical difficulties, ultra-weak interactions remain to be characterized. Phosphorylation can take place via a K(D)≈25 mM interaction between two bacterial enzymes. Using paramagnetic NMR spectroscopy and with the introduction of a novel Gd(III)-based probe, we determined the structure of the resulting complex to atomic resolution. The structure accounts for the mechanism of phosphoryl transfer between the two enzymes and demonstrates the physical basis for their ultra-weak interaction. Further, molecular dynamics (MD) simulations suggest that the complex has a lifetime in the micro- to millisecond regimen. Hence such interaction is termed a fleeting interaction. From mathematical modeling, we propose that an ultra-weak fleeting interaction enables rapid flux of phosphoryl signal, providing a high effective protein concentration.