Search

Diffuse large B-cell lymphoma (DLBCL) is the most common subtype of non-Hodgkin lymphoma (NHL) and is characterized by high heterogeneity. Assessment of its prognosis and genetic subtyping hold significant clinical implications. However, existing DLBCL prognostic models are mainly based on transcriptomic profiles, while genetic variation detection is more commonly used in clinical practice. In addition, current clustering-based subtyping methods mostly focus on genes with high mutation frequencies, providing insufficient explanations for the heterogeneity of DLBCL. Here, we proposed VNNSurv (https://bio-web1.nscc-gz.cn/app/VNNSurv), a survival model for DLBCL patients based on a biologically informed visible neural network (VNN). VNNSurv achieved an average C-index of 0.72 on the cross-validation set (HMRN cohort, n = 928), outperforming the baseline methods. The remarkable interpretability of VNNSurv facilitated the identification of the most impactful genes and the underlying pathways through which they act on patient outcomes. When only the 30 highest-impact genes were used as genetic input, the overall performance of VNNSurv improved, and a C-index of 0.70 was achieved on the external TCGA cohort (n = 48). Leveraging these high-impact genes, including 16 genes with low (

Abstract Image-based spatial transcriptomic sequencing technologies have enabled the measurement of gene expression at single-cell resolution, but with a limited number of genes. Current computational approaches attempt to overcome these limitations by imputing missing genes, but face challenges regarding prediction accuracy and identification of cell populations due to the neglect of gene-gene relationships. In this context, we present stImpute, a method to impute spatial transcriptomics according to reference scRNA-seq data based on the gene network constructed from the protein language model ESM-2. Specifically, stImpute employs an autoencoder to create gene expression embeddings for both spatial transcriptomics and scRNA-seq data, which are used to identify the nearest neighboring cells between scRNA-seq and spatial transcriptomics datasets. According to the neighbored cells, the gene expressions of spatial transcriptomics cells are imputed through a graph neural network, where nodes are genes, and edges are based on cosine similarity between the ESM-2 embeddings of the gene-encoding proteins. The gene prediction uncertainty is further measured through a deep learning model. stImpute was shown to consistently outperform state-of-the-art methods across multiple datasets concerning imputation and clustering. stImpute also demonstrates robustness in producing consistent results that are insensitive to model parameters.

Abstract Cryo-electron microscopy (cryo-EM) technique is widely used for protein structure determination. Current automatic cryo-EM protein complex modeling methods mostly rely on prior chain separation. However, chain separation without sequence guidance often suffers from errors caused by cross-chain interaction or noise densities, which would accumulate and mislead the subsequent steps. Here, we present EModelX, a fully automated cryo-EM protein complex structure modeling method, which achieves sequence-guiding modeling through cross-modal alignments between cryo-EM maps and protein sequences. EModelX first employs multi-task deep learning to predict Cα atoms, backbone atoms, and amino acid types from cryo-EM maps, which is subsequently used to sample Cα traces with amino acid profiles. The profiles are then aligned with protein sequences to obtain initial structural models, which yielded an average RMSD of 1.17 Å in our test set, approaching atomic-level precision in recovering PDB-deposited structures. After filling unmodeled gaps through sequence-guiding Cα threading, the final models achieved an average TM-score of 0.808, outperforming the state-of-the-art method. The further combination with AlphaFold can improve the average TM-score to 0.911. Analyzes conducted by comparing some EModelX-built models and PDB structures highlight its potential to improve PDB structures. EModelX is accessible at https://bio-web1.nscc-gz.cn/app/EModelX .

Abstract Despite recent advancement in 3D molecule conformation generation driven by diffusion models, its high computational cost in iterative diffusion/denoising process limits its application. Here, an equivariant consistency model (EC-Conf) was proposed as a fast diffusion method for low-energy conformation generation. In EC-Conf, a modified SE (3)-equivariant transformer model was directly used to encode the Cartesian molecular conformations and a highly efficient consistency diffusion process was carried out to generate molecular conformations. It was demonstrated that, with only one sampling step, it can already achieve comparable quality to other diffusion-based models running with thousands denoising steps. Its performance can be further improved with a few more sampling iterations. The performance of EC-Conf is evaluated on both GEOM-QM9 and GEOM-Drugs sets. Our results demonstrate that the efficiency of EC-Conf for learning the distribution of low energy molecular conformation is at least two magnitudes higher than current SOTA diffusion models and could potentially become a useful tool for conformation generation and sampling. Scientific Contributions In this work, we proposed an equivariant consistency model that significantly improves the efficiency of conformation generation in diffusion-based models while maintaining high structural quality. This method serves as a general framework and can be further extended to more complex structure generation and prediction tasks, including those involving proteins, in future steps.

Abstract Enzymes are crucial in numerous biological processes, with the Enzyme Commission (EC) number being a commonly used method for defining enzyme function. However, current EC number prediction technologies have not fully recognized the importance of enzyme active sites and structural characteristics. Here, we propose GraphEC, a geometric graph learning-based EC number predictor using the ESMFold-predicted structures and a pre-trained protein language model. Specifically, we first construct a model to predict the enzyme active sites, which is utilized to predict the EC number. The prediction is further improved through a label diffusion algorithm by incorporating homology information. In parallel, the optimum pH of enzymes is predicted to reflect the enzyme-catalyzed reactions. Experiments demonstrate the superior performance of our model in predicting active sites, EC numbers, and optimum pH compared to other state-of-the-art methods. Additional analysis reveals that GraphEC is capable of extracting functional information from protein structures, emphasizing the effectiveness of geometric graph learning. This technology can be used to identify unannotated enzyme functions, as well as to predict their active sites and optimum pH, with the potential to advance research in synthetic biology, genomics, and other fields.

Non-starch polysaccharides are major bioactive components in chestnuts, and can serve as water-soluble polysaccharides with potential prebiotic properties. This study aims to establish an in vitro digestion and fermentation model to reveal the digestive and fermentative characteristics of Non-starch polysaccharides from chestnut kernels (NSPCK). The results indicated that under simulated digestion, NSPCK was partially digested in gastric juice but remained significantly undigested in saliva and intestinal juice, demonstrating considerable resilience against hydrolysis. After digestion, NSPCK still exhibited stable rough, lamellar, and porous structure and maintained strong antioxidant capacity. Animal experiments revealed positive effects of NSPCK on blood lipid level, and colon tissue of mice. Moreover, NSPCK enhanced the accumulation of short-chain fatty acids during fermentation, particularly acetic acid, propionic acid, and butyric acid. Furthermore, NSPCK intervention increased the abundance of beneficial bacteria such as Lactobacillus and Bifidobacterium, and at the same time reduced that of harmful bacteria such as Enterococcus.

Abstract Protein functions are characterized by interactions with proteins, drugs, and other biomolecules. Understanding these interactions is essential for deciphering the molecular mechanisms underlying biological processes and developing new therapeutic strategies. Current computational methods mostly predict interactions based on either molecular network or structural information, without integrating them within a unified multi-scale framework. While a few multi-view learning methods are devoted to fusing the multi-scale information, these methods tend to rely intensively on a single scale and under-fitting the others, likely attributed to the imbalanced nature and inherent greediness of multi-scale learning. To alleviate the optimization imbalance, we present MUSE, a multi-scale representation learning framework based on a variant expectation maximization to optimize different scales in an alternating procedure over multiple iterations. This strategy efficiently fuses multi-scale information between atomic structure and molecular network scale through mutual supervision and iterative optimization. MUSE outperforms the current state-of-the-art models not only in molecular interaction (protein-protein, drug-protein, and drug-drug) tasks but also in protein interface prediction at the atomic structure scale. More importantly, the multi-scale learning framework shows potential for extension to other scales of computational drug discovery.

Abstract Constructing discriminative representations of molecules lies at the core of a number of domains such as drug discovery, chemistry, and medicine. State‐of‐the‐art methods employ graph neural networks and self‐supervised learning (SSL) to learn unlabeled data for structural representations, which can then be fine‐tuned for downstream tasks. Albeit powerful, these methods are pre‐trained solely on molecular structures and thus often struggle with tasks involved in intricate biological processes. Here, it is proposed to assist the learning of molecular representation by using the perturbed high‐content cell microscopy images at the phenotypic level. To incorporate the cross‐modal pre‐training, a unified framework is constructed to align them through multiple types of contrastive loss functions, which is proven effective in the formulated novel tasks to retrieve the molecules and corresponding images mutually. More importantly, the model can infer functional molecules according to cellular images generated by genetic perturbations. In parallel, the proposed model can transfer non‐trivially to molecular property predictions, and has shown great improvement over clinical outcome predictions. These results suggest that such cross‐modality learning can bridge molecules and phenotype to play important roles in drug discovery.

Abstract Background Predicting outcome of breast cancer is important for selecting appropriate treatments and prolonging the survival periods of patients. Recently, different deep learning-based methods have been carefully designed for cancer outcome prediction. However, the application of these methods is still challenged by interpretability. In this study, we proposed a novel multitask deep neural network called UISNet to predict the outcome of breast cancer. The UISNet is able to interpret the importance of features for the prediction model via an uncertainty-based integrated gradients algorithm. UISNet improved the prediction by introducing prior biological pathway knowledge and utilizing patient heterogeneity information. Results The model was tested in seven public datasets of breast cancer, and showed better performance (average C-index = 0.691) than the state-of-the-art methods (average C-index = 0.650, ranged from 0.619 to 0.677). Importantly, the UISNet identified 20 genes as associated with breast cancer, among which 11 have been proven to be associated with breast cancer by previous studies, and others are novel findings of this study. Conclusions Our proposed method is accurate and robust in predicting breast cancer outcomes, and it is an effective way to identify breast cancer-associated genes. The method codes are available at: https://github.com/chh171/UISNet .

The aim of this work was to apply response surface methodology (RSM) to model and optimize the accelerated solvent extraction (ASE) technique for extracting the crude polysaccharides (GLP-K) from Gracilaria lemaneiformis. In terms of the yield of polysaccharide, single factor tests and Box-Behnken design response surface method were developed. The structure of the prepared polysaccharides was characterized by Fourier transform infrared (FT-IR) and high performance liquid chromatography (HPLC). And the anti-inflammatory potential of GLP-K in lipopolysaccharide (LPS)-induced RAW264.7 macrophages was also explored. Briefly, the optimal extraction conditions for GLP-K were as follows: 70 °C of extraction temperature, 8.5 min of extraction time, and 4 extraction cycles. Under the condition, the experimental yield of polysaccharide was 9.58%±0.31%. FT-IR showed that the polysaccharide contained uronic acid, and the weight-average molecular weight ranged from 4.4 to 747.1 kDa. GLP-K had no significant cytotoxic effects at or below the concentrations of 1000 μg/mL (P

Recent studies suggest RNAs act as promising drug targets. However, limited development has been achieved in detecting RNA–ligand interactions. To guide the discovery of RNA-binding ligands, it is necessary to characterize them comprehensively, especially in the binding specificity, binding affinity and drug-like properties. We established a database, RNALID (http://biomed.nscc-gz.cn/RNALID/html/index.html#/database), which collects RNA–ligand interactions validated by low-throughput experiment. RNALID contains 358 RNA–ligand interactions. Comparing to the fellow database, 94.5% of ligands in RNALID are completely or partially novel collections, and 51.78% have novel two-dimensional (2D) structures. Through the analysis of ligand structure, binding affinity and cheminformatic parameters we found that multivalent (MV) ligands mainly binding to RNA repeats are more structurally conserved in both 2D and 3D structures than other ligand types, exhibit higher binding specificity and binding affinity than ligands binding to non-repeat RNAs, but deviate far from the Lipinski’s rule of five. In contrary, small molecule (SM) ligands binding to virus RNA exhibit higher affinity and more resemble protein-ligands, but potentially possess low binding specificity. Further analysis on 28 detailed drug-likeness properties indicated that RNA-ligands’ development need to balance between the binding affinity and the drug-likeness because of the significant linear co-relationship between the two. Comparing RNALID ligands to FDA-approved drugs and ligands without bioactivity indicated that RNA-binding ligands are different from them in chemical properties, structural properties and drug-likeness. Thus, characterizing the RNA–ligand interactions in RNALID in multiple respects provides new insights into discovering and designing druggable ligands binding with RNA.

Abstract Background With the development of cancer precision medicine, a huge amount of high-dimensional cancer information has rapidly accumulated regarding gene alterations, diseases, therapeutic interventions and various annotations. The information is highly fragmented across multiple different sources, making it highly challenging to effectively utilize and exchange the information. Therefore, it is essential to create a resource platform containing well-aggregated, carefully mined, and easily accessible data for effective knowledge sharing. Methods In this study, we have developed “Consensus Cancer Core” (Tri©DB), a new integrative cancer precision medicine knowledgebase and reporting system by mining and harmonizing multifaceted cancer data sources, and presenting them in a centralized platform with enhanced functionalities for accessibility, annotation and analysis. Results The knowledgebase provides the currently most comprehensive information on cancer precision medicine covering more than 40 annotation entities, many of which are novel and have never been explored previously. Tri©DB offers several unique features: (i) harmonizing the cancer-related information from more than 30 data sources into one integrative platform for easy access; (ii) utilizing a variety of data analysis and graphical tools for enhanced user interaction with the high-dimensional data; (iii) containing a newly developed reporting system for automated annotation and therapy matching for external patient genomic data. Benchmark test indicated that Tri©DB is able to annotate 46% more treatments than two officially recognized resources, oncoKB and MCG. Tri©DB was further shown to have achieved 94.9% concordance with administered treatments in a real clinical trial. Conclusions The novel features and rich functionalities of the new platform will facilitate full access to cancer precision medicine data in one single platform and accommodate the needs of a broad range of researchers not only in translational medicine, but also in basic biomedical research. We believe that it will help to promote knowledge sharing in cancer precision medicine. Tri©DB is freely available at www.biomeddb.org , and is hosted on a cutting-edge technology architecture supporting all major browsers and mobile handsets.

Revealing protein binding sites with other molecules, such as nucleic acids, peptides, or small ligands, sheds light on disease mechanism elucidation and novel drug design. With the explosive growth of proteins in sequence databases, how to accurately and efficiently identify these binding sites from sequences becomes essential. However, current methods mostly rely on expensive multiple sequence alignments or experimental protein structures, limiting their genome-scale applications. Besides, these methods haven’t fully explored the geometry of the protein structures. Here, we propose GPSite, a multi-task network for simultaneously predicting binding residues of DNA, RNA, peptide, protein, ATP, HEM, and metal ions on proteins. GPSite was trained on informative sequence embeddings and predicted structures from protein language models, while comprehensively extracting residual and relational geometric contexts in an end-to-end manner. Experiments demonstrate that GPSite substantially surpasses state-of-the-art sequence-based and structure-based approaches on various benchmark datasets, even when the structures are not well-predicted. The low computational cost of GPSite enables rapid genome-scale binding residue annotations for over 568,000 sequences, providing opportunities to unveil unexplored associations of binding sites with molecular functions, biological processes, and genetic variants. The GPSite webserver and annotation database can be freely accessed at https://bio-web1.nscc-gz.cn/app/GPSite.

Tumor mutation burden (TMB) has emerged as an essential biomarker for assessing the efficacy of cancer immunotherapy. However, due to the inherent complexity of tumors, TMB is not always correlated with the responsiveness of immune checkpoint inhibitors (ICIs). Thus, refining the interpretation and contextualization of TMB is a requisite for enhancing clinical outcomes. In this study, we conducted a comprehensive investigation of the relationship between TMB and multi-omics data across 33 human cancer types. Our analysis revealed distinct biological changes associated with varying TMB statuses in STAD, COAD, and UCEC. While multi-omics data offer an opportunity to dissect the intricacies of tumors, extracting meaningful biological insights from such massive information remains a formidable challenge. To address this, we developed and implemented the PGLCN, a biologically informed graph neural network based on pathway interaction information. This model facilitates the stratification of patients into subgroups with distinct TMB statuses and enables the evaluation of driver biological processes through enhanced interpretability. By integrating multi-omics data for TMB prediction, our PGLCN model outperformed previous traditional machine learning methodologies, demonstrating superior TMB status prediction accuracy (STAD AUC: 0.976 ± 0.007; COAD AUC: 0.994 ± 0.007; UCEC AUC: 0.947 ± 0.023) and enhanced interpretability (BA-House: 1.0; BA-Community: 0.999; BA-Grid: 0.994; Tree-Cycles: 0.917; Tree-Grids: 0.867). Furthermore, the biological interpretability inherent to PGLCN identified the Toll-like receptor family and DNA repair pathways as potential combined biomarkers in conjunction with TMB status in gastric cancer. This finding suggests a potential synergistic targeting strategy with immunotherapy for gastric cancer, thus advancing the field of precision oncology.

Dynamic changes of the paired heavy and light chain B cell receptor (BCR) repertoire provide an essential insight into understanding the humoral immune response post-SARS-CoV-2 infection and vaccination. However, differences between the endogenous paired BCR repertoire kinetics in SARS-CoV-2 infection and previously recovered/naïve subjects treated with the inactivated vaccine remain largely unknown. We performed single-cell V(D)J sequencing of B cells from six healthy donors with three shots of inactivated SARS-CoV-2 vaccine (BBIBP-CorV), five people who received the BBIBP-CorV vaccine after having recovered from COVID-19, five unvaccinated COVID-19 recovered patients and then integrated with public data of B cells from four SARS-CoV-2-infected subjects. We discovered that BCR variable (V) genes were more prominently used in the SARS-CoV-2 exposed groups (both in the group with active infection and in the group that had recovered) than in the vaccinated groups. The VH gene that expanded the most after SARS-CoV-2 infection was IGHV3-33, while IGHV3-23 in the vaccinated groups. SARS-CoV-2-infected group enhanced more BCR clonal expansion and somatic hypermutation than the vaccinated healthy group. A small proportion of public clonotypes were shared between the SARS-CoV-2 infected, vaccinated healthy, and recovered groups. Moreover, several public antibodies had been identified against SARS-CoV-2 spike protein. We comprehensively characterize the paired heavy and light chain BCR repertoire from SARS-CoV-2 infection to vaccination, providing further guidance for the development of the next-generation precision vaccine.

Abstract Polysaccharides are structurally varied groups of biological macromolecules which broadly occurrence in nature. The structural features like detailed glycosidic linkages, monosaccharide composition, molecular weight (MW), and chain conformation could impact their functions. This paper attempts to review the current development on conformational characterization of polysaccharides and includes the influence of chain conformations on their biological activities. The chain conformations characterization methods mainly include static and dynamic light scattering, multiple angle laser light scattering (MALLS) detector, high‐performance size exclusion chromatography (HPSEC), viscosity analysis in the base of the theory of dilute polymer solution, circular dichroism (CD), Congo red test, differential scanning calorimetry (DSC), atomic force microscopy (AFM), single molecular AFM, fluorescence correlation spectroscopy (FCS), single‐molecule force spectroscopy (SMFS), nuclear magnetic resonance (NMR) spectroscopy, HPSEC‐MALLS analysis, as well as HPSEC‐MALLS‐refractive index (RI)‐viscosity detector (VS) system. This would be also helpful for development and application of polysaccharides from different natural sources.

Observational studies consistently disclose brain imaging-derived phenotypes (IDPs) as critical markers for early diagnosis of both brain disorders and cardiovascular diseases. However, it remains unclear about the shared genetic landscape between brain IDPs and the risk of brain disorders and cardiovascular diseases, restricting the applications of potential diagnostic techniques through brain IDPs. Here, we reported genetic correlations and putative causal relationships between 921 brain IDPs, 20 brain disorders and six cardiovascular diseases by leveraging their large-scale genome-wide association study (GWAS) summary statistics. Applications of Mendelian randomization (MR) identified significant putative causal effects of multiple region-specific brain IDPs in relation to the increased risks for amyotrophic lateral sclerosis (ALS), major depressive disorder (MDD), autism spectrum disorder (ASD) and schizophrenia (SCZ). We also found brain IDPs specifically from temporal lobe as a putatively causal consequence of hypertension. The genome-wide colocalization analysis identified three genomic regions in which MDD, ASD and SCZ colocalized with the brain IDPs, and two novel SNPs to be associated with ASD, SCZ, and multiple brain IDPs. Furthermore, we identified a list of candidate genes involved in the shared genetics underlying pairs of brain IDPs and MDD, ASD, SCZ, ALS and hypertension. Our results provide novel insights into the genetic relationships between brain disorders and cardiovascular diseases and brain IDP, which may server as clues for using brain IDPs to predict risks of diseases.

Summary: The medicinal effect of a drug acts through a series of genes, and the pathological mechanism of a disease is also related to genes with certain biological functions. However, the complex information between drug or disease and a series of genes is neglected by traditional message passing methods. In this study, we proposed a new framework using two different strategies for gene-drug/disease and drug-disease networks, respectively. We employ long short-term memory (LSTM) network to extract the flow of message from series of genes (gene path) to drug/disease. Incorporating the resulting information of gene paths into drug-disease network, we utilize graph convolutional network (GCN) to predict drug-disease associations. Experimental results showed that our method GeneDR (gene-based drug repurposing) makes better use of the information in gene paths, and performs better in predicting drug-disease associations.

The wheat scab caused by Fusarium graminearum (F. graminearum) has seriously affected the yield and quality of wheat in China. In this study, gallic acid (GA), a natural polyphenol, was used to synthesize three azole-modified gallic acid derivatives (AGAs1–3). The antifungal activity of GA and its derivatives against F. graminearum was studied through mycelial growth rate experiments and field efficacy experiments. The results of the mycelial growth rate test showed that the EC50 of AGAs–2 was 0.49 mg/mL, and that of AGAs–3 was 0.42 mg/mL. The biological activity of AGAs–3 on F. graminearum is significantly better than that of GA. The results of field efficacy tests showed that AGAs–2 and AGAs–3 significantly reduced the incidence rate and disease index of wheat scab, and the control effect reached 68.86% and 72.11%, respectively. In addition, preliminary investigation was performed on the possible interaction between AGAs–3 and F. graminearum using density functional theory (DFT). These results indicate that compound AGAs–3, because of its characteristic of imidazolium salts, has potential for use as a green and environmentally friendly plant-derived antifungal agent for plant pathogenic fungi.

Abstract Background Observational studies have revealed that type 2 diabetes (T2D) is associated with an increased risk of peripheral artery disease (PAD). However, whether the two diseases share a genetic basis and whether the relationship is causal remain unclear. It is also unclear as to whether these relationships differ between ethnic groups. Methods By leveraging large-scale genome-wide association study (GWAS) summary statistics of T2D (European-based: N case = 21,926, N control = 342,747; East Asian-based: N case = 36,614, N control = 155,150) and PAD (European-based: N case = 5673, N control = 359,551; East Asian-based: N case = 3593, N control = 208,860), we explored the genetic correlation and putative causal relationship between T2D and PAD in both Europeans and East Asians using linkage disequilibrium score regression and seven Mendelian randomization (MR) models. We also performed multi-trait analysis of GWAS and two gene-based analyses to reveal candidate variants and risk genes involved in the shared genetic basis between T2D and PAD. Results We observed a strong genetic correlation (r g ) between T2D and PAD in both Europeans (r g = 0.51; p-value = 9.34 × 10−15) and East Asians (r g = 0.46; p-value = 1.67 × 10−12). The MR analyses provided consistent evidence for a causal effect of T2D on PAD in both ethnicities (odds ratio [OR] = 1.05 to 1.28 for Europeans and 1.15 to 1.27 for East Asians) but not PAD on T2D. This putative causal effect was not influenced by total cholesterol, body mass index, systolic blood pressure, or smoking initiation according to multivariable MR analysis, and the genetic overlap between T2D and PAD was further explored employing an independent European sample through polygenic risk score regression. Multi-trait analysis of GWAS revealed two novel European-specific single nucleotide polymorphisms (rs927742 and rs1734409) associated with the shared genetic basis of T2D and PAD. Gene-based analyses consistently identified one gene ANKFY1 and gene-gene interactions (e.g., STARD10 [European-specific] to AP3S2 [East Asian-specific]; KCNJ11 [European-specific] to KCNQ1 [East Asian-specific]) associated with the trans-ethnic genetic overlap between T2D and PAD, reflecting a common genetic basis for the co-occurrence of T2D and PAD in both Europeans and East Asians. Conclusions Our study provides the first evidence for a genetically causal effect of T2D on PAD in both Europeans and East Asians. Several candidate variants and risk genes were identified as being associated with this genetic overlap. Our findings emphasize the importance of monitoring PAD status in T2D patients and suggest new genetic biomarkers for screening PAD risk among patients with T2D.

Chinese chestnut (Castanea mollissima) is an important nut tree species, and its embryo is rich in sugar. We combined metabolomic and transcriptomic data to analyze metabolites and genes related to sugar in two Chinese chestnut cultivars at 60, 70, 80, 90 and 100 days after flowering (DAF). The soluble sugar content of high-sugar cultivar at maturity is 1.5 times that of low-sugar cultivar. Thirty sugar metabolites were identified in embryo, with the most dominant being sucrose. Analysis of the gene expression patterns revealed that the high-sugar cultivar promoted the conversion of starch to sucrose by up-regulating genes related to starch degradation and sucrose synthesis at 90-100 DAF. It also strongly increased the enzyme activity of SUS-synthetic, which may promote sucrose synthesis. Gene co-expression network analysis showed that ABA and peroxide were related to starch decomposition during Chinese chestnut ripening. Our study analyzed the composition and molecular synthesis mechanism of sugar in Chinese chestnut embryos, and provided a new insight into the regulation pattern of high sugar accumulation in Chinese chestnut nuts.

The complete biosynthetic pathway from most natural products (NPs) are unknown. Here, the authors report BioNavi-NP, a computational toolkit for bio-retrosynthetic pathway elucidation or reconstruction for both NPs and NP-like compounds.

In order to improve the comprehensive utilization value of Anli resources, the extraction technology of soluble dietary fiber from Anli pomace was studied with a by-product of food processing in this work. On the basis of single factor test, response surface methodology was used to optimize the microwave-assisted composite enzymatic extraction process of soluble dietary fiber from Anli fruit pomace (ALDF). The structure of the prepared dietary fiber was characterized by scanning electron microscopy (SEM), fourier transform infrared (FT-IR) and high performance liquid chromatography (HPLC). The optimization results showed that when the microwave power was 370 W, the liquid-to-material ratio was 14.4:1 mL/g, the added amount of enzyme was 1.6%, and the pH was 7.0, the yield of ALDF was the highest level of 8.07%. The SEM showed that the ALDF had a spindle shape of about 5 μm in length, FT-IR showed that it had the characteristic peak of polysaccharide, and the molecular weight ranged from 5 to 2.076×104 kDa. The results showed that the microwave-assisted enzymatic method had a good extraction effect on the ALDF with a certain industrial application prospect. Therefore, this method could provide a theoretical basis for the high-value development and utilization of Anli fruit pomace.

Abstract The rapid spread of the novel coronavirus disease 2019 (COVID‐19) causes a significant impact on public health. It is critical to diagnose COVID‐19 patients so that they can receive reasonable treatments quickly. The doctors can obtain a precise estimate of the infection's progression and decide more effective treatment options by segmenting the CT images of COVID‐19 patients. However, it is challenging to segment infected regions in CT slices because the infected regions are multi‐scale, and the boundary is not clear due to the low contrast between the infected area and the normal area. In this paper, a coarse‐refine segmentation network is proposed to address these challenges. The coarse‐refine architecture and hybrid loss is used to guide the model to predict the delicate structures with clear boundaries to address the problem of unclear boundaries. The atrous spatial pyramid pooling module in the network is added to improve the performance in detecting infected regions with different scales. Experimental results show that the model in the segmentation of COVID‐19 CT images outperforms other familiar medical segmentation models, enabling the doctor to get a more accurate estimate on the progression of the infection and thus can provide more reasonable treatment options.

Abstract Scaffold hopping is a central task of modern medicinal chemistry for rational drug design, which aims to design molecules of novel scaffolds sharing similar target biological activities toward known hit molecules. Traditionally, scaffolding hopping depends on searching databases of available compounds that can't exploit vast chemical space. In this study, we have re-formulated this task as a supervised molecule-to-molecule translation to generate hopped molecules novel in 2D structure but similar in 3D structure, as inspired by the fact that candidate compounds bind with their targets through 3D conformations. To efficiently train the model, we curated over 50 thousand pairs of molecules with increased bioactivity, similar 3D structure, but different 2D structure from public bioactivity database, which spanned 40 kinases commonly investigated by medicinal chemists. Moreover, we have designed a multimodal molecular transformer architecture by integrating molecular 3D conformer through a spatial graph neural network and protein sequence information through Transformer. The trained DeepHop model was shown able to generate around 70% molecules having improved bioactivity together with high 3D similarity but low 2D scaffold similarity to the template molecules. This ratio was 1.9 times higher than other state-of-the-art deep learning methods and rule- and virtual screening-based methods. Furthermore, we demonstrated that the model could generalize to new target proteins through fine-tuning with a small set of active compounds. Case studies have also shown the advantages and usefulness of DeepHop in practical scaffold hopping scenarios.

In this study, IDF from Pyrus ussuriensis Maxim pomace (PUIDF) was prepared, and its routine constituents as well as physicochemical properties and structure were analyzed. Moreover, the effect of PUIDF on obesity and gut microbiota (GM) through a high-fat diet (HFD) mouse model was investigated. The results demonstrated that PUIDF had a loose and porous surface structure and typical cellulose polysaccharide functional groups with good water holding, oil holding and swelling capacity, as well as glucose and cholesterol adsorption performance in vitro. PUIDF could significantly reduce weight gain and lipid and cholesterol levels in HFD-induced obese mice. More importantly, PUIDF intervention greatly improved the diversity of GM in mice, as well as regulated intestinal microbiota composition and increasing the abundance of short-chain fatty acid producing bacteria. These results suggest that PUIDF may have a positive effect to alleviate hyperlipidemia and improve intestinal bacterial ecosystems.

Shigella dysenteriae is a highly pathogenic microorganism that can cause human bacillary dysentery by contaminating food and drinking water. This study investigated the antibacterial activity of chestnut bur polyphenol extract (CBPE) on S. dysenteriae and the underlying mechanism. The results showed that the minimum inhibitory concentration (MIC) of CBPE for S. dysenteriae was 0.4 mg/mL, and the minimum bactericidal concentration (MBC) was 1.6 mg/mL. CBPE treatment irreversibly disrupted cell morphology, decreased cell activity, and increased cell membrane permeability, cell membrane depolarization, and cell content leakage of S. dysenteriae, indicating that CBPE has obvious destructive effects on the cell membrane and cell wall of S. dysenteriae. Combined transcriptomic and metabolomics analysis revealed that CBPE inhibits S. dysenteriae by interfering with ABC protein transport, sulfur metabolism, purine metabolism, amino acid metabolism, glycerophospholipid metabolism, and some other pathways. These findings provide a theoretical basis for the prevention and treatment of S. dysenteriae infection with extract from chestnut burs.

Abstract Fucoidan is a kind of sulfated polysaccharide that consists mainly of fucose and is found in marine brown alga. Fucoidan is a major bioactive compound in brown seaweeds known to have biological activities including anticancer, anti‐inflammatory, antioxidant, anticoagulant, antithrombotic, antivirus, hypolipidemic, and immune‐modulating properties. Due to the adhesive property of fucoidan, it can increase endothelial cell adhesion and proliferation. An increasing number of investigations in the past few years have demonstrated the extraction process of fucoidan. However, the critical review of fucoidan in health benefits has been less explored. In this review, the recent advances of fucoidan were summarized. Moreover, the extraction and purification processes, the characterization of structure, as well as its physicochemical activities, biological activities, and industrial applications were discussed. This review will provide fundamental information for future investigation on controlling the structure, purity, and consistency of the fucoidan product.

Abstract B cell response plays a critical role against SARS-CoV-2 infection. However, little is known about the diversity and frequency of the paired SARS-CoV-2 antigen-specific BCR repertoire after SARS-CoV-2 infection. Here, we performed single-cell RNA sequencing and VDJ sequencing using the memory and plasma B cells isolated from five convalescent COVID-19 patients, and analyzed the spectrum and transcriptional heterogeneity of antibody immune responses. Via linking BCR to antigen specificity through sequencing (LIBRA-seq), we identified a distinct activated memory B cell subgroup (CD11c high CD95 high ) had a higher proportion of SARS-CoV-2 antigen-labeled cells compared with memory B cells. Our results revealed the diversity of paired BCR repertoire and the non-stochastic pairing of SARS-CoV-2 antigen-specific immunoglobulin heavy and light chains after SARS-CoV-2 infection. The public antibody clonotypes were shared by distinct convalescent individuals. Moreover, several antibodies isolated by LIBRA-seq showed high binding affinity against SARS-CoV-2 receptor-binding domain (RBD) or nucleoprotein (NP) via ELISA assay. Two RBD-reactive antibodies C14646P3S and C2767P3S isolated by LIBRA-seq exhibited high neutralizing activities against both pseudotyped and authentic SARS-CoV-2 viruses in vitro. Our study provides fundamental insights into B cell response following SARS-CoV-2 infection at the single-cell level.

Abstract Schizophrenia (SCZ) is a polygenic disease with a heritability approaching 80%. Over 100 SCZ-related loci have so far been identified by genome-wide association studies (GWAS). However, the risk genes associated with these loci often remain unknown. We present a new risk gene predictor, rGAT-omics, that integrates multi-omics data under a Bayesian framework by combining the Hotelling and Box–Cox transformations. The Bayesian framework was constructed using gene ontology, tissue-specific protein–protein networks, and multi-omics data including differentially expressed genes in SCZ and controls, distance from genes to the index single-nucleotide polymorphisms (SNPs), and de novo mutations. The application of rGAT-omics to the 108 loci identified by a recent GWAS study of SCZ predicted 103 high-risk genes (HRGs) that explain a high proportion of SCZ heritability (Enrichment = 43.44 and $$p = 9.30 \times 10^{ - 9}$$ p = 9.30 × 1 0 − 9 ). HRGs were shown to be significantly ( $$p_{\mathrm{adj}} = 5.35 \times 10^{ - 7}$$ p adj = 5.35 × 1 0 − 7 ) enriched in genes associated with neurological activities, and more likely to be expressed in brain tissues and SCZ-associated cell types than background genes. The predicted HRGs included 16 novel genes not present in any existing databases of SCZ-associated genes or previously predicted to be SCZ risk genes by any other method. More importantly, 13 of these 16 genes were not the nearest to the index SNP markers, and them would have been difficult to identify as risk genes by conventional approaches while ten out of the 16 genes are associated with neurological functions that make them prime candidates for pathological involvement in SCZ. Therefore, rGAT-omics has revealed novel insights into the molecular mechanisms underlying SCZ and could provide potential clues to future therapies.

Abstract Protein solubility is significant in producing new soluble proteins that can reduce the cost of biocatalysts or therapeutic agents. Therefore, a computational model is highly desired to accurately predict protein solubility from the amino acid sequence. Many methods have been developed, but they are mostly based on the one-dimensional embedding of amino acids that is limited to catch spatially structural information. In this study, we have developed a new structure-aware method GraphSol to predict protein solubility by attentive graph convolutional network (GCN), where the protein topology attribute graph was constructed through predicted contact maps only from the sequence. GraphSol was shown to substantially outperform other sequence-based methods. The model was proven to be stable by consistent $${\text{R}}^{2}$$ R 2 of 0.48 in both the cross-validation and independent test of the eSOL dataset. To our best knowledge, this is the first study to utilize the GCN for sequence-based protein solubility predictions. More importantly, this architecture could be easily extended to other protein prediction tasks requiring a raw protein sequence.

Heart failure is characterized by the inability of the heart to pump effectively and generate proper blood circulation to meet the body’s needs; it is a devastating condition that affects more than 100 million people globally. In spite of this, little is known about the mechanisms regulating the transition from cardiac hypertrophy to heart failure. Previously, we identified a cardiomyocyte-enriched gene, CIP, which regulates cardiac homeostasis under pathological stimulation. Here, we show that the cardiac transcriptional factor GATA4 binds the promotor of CIP gene and regulates its expression. We further determined that both CIP mRNA and protein decrease in diseased human hearts. In a mouse model, induced cardiac-specific overexpression of CIP after the establishment of cardiac hypertrophy protects the heart by inhibiting disease progression toward heart failure. Transcriptome analyses revealed that the IGF, mTORC2 and TGFβ signaling pathways mediate the inhibitory function of CIP on pathologic cardiac remodeling. Our study demonstrates GATA4 as an upstream regulator of CIP gene expression in cardiomyocytes, as well as the clinical significance of CIP expression in human heart disease. More importantly, our investigation suggests CIP is a key regulator of the transition from cardiac hypertrophy to heart failure. The ability of CIP to intervene in the onset of heart failure suggests a novel therapeutic avenue of investigation for the prevention of heart disease progression.

Sulfated polysaccharide (SP) from sea cucumber possesses unique structure with a wide range of biological activities and health benefits. Many SPs from sea cucumbers have been extracted, purified and fractionated by different methods. However, commercial processing technologies of SPs from sea cucumbers have not been well established. This review summarizes the methods of pretreatment, extraction, separation, purification, and fractionation of various SPs from different sea cucumber species, and deficiencies and limitations of the methods are pointed out. Besides, suggestion on the combination of the processing methods and technologies of SPs are put forward, and SPs extraction in combination with traditionally well-established sea cucumbers food processing technology should be considered. In addition, standardized processing methods have to be confirmed and commercial applications need to be officially approved for SPs. Finally, a promising future perspective of SPs is optimistically forecast.

Scope: The dysbiosis of intestinal microecology plays an important pathogenic role in the development of inflammatory bowel disease. Methods and Results: A polysaccharide named Fuc-S, with a molecular weight of 156 kDa, was prepared by the ultrasonic degradation of fucoidan. Monosaccharide composition, FTIR, methylation, and NMR spectral analysis indicated that Fuc-S may have a backbone consisting of →3)-α-L-Fucp-(1→, →4)-α-L-Fucp-(1→ and →3, 4)-α-D-Glcp-(1→. Moreover, male C57BL/6 mice were fed three cycles of 1.8% dextran sulfate sodium (DSS) for 5 days and then water for 7 days to induce colitis. The longitudinal microbiome alterations were evaluated using 16S amplicon sequencing. In vivo assays showed that Fuc-S significantly improved clinical manifestations, colon shortening, colon injury, and colonic inflammatory cell infiltration associated with DSS-induced chronic colitis in mice. Further studies revealed that these beneficial effects were associated with the inhibition of Akt, p-38, ERK, and JNK phosphorylation in the colon tissues, regulating the structure and abundance of the gut microbiota, and modulating the host–microbe tryptophan metabolism of the mice with chronic colitis. Conclusion: Our data confirmed the presence of glucose in the backbone of fucoidan and provided useful information that Fuc-S can be applied as an effective functional food and pharmaceutical candidate for IBD treatment.

Due to the large molecular weight and complex structure of Laminaria japonica polysaccharides (LJP), which limit their absorption and utilization by the body, methods to effectively degrade polysaccharides had received more and more attention. In the present research, hot water extraction coupled with three-phase partitioning (TPP) was developed to extract and isolate LJP. Ultrasonic L. japonica polysaccharides (ULJP) were obtained by ultrasonic degradation. In addition, their physicochemical characteristics and in vitro biological activities were investigated. Results indicated that ULJP had lower weight-average molecular weight (153 kDa) and looser surface morphology than the LJP. The primary structures of LJP and ULJP were basically unchanged, both contained α-hexo-pyranoses and were mainly connected by 1,4-glycosidic bonds. Compared with LJP, ULJP had stronger antioxidant activity, α-amylase inhibitory effect and anti-inflammatory effect on RAW264.7 macrophages. The scavenging rate of DPPH free radicals by ULJP is 35.85%. Therefore, ultrasonic degradation could effectively degrade LJP and significantly improve the biological activity of LJP, which provided a theoretical basis for the in-depth utilization and research and development of L. japonica in the fields of medicine and food.

Congenital heart defects (CHDs) represent the most common human birth defects. Our previous study indicates that the malfunction of microRNAs (miRNAs) in cardiac neural crest cells (NCCs), which contribute to the development of the heart and the connected great vessels, is likely linked to the pathogenesis of human CHDs. In this study, we attempt to further search for causative single-nucleotide variants (SNVs) from CHD patients that mediate the mis-regulating of miRNAs on their downstream target genes in the pathogenesis of CHDs. As a result, a total of 2,925 3′UTR SNVs were detected from a CHD cohort. In parallel, we profiled the expression of miRNAs in cardiac NCCs and found 201 expressed miRNAs. A combined analysis with these data further identified three 3′UTR SNVs, including NFATC1 c.*654C>T, FGFRL1 c.*414C>T, and CTNNB1 c.*729_*730insT, which result in the malfunction of miRNA-mediated gene regulation. The dysregulations were further validated experimentally. Therefore, our study indicates that miRNA-mediated gene dysregulation in cardiac NCCs could be an important etiology of congenital heart disease, which could lead to a new direction of diagnostic and therapeutic investigation on congenital heart disease.

BackgroundPredicting hepatocellular carcinoma (HCC) prognosis is important for treatment selection, and it is increasingly interesting to predict prognosis through gene expression data. Currently, the prognosis remains of low accuracy due to the high dimension but small sample size of liver cancer omics data. In previous studies, a transfer learning strategy has been developed by pre-training models on similar cancer types and then fine-tuning the pre-trained models on the target dataset. However, transfer learning has limited performance since other cancer types are similar at different levels, and it is not trivial to balance the relations with different cancer types.MethodsHere, we propose an adaptive transfer-learning-based deep Cox neural network (ATRCN), where cancers are represented by 12 phenotype and 10 genotype features, and suitable cancers were adaptively selected for model pre-training. In this way, the pre-trained model can learn valuable prior knowledge from other cancer types while reducing the biases.ResultsATRCN chose pancreatic and stomach adenocarcinomas as the pre-training cancers, and the experiments indicated that our method improved the C-index of 3.8% by comparing with traditional transfer learning methods. The independent tests on three additional HCC datasets proved the robustness of our model. Based on the divided risk subgroups, we identified 10 HCC prognostic markers, including one new prognostic marker, TTC36. Further wet experiments indicated that TTC36 is associated with the progression of liver cancer cells.ConclusionThese results proved that our proposed deep-learning-based method for HCC prognosis prediction is robust, accurate, and biologically meaningful.

Abstract Performance of structure-based molecular docking largely depends on the accuracy of scoring functions. One important type of scoring functions are knowledge-based potentials derived from known three-dimensional structures of proteins and/or protein–ligand complex structures. This study seeks to improve a knowledge-based protein–ligand potential based on a distance-scale finite ideal-gas reference (DFIRE) state (DLIGAND) by expanding the representation of protein atoms from 13 mol2 atom types to 167 residue-specific atom types, and employing a recently updated dataset containing 12,450 monomer protein chains for training. We found that the updated version DLIGAND2 has a consistent improvement over DLIGAND in predicting binding affinities for either native complex structures or docking-generated poses. More importantly, DLIGAND2 has a 52% increase over DLIGAND in enrichment factors in top 1% predictions based on the DUD-E decoy set, and consistently improves over Autodock Vina and other statistical energy functions in all three benchmark tests. We further found that DLIGAND2 outperforms empirical and machine-learning methods compared for virtual screening on new targets that are not homologous to the DUD-E training set. Given the best performance as a parameter-free statistical potential and among the best in all performance measures, DLIGAND2 should be useful for re-assessing the poses generated by docking software, or acting as one term in other scoring functions. The program is available at https://github.com/sysu-yanglab/DLIGAND2.

Multiple in vitro and in vivo model investigations have suggested a broad spectrum of potential mechanisms by which plant/macrofungi-derived non-starch polysaccharides may play a role in the treatment of inflammatory bowel disease (IBD). This article reviews the in vivo and in vitro evidence of different plant-derived polysaccharides for IBD therapy. Their underlying mechanisms, particularly the molecular mechanisms associated with protective effects in the treatment and prevention of IDB, have been well summarized, including anti-inflammatory, epithelial barrier repair, and the regulation of intestinal flora. Emerging studies have observed the potent role of probiotics in IBD, particularly its ability to modulate gut microbiota, a well-known key factor for IBD. In summary, plant/macrofungi-derived polysaccharides have the potential to be a promising agent for the adjuvant treatment and prevention of IBD and will contribute to the design of well-designed clinical intervention trials that will ultimately improve the therapy of IBD.

Summary: Single-cell RNA sequencing technology promotes the profiling of single-cell transcriptomes at an unprecedented throughput and resolution. However, in scRNA-seq studies, only a low amount of sequenced mRNA in each cell leads to missing detection for a portion of mRNA molecules, i.e. the dropout problem which hinders various downstream analyses. Therefore, it is necessary to develop robust and effective imputation methods for the increasing scRNA-seq data. In this study, we have developed an imputation method (GraphSCI) to impute the dropout events in scRNA-seq data based on the graph convolution networks. Extensive experiments demonstrated that GraphSCI outperforms other state-of-the-art methods for imputation on both simulated and real scRNA-seq data. Meanwhile, GraphSCI is able to accurately infer gene-to-gene relationships and the inferred gene-to-gene relationships could also provide powerful assistance for imputation dynamically during the training process, which is a key promotion of GraphSCI compared with other imputation algorithms.

Castanea mollissima Blume has been used to treat a variety of diseases in Chinese traditional medicine. This study was aimed to examine the anti-fatigue effect of purified Chestnut flower flavonoids (CFF) in mice and its underlying mechanisms using the weight-loaded swimming test. The study revealed that the content of total flavonoids in CFF was 754.20 mgRE/gdm, which identified six main types of flavonoids, myricetin-3-O-glucoside, quercetin-3-O-rutinoside, quercetin-O-glucoside, kaempherol-3-O-rutinoside, isorhamnetin-3-O-rutinoside, kaempferol, respectively. CFF showed excellent free radical scavenging and anti-fatigue effect. Notably, CFF significantly prolonged the swimming time of mice, while down-regulating the levels of lactate dehydrogenase (LDH), blood lactic acid (BLA), blood urea nitrogen (BUN), and creatine kinase (CK), and increasing liver and muscle glycogen. Meanwhile, CFF suppressed exercise-elicited oxidative stress through promoting the SOD activity and reducing the MDA level. Furthermore, CFF up-regulated PGC-1α and PPARα expression in mouse skeletal muscle. In conclusion, these findings indicate that CFF exerts an anti-fatigue effect presumably via energy reserves and oxidative stress inhibition.

In this study, soluble dietary fiber (SDF) and insoluble dietary fiber (IDF) were extracted from Pyrus ussuriensis Maxim pomace via three methods including enzymic extraction (EE), microwave-assisted enzymatic extraction (MEE), and three-phase partitioning (TPP). The effects of different extraction methods on the structure, physicochemical property, and functional activity of the extracted dietary fiber were evaluated. The results showed that different extraction methods had significant effects on the extraction yield, molecular weight distribution, thermal stability, antioxidant activity, and hypoglycemic activity in vitro, but resulted in no difference in the structure and composition of functional groups. It is noteworthy that SDF extracted by TPP has a more complex and porous structure, lower molecular weight, and higher thermal stability, as well as better physicochemical properties and in vitro hypoglycemic activity. IDF extracted by MEE showed the greatest water and oil holding capacity; the highest adsorption capacity for glucose, cholesterol, and nitrite ion; as well as the strongest inhibitory activity on α-amylase. These results suggest that PUP may be a source of cheap natural dietary fiber.

Abstract Biogenic compounds are important materials for drug discovery and chemical biology. In this work, we report a quasi-biogenic molecule generator (QBMG) to compose virtual quasi-biogenic compound libraries by means of gated recurrent unit recurrent neural networks. The library includes stereo-chemical properties, which are crucial features of natural products. QMBG can reproduce the property distribution of the underlying training set, while being able to generate realistic, novel molecules outside of the training set. Furthermore, these compounds are associated with known bioactivities. A focused compound library based on a given chemotype/scaffold can also be generated by this approach combining transfer learning technology. This approach can be used to generate virtual compound libraries for pharmaceutical lead identification and optimization.

The present work is conducted to investigate the optimal extraction technology of polysaccharide from chestnut mushroom (Agrocybe aegerita) using a new method based on accelerated solvent extraction combined with response surface methodology (ASE-RSM). The conventional reflux extraction (CRE) method and ultrasonic-assisted extraction (UAE) method were also carried out. Additionally, the in vitro antioxidant activities, including ABTS and DPPH assay, were evaluated. The RSM method, based on a three level and three variable Box–Behnken design (BBD), was developed to obtain the optimal combination of extraction conditions. In brief, the polysaccharide was optimally extracted with water as extraction solvent, extraction temperature of 71 °C, extraction time of 6.5 min, number of cycles of 3, and extraction pressure of 10 MPa. The 3D response surface plot and the contour plot derived from the mathematical models were applied to determine the optimal conditions. Under the above conditions, the experimental value of polysaccharide yield was 19.77 ± 0.12%, which is in close agreement with the value (19.81%) predicted by the model. These findings demonstrate that ASE-RSM produce much higher polysaccharide and consumed environmentally friendly extraction and solvent systems, have less extraction discrimination and shorter time and provide scientific basis for industrialization of polysaccharide extraction. Moreover, it was proved that the polysaccharide had the potential ability to scavenge ABTS and DPPH.

Abstract Background Protein structure can be described by backbone torsion angles: rotational angles about the N-Cα bond (φ) and the Cα-C bond (ψ) or the angle between Cαi-1-Cαi-Cαi + 1 (θ) and the rotational angle about the Cαi-Cαi + 1 bond (τ). Thus, their accurate prediction is useful for structure prediction and model refinement. Early methods predicted torsion angles in a few discrete bins whereas most recent methods have focused on prediction of angles in real, continuous values. Real value prediction, however, is unable to provide the information on probabilities of predicted angles. Results Here, we propose to predict angles in fine grids of 5° by using deep learning neural networks. We found that this grid-based technique can yield 2–6% higher accuracy in predicting angles in the same 5° bin than existing prediction techniques compared. We further demonstrate the usefulness of predicted probabilities at given angle bins in discrimination of intrinsically disorder regions and in selection of protein models. Conclusions The proposed method may be useful for characterizing protein structure and disorder. The method is available at http://sparks-lab.org/server/SPIDER2/ as a part of SPIDER2 package.

Abstract Background Ascochyta blight, caused by the fungus Ascochyta lentis, is one of the most destructive lentil diseases worldwide, resulting in over $16 million AUD annual loss in Australia alone. The use of resistant cultivars is currently considered the most effective and environmentally sustainable strategy to control this disease. However, little is known about the genes and molecular mechanisms underlying lentil resistance against A. lentis. Results To uncover the genetic basis of lentil resistance to A. lentis, differentially expressed genes were profiled in lentil plants during the early stages of A. lentis infection. The resistant ‘ILL7537’ and susceptible ‘ILL6002’ lentil genotypes were examined at 2, 6, and 24 h post inoculation utilising high throughput RNA-Sequencing. Genotype and time-dependent differential expression analysis identified genes which play key roles in several functions of the defence response: fungal elicitors recognition and early signalling; structural response; biochemical response; transcription regulators; hypersensitive reaction and cell death; and systemic acquired resistance. Overall, the resistant genotype displayed an earlier and faster detection and signalling response to the A. lentis infection and demonstrated higher expression levels of structural defence-related genes. Conclusions This study presents a first-time defence-related transcriptome of lentil to A. lentis, including a comprehensive characterisation of the molecular mechanism through which defence against A. lentis is induced in the resistant lentil genotype.

Abstract Genome-wide association studies (GWAS) have successfully identified single variants associated with diseases. To increase the power of GWAS, gene-based and pathway-based tests are commonly employed to detect more risk factors. However, the gene- and pathway-based association tests may be biased towards genes or pathways containing a large number of single-nucleotide polymorphisms (SNPs) with small P-values caused by high linkage disequilibrium (LD) correlations. To address such bias, numerous pathway-based methods have been developed. Here we propose a novel method, DGAT-path, to divide all SNPs assigned to genes in each pathway into LD blocks, and to sum the chi-square statistics of LD blocks for assessing the significance of the pathway by permutation tests. The method was proven robust with the type I error rate >1.6 times lower than other methods. Meanwhile, the method displays a higher power and is not biased by the pathway size. The applications to the GWAS summary statistics for schizophrenia and breast cancer indicate that the detected top pathways contain more genes close to associated SNPs than other methods. As a result, the method identified 17 and 12 significant pathways containing 20 and 21 novel associated genes, respectively for two diseases. The method is available online by http://sparks-lab.org/server/DGAT-path .

Moraxella catarrhalis is a human-adapted, opportunistic bacterial pathogen of the respiratory mucosa. Although asymptomatic colonization of the nasopharynx is common, M. catarrhalis can ascend into the middle ear, where it is a prevalent causative agent of otitis media in children, or enter the lower respiratory tract, where it is associated with acute exacerbations of chronic obstructive pulmonary disease in adults. Phase variation is the high frequency, random, reversible switching of gene expression that allows bacteria to adapt to different host microenvironments and evade host defences, and is most commonly mediated by simple DNA sequence repeats. Bioinformatic analysis of five closed M. catarrhalis genomes identified 17 unique simple DNA sequence repeat tracts that were variable between strains, indicating the potential to mediate phase variable expression of the associated genes. Assays designed to assess simple sequence repeat variation under conditions mimicking host infection demonstrated that phase variation of uspA1 (ubiquitous surface protein A1) from high to low expression occurs over 72 hours of biofilm passage, while phase variation of uspA2 (ubiquitous surface protein A2) to high expression variants occurs during repeated exposure to human serum, as measured by mRNA levels. We also identify and confirm the variable expression of two novel phase variable genes encoding a Type III DNA methyltransferase (modO), and a conserved hypothetical permease (MC25239_RS00020). These data reveal the repertoire of phase variable genes mediated by simple sequence repeats in M. catarrhalis and demonstrate that modulation of expression under conditions mimicking human infection is attributed to changes in simple sequence repeat length.

Background: Diseases of the nervous system are widely considered to be caused by genetic mutations, and they have been shown to share pathogenic genes. Discovering the shared mechanisms of these diseases is useful for designing common treatments.Method: In this study, by reviewing 518 articles published after 2007 on 20 diseases of the nervous system, we compiled data on 1607 mutations occurring in 365 genes, totals that are 1.9 and 3.2 times larger than those collected in the Clinvar database, respectively. A combination with the Clinvar data gives 2434 pathogenic mutations and 424 genes. Using this information, we measured the genetic similarities between the diseases according to the number of genes causing two diseases simultaneously. Further detection was carried out on the similarity between diseases in terms of cell types. Disease-related cell types were defined as those with disease-related gene enrichment among the marker genes of cells, as ascertained by analyzing single-cell sequencing data. Enrichment profiles of the disease-related genes over 25 cell types were constructed. The disease similarity in terms of cell types was obtained by calculating the distances between the enrichment profiles of these genes. The same strategy was applied to measure the disease similarity in terms of brain regions by analyzing the gene expression data from 10 brain regions.Results: The disease similarity was first measured in terms of genes. The result indicated that the proportions of overlapped genes between diseases were significantly correlated to the DMN scores (phenotypic similarity), with a Pearson correlation coefficient of 0.40 and P-value = 6.0×10-3. The disease similarity analysis for cell types identified that the distances between enrichment profiles of the disease-related genes were negatively correlated to the DMN scores, with Spearman correlation coefficient = -0.26 (P-value = 1.5 × 10-2). However, the brain region enrichment profile distances of the disease-related genes were not significantly correlated with the DMN score. Besides the similarity of diseases, this study identified novel relationships between diseases and cell types.Conclusion: We manually constructed the most comprehensive dataset to date for genes with mutations related to 20 nervous system diseases. By using this dataset, the similarities between diseases in terms of genes and cell types were found to be significantly correlated to their phenotypic similarity. However, the disease similarities in terms of brain regions were not significantly correlated with the phenotypic similarities. Thus, the phenotypic similarity between the diseases is more likely to be caused by dysfunctions of the same genes or the same types of neurons rather than the same brain regions. The data are collected into the database NeurodisM, which is available at http://biomed-ai.org/neurodism.

The dynamics of assembled structures are significantly dependent on joints. Joints parameters, owing to be difficultly measured, are always ignored by pure rigid in numerical modeling and it will result in unreliable or even error descriptions. 1~3 Hz deviations may cause resonance especially in engineering optimization issues; hence it is necessary to identify joints parameters for structural dynamics investigation. In present work, multiobjective optimization algorithms are used to identify joints parameters by approaching actual test results in each series of structure to decrease unreliable or error for numerical modeling. Taking automotive dynamics with seam-welding and spot-welding as examples, the relationship of joints parameters perturbation and structural dynamics is derived to give the selecting reason of parameters’ identification. The actual dynamics of an SUV’s frame and a thin-walled part in BIW (body in white) are utilized to validate the methodology. Results demonstrate that the validated model has enough accuracy to reflect the dynamics of the actual structure. The methodology provides reliable guarantee for dynamic analysis and the design of structure.