CMAUP database update 2024: extended functional and association information of useful plants for biomedical research.

Knowledge of the collective activities of individual plants together with the derived clinical effects and targeted disease associations is useful for plant-based biomedical research. To provide the information in complement to the established databases, we introduced a major update of CMAUP database, previously featured in NAR. This update includes (i) human transcriptomic changes overlapping with 1152 targets of 5765 individual plants, covering 74 diseases from 20 027 patient samples; (ii) clinical information for 185 individual plants in 691 clinical trials; (iii) drug development information for 4694 drug-producing plants with metabolites developed into approved or clinical trial drugs; (iv) plant and human disease associations (428 737 associations by target, 220 935 reversion of transcriptomic changes, 764 and 154121 associations by clinical trials of individual plants and plant ingredients); (v) the location of individual plants in the phylogenetic tree for navigating taxonomic neighbors, (vi) DNA barcodes of 3949 plants, (vii) predicted human oral bioavailability of plant ingredients by the established SwissADME and HobPre algorithm, (viii) 21-107% increase of CMAUP data over the previous version to cover 60 222 chemical ingredients, 7865 plants, 758 targets, 1399 diseases, 238 KEGG human pathways, 3013 gene ontologies and 1203 disease ontologies. CMAUP update version is freely accessible at https://bidd.group/CMAUP/index.html.

Exploring outer space biophysical phenomena via SpaceLID.

Extensive investigations in outer space have revealed not only how life adapts to the space environment, but also that interesting biophysical phenomena occur. These phenomena affect human health and other life forms (animals, plants, bacteria, and fungi), and to ensure the safety of future human space exploration need to be further investigated. This calls for joint research efforts between biologists and physicists, as these phenomena present cross-disciplinary barriers. Various national organizations provide useful forums for bridging this gap. Additional discussion avenues and database resources are helpful for facilitating the interdisciplinary investigations of these phenomena. In this paper, we present the newly established Space Life Investigation Database (SpaceLID, https://bidd.group/spacelid/ ) which provides information about biophysical phenomena occurring in space. Examples obtained using the database are given while discussing the underlying causes of these phenomena and their implications for the physiology and health of life in space.

Database of space life investigations and information on spaceflight plant biology.

Extensive spaceflight life investigations (SLIs) have revealed observable space effects on plants, particularly their growth, nutrition yield, and secondary metabolite production. Knowledge of these effects not only facilitates space agricultural and biopharmaceutical technology development but also provides unique perspectives to ground-based investigations. SLIs are specialized experimental protocols and notable biological phenomena. These require specialized databases, leading to the development of the NASA Science Data Archive, Erasmus Experiment Archive, and NASA GeneLab. The increasing interests of SLIs across diverse fields demand resources with comprehensive content, convenient search facilities, and friendly information presentation. A new database SpaceLID (Space Life Investigation Database http://bidd.group/spacelid/ ) was developed with detailed menu search tools and categorized contents about the phenomena, protocols, and outcomes of 459 SLIs (including 106 plant investigations) of 92 species, where 236 SLIs and 57 plant investigations are uncovered by the existing databases. The usefulness of SpaceLID as an SLI information source is illustrated by the literature-reported analysis of metabolite, nutrition, and symbiosis variations of spaceflight plants. In conclusion, this study extensively investigated the impact of the space environment on plant biology, utilizing SpaceLID as an information source and examining various plant species, including Arabidopsis thaliana, Brassica rapa L., and Glycyrrhiza uralensis Fisch. The findings provide valuable insights into the effects of space conditions on plant physiology and metabolism.

Infrared Spectral Analysis for Prediction of Functional Groups Based on Feature-Aggregated Deep Learning.

Infrared (IR) spectroscopy is a powerful and versatile tool for analyzing functional groups in organic compounds. A complex and time-consuming interpretation of massive unknown spectra usually requires knowledge of chemistry and spectroscopy. This paper presents a new deep learning method for transforming IR spectral features into intuitive imagelike feature maps and prediction of major functional groups. We obtained 8272 gas-phase IR spectra from the NIST Chemistry WebBook. Feature maps are constructed using the intrinsic correlation of spectral data, and prediction models are developed based on convolutional neural networks. Twenty-one major functional groups for each molecule are successfully identified using binary and multilabel models without expert guidance and feature selection. The multilabel classification model can produce all prediction results simultaneously for rapid characterization. Further analysis of the detailed substructures indicates that our model is capable of obtaining abundant structural information from IR spectra for a comprehensive investigation. The interpretation of our model reveals that the peaks of most interest are similar to those often considered by spectroscopists. In addition to demonstrating great potential for spectral identification, our method may contribute to the development of automated analyses in many fields.

Deep learning of 2D-Restructured gene expression representations for improved low-sample therapeutic response prediction.

Clinical outcome prediction is important for stratified therapeutics. Machine learning (ML) and deep learning (DL) methods facilitate therapeutic response prediction from transcriptomic profiles of cells and clinical samples. Clinical transcriptomic DL is challenged by the low-sample sizes (34-286 subjects), high-dimensionality (up to 21,653 genes) and unordered nature of clinical transcriptomic data. The established methods rely on ML algorithms at accuracy levels of 0.6-0.8 AUC/ACC values. Low-sample DL algorithms are needed for enhanced prediction capability. Here, an unsupervised manifold-guided algorithm was employed for restructuring transcriptomic data into ordered image-like 2D-representations, followed by efficient DL of these 2D-representations with deep ConvNets. Our DL models significantly outperformed the state-of-the-art (SOTA) ML models on 82% of 17 low-sample benchmark datasets (53% with >0.05 AUC/ACC improvement). They are more robust than the SOTA models in cross-cohort prediction tasks, and in identifying robust biomarkers and response-dependent variational patterns consistent with experimental indications.

NPASS database update 2023: quantitative natural product activity and species source database for biomedical research.

Quantitative activity and species source data of natural products (NPs) are important for drug discovery, medicinal plant research, and microbial investigations. Activity values of NPs against specific targets are useful for discovering targeted therapeutic agents and investigating the mechanism of medicinal plants. Composition/concentration values of NPs in individual species facilitate the assessments and investigations of the therapeutic quality of herbs and phenotypes of microbes. Here, we describe an update of the NPASS natural product activity and species source database previously featured in NAR. This update includes: (i) new data of ∼95 000 records of the composition/concentration values of ∼1 490 NPs/NP clusters in ∼390 species, (ii) extended data of activity values of ∼43 200 NPs against ∼7 700 targets (∼40% and ∼32% increase, respectively), (iii) extended data of ∼31 600 species sources of ∼94 400 NPs (∼26% and ∼32% increase, respectively), (iv) new species types of ∼440 co-cultured microbes and ∼420 engineered microbes, (v) new data of ∼66 600 NPs without experimental activity values but with estimated activity profiles from the established chemical similarity tool Chemical Checker, (vi) new data of the computed drug-likeness properties and the absorption, distribution, metabolism, excretion and toxicity (ADMET) properties for all NPs. NPASS update version is freely accessible at http://bidd.group/NPASS.

Enhanced metagenomic deep learning for disease prediction and consistent signature recognition by restructured microbiome 2D representations.

Metagenomic analysis has been explored for disease diagnosis and biomarker discovery. Low sample sizes, high dimensionality, and sparsity of metagenomic data challenge metagenomic investigations. Here, an unsupervised microbial embedding, grouping, and mapping algorithm (MEGMA) was developed to transform metagenomic data into individualized multichannel microbiome 2D representation by manifold learning and clustering of microbial profiles (e.g., composition, abundance, hierarchy, and taxonomy). These 2D representations enable enhanced disease prediction by established ConvNet-based AggMapNet models, outperforming the commonly used machine learning and deep learning models in metagenomic benchmark datasets. These 2D representations combined with AggMapNet explainable module robustly identified more reliable and replicable disease-prediction microbes (biomarkers). Employing the MEGMA-AggMapNet pipeline for biomarker identification from 5 disease datasets, 84% of the identified biomarkers have been described in over 74 distinct works as important for these diseases. Moreover, the method also discovered highly consistent sets of biomarkers in cross-cohort colorectal cancer (CRC) patients and microbial shifts in different CRC stages.

Toward ordered -omics data science: Researchers on the magic of turning metagenomic chaos into image-like patterns.

Wan Xiang Shen, a postdoctoral researcher at National University of Singapore, and Yu Zong Chen, the PI of the Bioinformatics and Drug Design (BIDD) group, have developed an AI pipeline for enhanced deep learning of metagenomic data. Their Patterns paper highlights the advantages of unsupervised data restructuring in microbiome-based disease prediction and biomarker discovery. They talk about their view of data science and the backstory of the article published in Patterns.

Database of space life investigations and bioinformatics of microbiology in extreme environments.

Biological experiments performed in space crafts like space stations, space shuttles, and recoverable satellites has enabled extensive spaceflight life investigations (SLIs). In particular, SLIs have revealed distinguished space effects on microbial growth, survival, metabolite production, biofilm formation, virulence development and drug resistant mutations. These provide unique perspectives to ground-based microbiology and new opportunities for industrial pharmaceutical and metabolite productions. SLIs are with specialized experimental setups, analysis methods and research outcomes, which can be accessed by established databases National Aeronautics and Space Administration (NASA) Life Science Data Archive, Erasmus Experiment Archive, and NASA GeneLab. The increasing research across diverse fields may be better facilitated by databases of convenient search facilities and categorized presentation of comprehensive contents. We therefore developed the Space Life Investigation Database (SpaceLID) http://bidd.group/spacelid/, which collected SLIs from published academic papers. Currently, this database provides detailed menu search facilities and categorized contents about the studied phenomena, materials, experimental procedures, analysis methods, and research outcomes of 448 SLIs of 90 species (microbial, plant, animal, human), 81 foods and 106 pharmaceuticals, including 232 SLIs not covered by the established databases. The potential applications of SpaceLID are illustrated by the examples of published experimental design and bioinformatic analysis of spaceflight microbial phenomena.

In silico approaches uncovering the systematic function of N-phosphorylated proteins in human cells.

Phosphorylation plays a key role in the regulation of protein function. In addition to the extensively studied O-phosphorylation of serine, threonine, and tyrosine, emerging evidence suggests that the non-canonical phosphorylation of histidine, lysine, and arginine termed N-phosphorylation, exists widely in eukaryotes. At present, the study of N-phosphorylation is still in its infancy, and its regulatory role and specific biological functions in mammalian cells are still unknown. Here, we report the in silico analysis of the systematic biological significance of N-phosphorylated proteins in human cells. The protein structural and functional domain enrichment analysis revealed that N-phosphorylated proteins are rich in RNA recognition motif, nucleotide-binding and alpha-beta plait domains. The most commonly enriched biological pathway is the metabolism of RNA. Besides, arginine phosphorylated (pArg) proteins are highly related to DNA repair, while histidine phosphorylated (pHis) proteins may play a role in the regulation of the cell cycle, and lysine phosphorylated (pLys) proteins are linked to cellular stress response, intracellular signal transduction, and intracellular transport, which are of great significance for maintaining cell homeostasis. Protein-protein interaction (PPI) network analysis revealed important hub proteins (i.e., SRSF1, HNRNPA1, HNRNPC, SRSF7, HNRNPH1, SRSF2, SRSF11, HNRNPD, SRRM2 and YBX1) which are closely related to neoplasms, nervous system diseases, and virus infection and have potential as therapeutic targets. Those proteins with clinical significance are worthy of attention, and the rational considerations of N-phosphorylation in occurrence and progression of diseases might be beneficial for further translational applications.

Fluorescence Analysis of Circulating Exosomes for Breast Cancer Diagnosis Using a Sensor Array and Deep Learning.

Emerging liquid biopsy methods for investigating biomarkers in bodily fluids such as blood, saliva, or urine can be used to perform noninvasive cancer detection. However, the complexity and heterogeneity of exosomes require improved methods to achieve the desired sensitivity and accuracy. Herein, we report our study on developing a breast cancer liquid biopsy system, including a fluorescence sensor array and deep learning (DL) tool AggMapNet. In particular, we used a 12-unit sensor array composed of conjugated polyelectrolytes, fluorophore-labeled peptides, and monosaccharides or glycans to collect fluorescence signals from cells and exosomes. Linear discriminant analysis (LDA) processed the fluorescence spectral data of cells and cell-derived exosomes, demonstrating successful discrimination between normal and different cancerous cells and 100% accurate classification of different BC cells. For heterogeneous plasma-derived exosome analysis, CNN-based DL tool AggMapNet was applied to transform the unordered fluorescence spectra into feature maps (Fmaps), which gave a straightforward visual demonstration of the difference between healthy donors and BC patients with 100% prediction accuracy. Our work indicates that our fluorescent sensor array and DL model can be used as a promising noninvasive method for BC diagnosis.

Chinese Medicine Meets Conventional Medicine in Targeting COVID-19 Pathophysiology, Complications and Comorbidities.

OBJECTIVE: To investigate how the National Health Commission of China (NHCC)-recommended Chinese medicines (CMs) modulate the major maladjustments of coronavirus disease 2019 (COVID-19), particularly the clinically observed complications and comorbidities. METHODS: By focusing on the potent targets in common with the conventional medicines, we investigated the mechanisms of 11 NHCC-recommended CMs in the modulation of the major COVID-19 pathophysiology (hyperinflammations, viral replication), complications (pain, headache) and comorbidities (hypertension, obesity, diabetes). The constituent herbs of these CMs and their chemical ingredients were from the Traditional Chinese Medicine Information Database. The experimentally-determined targets and the activity values of the chemical ingredients of these CMs were from the Natural Product Activity and Species Source Database. The approved and clinical trial drugs against these targets were searched from the Therapeutic Target Database and DrugBank Database. Pathways of the targets was obtained from Kyoto Encyclopedia of Genes and Genomes and additional literature search. RESULTS: Overall, 9 CMs modulated 6 targets discovered by the COVID-19 target discovery studies, 8 and 11 CMs modulated 8 and 6 targets of the approved or clinical trial drugs for the treatment of the major COVID-19 complications and comorbidities, respectively. CONCLUSION: The coordinated actions of each NHCC-recommended CM against a few targets of the major COVID-19 pathophysiology, complications and comorbidities, partly have common mechanisms with the conventional medicines.

AggMapNet: enhanced and explainable low-sample omics deep learning with feature-aggregated multi-channel networks.

Omics-based biomedical learning frequently relies on data of high-dimensions (up to thousands) and low-sample sizes (dozens to hundreds), which challenges efficient deep learning (DL) algorithms, particularly for low-sample omics investigations. Here, an unsupervised novel feature aggregation tool AggMap was developed to Aggregate and Map omics features into multi-channel 2D spatial-correlated image-like feature maps (Fmaps) based on their intrinsic correlations. AggMap exhibits strong feature reconstruction capabilities on a randomized benchmark dataset, outperforming existing methods. With AggMap multi-channel Fmaps as inputs, newly-developed multi-channel DL AggMapNet models outperformed the state-of-the-art machine learning models on 18 low-sample omics benchmark tasks. AggMapNet exhibited better robustness in learning noisy data and disease classification. The AggMapNet explainable module Simply-explainer identified key metabolites and proteins for COVID-19 detections and severity predictions. The unsupervised AggMap algorithm of good feature restructuring abilities combined with supervised explainable AggMapNet architecture establish a pipeline for enhanced learning and interpretability of low-sample omics data.

Protein music of enhanced musicality by music style guided exploration of diverse amino acid properties.

Inspired by the traceable analogies between protein sequences and music notes, protein music has been composed from amino acid sequences for popularizing science and sourcing melodies. Despite the continuous development of protein-to-music algorithms, the musicality of protein music lags far behind human music. Musicality may be enhanced by fine-tuned protein-to-music mapping to the features of a specific music style. We analyzed the features of a music style (Fantasy-Impromptu style), and used the quantized musical features to guide broad exploration of diverse amino acid properties (104 properties, sequence patterns and variations) for developing a novel protein-to-music algorithm of enhanced musicality. This algorithm was applied to 18 proteins of various biological functions. The derived music pieces consistently exhibited enhanced musicality with respect to existing protein music. Music style guided exploration of diverse amino acid properties enable protein music composition of enhanced musicality, which may be further developed and applied to a wider variety of music styles.

Combining kinase inhibitors for optimally co-targeting cancer and drug escape by exploitation of drug target promiscuities.

Cancers resist targeted therapeutics by drug-escape signaling. Multitarget drugs co-targeting cancer and drug-escape mediators (DEMs) are clinically advantageous. DEM coverage may be expanded by drug combinations. This work evaluated to what extent the kinase DEMs (KDEMs) can be optimally co-targeted by drug combinations based on target promiscuities of individual drugs. We focused on 41 approved and 28 clinical trial small molecule kinase inhibitor drugs with available experimental kinome and clinical pharmacokinetic data. From the kinome inhibitory profiles of these drugs, drug combinations were assembled for optimally co-targeting an established cancer target (EGFR, HER2, ABL1, or MEK1) and 9-16 target-associated KDEMs at comparable potency levels as that against the cancer target. Each set of two-, three-, and four-drug combinations co-target 36-71%, 44-89%, 50-88%, and 27-55% KDEMs of EGFR, HER2, ABL1, and MEK1, respectively, compared with the 36, 33, 38, and 18% KDEMs maximally co-targeted by an existing drug or drug combination approved or clinically tested for the respective cancer. Some co-targeted KDEMs are not covered by any existing drug or drug combination. Our work suggested that novel drug combinations may be constructed for optimally co-targeting cancer and drug escape by the exploitation of drug target promiscuities.

MASI: microbiota-active substance interactions database.

Xenobiotic and host active substances interact with gut microbiota to influence human health and therapeutics. Dietary, pharmaceutical, herbal and environmental substances are modified by microbiota with altered bioavailabilities, bioactivities and toxic effects. Xenobiotics also affect microbiota with health implications. Knowledge of these microbiota and active substance interactions is important for understanding microbiota-regulated functions and therapeutics. Established microbiota databases provide useful information about the microbiota-disease associations, diet and drug interventions, and microbiota modulation of drugs. However, there is insufficient information on the active substances modified by microbiota and the abundance of gut bacteria in humans. Only ∼7% drugs are covered by the established databases. To complement these databases, we developed MASI, Microbiota-Active Substance Interactions database, for providing the information about the microbiota alteration of various substances, substance alteration of microbiota, and the abundance of gut bacteria in humans. These include 1,051 pharmaceutical, 103 dietary, 119 herbal, 46 probiotic, 142 environmental substances interacting with 806 microbiota species linked to 56 diseases and 784 microbiota-disease associations. MASI covers 11 215 bacteria-pharmaceutical, 914 bacteria-herbal, 309 bacteria-dietary, 753 bacteria-environmental substance interactions and the abundance profiles of 259 bacteria species in 3465 patients and 5334 healthy individuals. MASI is freely accessible at http://www.aiddlab.com/MASI.