Biology subjects

Liu, Y.

Publications and source records attributed to Liu, Y..

At least 1,549 records · Page 86Linked to original sources

bioRxiv2023

Biochemical and chemical biological approaches to mammalian sleep: roles of calcineurin in site-specific dephosphorylation and sleep regulation

Understanding of sleep mechanisms traditionally rely on electrophysiology and genetics but here we have initiated biochemical and chemical biological studies. Sleep was increased in mouse mutants with an alanine replacing threonine at residue 469 (T469A) of the salt inducible kinase 3 (SIK3). We searched for T469 phosphatases by classic purification with HEK293 cells and by a new photo-crosslinking method with mouse brains. Both led to PPP3CA, a catalytic subunit of calcium/calmodulin activated phosphatase (calcineurin). It dephosphorylated T469 and serine (S) 551 but not T221 in SIK3 in vitro. PPP3CA knockdown increased phosphorylation of T469 and S551 but not T221 in mouse brains. Knockdown of its regulatory subunit PPP3R1 significantly reduced daily sleep by more than 5 hours, exceeding other known mouse mutants. Our results have uncovered in vitro and in vivo evidence for site-specific SIK3 dephosphorylation by calcineurin, demonstrated a physiological role for calcineurin in sleep, and suggested sleep control by calcium dependent dephosphorylation.

neuroscience↗

bioRxiv2023

SCFFBXW11 complex targets interleukin-17 receptor A for ubiquitin-proteasome-mediated degradation.

Interleukin-17 (IL-17) is a pro-inflammatory cytokine, participating in innate and adaptive immune responses, that plays an important role in host defense, autoimmune diseases, tissue regeneration, metabolic regulation, and tumor progression. Post-translational modifications (PTMs) are crucial for protein function, stability, cellular localization, cellular transduction, and cell death. However, PTMs of IL-17 receptor A (IL-17RA) have not been investigated. Here, we showed that human IL-17RA was targeted by F-box and WD repeats domain containing 11 (FBXW11) for ubiquitination, followed by proteasome-mediated degradation. We used bioinformatics tools and biochemical techniques to determine that FBXW11 ubiquitinated IL-17RA through a lysine 27-linked polyubiquitin chain, targeting IL-17RA for proteasomal degradation. Domain 665-804 of IL-17RA was critical for interaction with FBXW11 and subsequent ubiquitination. Our study demonstrates that FBXW11 regulates IL-17 signaling pathways at IL-17RA level.

biochemistry↗

bioRxiv2023

Distinguished biological adaptation architecture aggravated population differentiation of Tibeto-Burman-speaking people inferred from 500 whole genome data from 39 populations

Tibeto-Burman (TB) people have tried to adapt to the hypoxic, cold and high-UV high-altitude Tibetan Plateau and complex disease exposure in the lowland wet and hot rainforest since the late Paleolithic period. However, the full landscape of genetic history and biological adaptation of geographically diverse TB people and their interaction mechanism remained unknown. We generated a whole-genome-based meta-database of 500 individuals from 39 TB populations from East Asia and Southeast Asia and presented a comprehensive landscape of genetic diversity, admixture history and differentiated adaptative features of geographically different TB people. We identified geography/language-related genetic differentiation among Tibetan Plateau, Tibetan-Yi-Corridor (TYC) and Southeast Asian TB people, consistent with their differentiated admixture process with incoming or indigenous ancestral source populations. A robust genetic connection between TYC people and ancient YR people supported the Northern origin hypothesis of TB people. We reported population substructure-related differentiated biological adaptative signatures between highland Tibetans and lowland TB people and between geographically different Lolo speakers. Highland adaptative EPAS1 and EGLN variants riched in Tibetans but lacked in TYC people whose adaptation is associated with the physical features and skin pigmentation (EDAR and SLC24A5), hepatic alcohol metabolism (ALDH9A1), regulation of cell-cell adhesion of muscle cells (CTNNA3) and immune/fat metabolism-related adaptative signature. TB-related genomic resources provided new insights into the genetic basis of phenotype differences and better reference for the anthropologically-informed sampling design in biomedical and genomic cohort research.

genetics↗

bioRxiv2023

Systematic single-cell analysis reveals dynamic control of transposable element activity orchestrating the endothelial-to-hematopoietic transition

BackgroundThe endothelial-to-hematopoietic transition (EHT) process during definitive hematopoiesis in vertebrate is highly conserved. Stage-specific expression of transposable elements (TEs) has been detected during zebrafish EHT and may promote hematopoietic stem cell formation by activating inflammatory signaling. However, little is known about how TEs contribute to the EHT process in human and mouse. ResultsWe reconstructed the single-cell EHT trajectories of human and mouse, and resolved the dynamic expression patterns of TEs during EHT. Most TEs presented a transient co-upregulation pattern along the conserved EHT trajectories. Enhanced TE activation was tightly associated with the temporal relaxation of epigenetic silencing systems. TE products can be sensed by multiple pattern recognition receptors, triggering inflammatory signaling to facilitate the emergence of hematopoietic stem cells. Furthermore, we observed that hypoxia-related signals were enriched in cells with higher TE expression. Additionally, we constructed the hematopoietic cis-regulatory network of accessible TEs and identified potential enhancers derived by TEs, which may boost the expression of specific EHT marker genes. ConclusionsOur study provides a systematic vision on how TEs are dynamically controlled to promote the hematopoietic fate decision through transcriptional and cis-regulatory networks, and pre-train the immunity of nascent hematopoietic stem cells.

bioinformatics↗

bioRxiv2023

Not everything, not everywhere, not all at once: a study of brain-wide encoding of movement

Activity related to movement is found throughout sensory and motor regions of the brain. However, it remains unclear how movement-related activity is distributed across the brain and whether systematic differences exist between brain areas. Here, we analyzed movement related activity in brain-wide recordings containing more than 50,000 neurons in mice performing a decision-making task. Using multiple techniques, from markers to deep neural networks, we find that movement-related signals were pervasive across the brain, but systematically differed across areas. Movement-related activity was stronger in areas closer to the motor or sensory periphery. Delineating activity in terms of sensory- and motor-related components revealed finer scale structures of their encodings within brain areas. We further identified activity modulation that correlates with decision-making and uninstructed movement. Our work charts out a largescale map of movement encoding and provides a roadmap for dissecting different forms of movement and decision-making related encoding across multi-regional neural circuits.

neuroscience↗

bioRxiv2023

Improving modelling for epidemic responses: reflections from members of the UK infectious disease modelling community on their experiences during the COVID-19 pandemic

The COVID-19 pandemic both relied and placed significant burdens on the experts involved from research and public health sectors. The sustained high pressure of a pandemic on responders, such as healthcare workers, can lead to lasting psychological impacts including acute stress disorder, post-traumatic stress disorder, burnout, and moral injury, which can impact individual wellbeing and productivity. As members of the infectious disease modelling community, we convened a reflective workshop to understand the professional and personal impacts of response work on our community and to propose recommendations for future epidemic responses. The attendees represented a range of career stages, institutions, and disciplines. This piece was collectively produced by those present at the session based on our collective experiences. Key issues we identified at the workshop were lack of institutional support, insecure contracts, unequal credit and recognition, and mental health impacts. Our recommendations include rewarding impactful work, fostering academia-public health collaboration, decreasing dependence on key individuals by developing teams, increasing transparency in decision-making, and implementing sustainable work practices. Despite limitations in representation, this workshop provided valuable insights into the UK COVID-19 modelling experience and guidance for future public health crises. Recognising and addressing the issues highlighted is crucial, in our view, for ensuring the effectiveness of epidemic response work in the future.

scientific communication and education↗

bioRxiv2023

Select EZH2 inhibitors enhance the viral mimicry effects of DNMT inhibition through a mechanism involving calcium-calcineurin-NFAT signaling

DNA methyltransferase (DNMT) inhibitors are FDA-approved for various hematological malignancies but have limited efficacy in solid tumors. DNA hypomethylation with these drugs is associated with elevated lysine 27 tri-methylation on histone H3 (H3K27me3). We hypothesized that this EZH2-dependent repressive mark limits the full potential of DNMT inhibition. Here, we show in cell line and tumoroid models of colorectal cancer, that low-dose DNMT inhibition sensitizes cells to selective EZH2 inhibitors that have limited single agent toxicity, and that EZH2 inhibition enhances DNMT inhibitor-driven molecular and therapeutic effects. Through integrative epigenomic analyses, we reveal that DNMT inhibition induces H3K27me3 accumulation at genomic regions poised with EZH2. Unexpectedly, combined treatment alters the epigenome landscape to promote transcriptional upregulation of the calcium-calcineurin-NFAT signaling pathway. Blocking this pathway limits the transcriptional activating effects of the drug combination, including expression of transposable elements and innate immune response genes within a viral defense pathway. Consistently, we demonstrate positive correlations between DNMT inhibitor- and innate immune response-associated transcription profiles and calcium signal activation in primary human colon cancer specimens. Collectively, our study demonstrates that compensatory EZH2 activity following DNA hypomethylation presents a barrier to the therapeutic action of DNMT inhibition in colon cancer, reveals a new application of EZH2 inhibitors beyond cancers associated with PRC2 hyperactivity, and links calcium-calcineurin-NFAT signaling to epigenetic therapy-induced viral mimicry. HighlightsO_LISelect EZH2 inhibitors enhance the transcriptional activating and antiproliferative effects of DNA hypomethylating agents in colon cancer cells. C_LIO_LIThe mechanism involves blockade of H3K27me3 accumulation in regions of the genome poised for PRC2 activity. C_LIO_LIDNMT inhibitor + EZH2 inhibitor treatment transcriptionally upregulates calcium-calcineurin- NFAT signaling, and this pathway is necessary for complete induction of viral mimicry and innate immune response pathways. C_LIO_LIThe therapeutic utility of EZH2 inhibitors may be extended beyond cancers with PRC2 hyperactivity in combination regimens with DNMT inhibitors. C_LI

cancer biology↗

bioRxiv2023

Blueprinting expandable nanomaterials with standardized protein building blocks

A wooden house frame consists of many different lumber pieces, but because of the regularity of these building blocks, the structure can be designed using straightforward geometrical principles. The design of multicomponent protein assemblies in comparison has been much more complex, largely due to the irregular shapes of protein structures1. Here we describe extendable linear, curved, and angled protein building blocks, as well as inter-block interactions that conform to specified geometric standards; assemblies designed using these blocks inherit their extendability and regular interaction surfaces, enabling them to be expanded or contracted by varying the number of modules, and reinforced with secondary struts. Using X-ray crystallography and electron microscopy, we validate nanomaterial designs ranging from simple polygonal and circular oligomers that can be concentrically nested, up to large polyhedral nanocages and unbounded straight "train track" assemblies with reconfigurable sizes and geometries that can be readily blueprinted. Because of the complexity of protein structures and sequence-structure relationships, it has not been previously possible to build up large protein assemblies by deliberate placement of protein backbones onto a blank 3D canvas; the simplicity and geometric regularity of our design platform now enables construction of protein nanomaterials according to "back of an envelope" architectural blueprints.

biochemistry↗

bioRxiv2023

Development of MAPT S305 mutation models exhibiting elevated 4R tau expression, resulting in altered neuronal and astrocytic function

Due to the importance of 4R tau in the pathogenicity of primary tauopathies, it has been challenging to model these diseases in iPSC-derived neurons, which express very low levels of 4R tau. To address this problem we have developed a panel of isogenic iPSC lines carrying the MAPT splice-site mutations S305S, S305I or S305N, derived from four different donors. All three mutations significantly increased the proportion of 4R tau expression in iPSC-neurons and astrocytes, with up to 80% 4R transcripts in S305N neurons from as early as 4 weeks of differentiation. Transcriptomic and functional analyses of S305 mutant neurons revealed shared disruption in glutamate signaling and synaptic maturity, but divergent effects on mitochondrial bioenergetics. In iPSC-astrocytes, S305 mutations induced lysosomal disruption and inflammation and exacerbated internalization of exogenous tau that may be a precursor to the glial pathologies observed in many tauopathies. In conclusion, we present a novel panel of human iPSC lines that express unprecedented levels of 4R tau in neurons and astrocytes. These lines recapitulate previously characterized tauopathy-relevant phenotypes, but also highlight functional differences between the wild type 4R and mutant 4R proteins. We also highlight the functional importance of MAPT expression in astrocytes. These lines will be highly beneficial to tauopathy researchers enabling a more complete understanding of the pathogenic mechanisms underlying 4R tauopathies across different cell types.

neuroscience↗

bioRxiv2023

Haplotype-resolve genome assembly and resequencing provide insights into the origin and domestication of modern rose

Modern rose (Rosa hybrida) is a recently formed interspecific hybrid and has become one of the most important and widely cultivated ornamentals. Here, we report the haplotype-resolved chromosome-scale genome assembly of the tetraploid R. hybrida Samantha and a genome variation map of 233 Rosa accessions involving various wild species, and old and modern cultivars. The two subgenomes of Samantha show no significant global bias in gene loss but substantial expression bias, and homoeologous exchanges are frequently observed between subgenomes. Population genomic and genomic composition analyses reveal the contributions of wild Rosa species to modern roses, and highlight that R. odorata and its derived cultivars contribute more to modern roses than R. chinensis Old Blush. Furthermore, selective sweeps during modern rose breeding associated with major agronomic traits, including continuous flowering, floral organ identity, flower senescence, and disease resistance, are identified. This study provides insights into the genetic basis of modern rose origin and breeding history, and offers unprecedented genomic resources for rose improvement.

genomics↗

bioRxiv2023

Cryo-EM structure of the human G-protein coupled receptor 1 (GPR1) - Gi protein complex bound to the chemerin C-terminal nonapeptide

Chemerin is an adipokine with chemotactic activity to a subset of leukocytes. Chemerin mainly acts through its C-terminal nonapeptide (YFPGQFAFS, C9) that bind to three G protein-coupled receptors including chemokine-like receptor 1 (CMKLR1), G-protein coupled receptor 1 (GPR1) and C-C chemokine receptor-like 2 (CCRL2). We examined C9 signaling through GPR1 and found this receptor capable of Gi signaling but very weak {beta}-arrestin signaling. Here we report high-resolution cryo-EM structures of GPR1-Gi complexes bound to full-length chemerin and to C9, respectively. Unlike C9 that inserts directly into a transmembrane binding pocket, full-length chemerin uses its N-terminal globular core for extensive interaction with the N terminus of GPR1. Within the binding pocket, the C terminal loop containing the nonapeptide takes the same "S-shape" pose as synthetic C9 nonapeptide. These findings explain why the nonapeptide is a full agonist of GPR1, and demonstrate that chemerin uses a "two-site" model for interaction with GPR1. An analysis of the GPR1-Gi protein interface found high similarities to the CMKLR1-Gi complex, as confirmed by site-directed mutagenesis with functional verifications. Our structural analysis demonstrates shared features with chemokines in that chemerin acts as a "reverse chemokine" with switched functions of its N and C termini in the interaction with GPR1.

pharmacology and toxicology↗

bioRxiv2023

Filament structure and subcellular organization of the bacterial intermediate filament-like protein crescentin

The coiled coil protein crescentin is required for the crescent shape of the freshwater bacterium Caulobacter crescentus (vibrioides). Crescentin forms a filamentous structure on the inner, concave side of the curved cells. It shares features with eukaryotic intermediate filament (IF) proteins, such as its ability to form filaments in vitro, the proteins length, sequence comparisons and the presence of a coiled coil discontinuity called the "stutter". Here, we have used electron cryomicroscopy (cryo-EM) to determine the structure of the full-length protein and its filament, exploiting a crescentin-specific nanobody. The filament is formed by two strands, related by two-fold symmetry that each consist of two dimers, resulting in an octameric assembly. Crescentin subunits form longitudinal contacts head-to-head and tail-to-tail, making the entire filament non-polar. Using in vivo site-directed cysteine crosslinking we demonstrated that contacts observed in the in vitro filament structure exist in cells. Electron cryotomography (cryo-ET) of cells expressing crescentin showed filaments on the concave side of the curved cells, close to the inner membrane, where they form a band. Comparison of our crescentin filament structure with current models of IF proteins and their filaments revealed similar coiled coil dimer formation as well as an absence of overall polarity. IF proteins form head-to-tail longitudinal contacts in contrast to crescentin and hence several inter-dimer contacts in IFs have no equivalents in crescentin filaments. Our work supports the idea that intermediate filament-like proteins achieve their shared polymerization and mechanical properties through a variety of filament architectures. SIGNIFICANCE STATEMENTCrescentin is a coiled coil protein that is required for the crescent cell shape of bacteria such as Caulobacter crescentus. Crescentin shares biochemical and cytoskeletal properties with intermediate filament (IF) proteins, which form the third major class of cytoskeletal proteins in eukaryotes. To better understand the relationship between crescentin and IF proteins, and the filaments they form, we have determined the three-dimensional structure of crescentin filaments by cryo-EM. This revealed the full-length structure of the parallel coiled coil dimer of crescentin and how dimers come together laterally and longitudinally, to form a non-polar, octameric filament. Differences in filament architecture highlight the versatility of intermediate filament-like proteins across the tree of life.

microbiology↗

bioRxiv2023

YAP/TAZ enhance P-body formation to promote tumorigenesis

The role of Processing bodies (P-bodies) in tumorigenesis and tumor progression is not well understood. Here, we showed that the oncogenes YAP/TAZ promote P-body formation in a series of cancer cell lines. Mechanistically, both transcriptional activation of the P- body-related genes SAMD4A, AJUBA, and WTIP and transcriptional suppression of the tumor suppressor gene PNRC1 are involved in enhancing the effects of YAP/TAZ on P- body formation in CRC cells. By reexpression of PNRC1 or knockdown of P-body core genes (DDX6, DCP1A, and LSM14A), we determined that disruption of P-bodies attenuates cell proliferation, cell migration and tumor growth induced by overexpression of YAP5SA in CRC. Analysis of a pancancer CRISPR screen database (DepMap) revealed codependencies between YAP/TEAD and the P-body core genes and correlations between the mRNA levels of SAMD4A, AJUBA, WTIP, PNRC1 and YAP target genes. Our study suggests that the P-body is a new downstream effector of YAP/TAZ, which implies that reexpression of PNRC1 or disruption of P-bodies is a potential therapeutic strategy for tumors with active YAP.

cell biology↗

bioRxiv2023

Morphological brain networks of white matter: mapping, evaluation, characterization and application

Neuroimaging-based connectomics studies have long focused on the wiring patterns between gray matter regions. In recent years, increasing evidence emerges that neural activity in specific sets of white matter (WM) tracts dynamically fluctuates in a coordinated manner. However, the structural basis underlying the coordination is poorly understood largely due to the lack of approaches for estimating structural relations between WM regions. Here, we developed an approach to construct morphological WM networks based on structural magnetic resonance imaging. We found that the morphological WM networks exhibited nontrivial organizational principles, presented good to excellent short- and long-term reliability, accounted for phenotypic interindividual differences (Motor and Cognition), and were under genetic control. Interestingly, highly heritable edges contributed largely to interindividual differences in phenotype. Through integration with other multimodal and multiscale data, we further showed that the morphological WM networks were able to predict regional profiles of hamodynamic coherence, metabolic synchronization, gene co-expression and chemoarchitectonic covariance. Moreover, the prediction followed functional connectomic hierarchy of WM for hamodynamic coherence, was driven by genes enriched in the forebrain neuron development and differentiation for gene co-expression, and was attributed to serotonergic system-related receptors and transporters for chemoarchitectonic covariance. Finally, applying our approach to multiple sclerosis and neuromyelitis optica spectrum disorders, we found that both diseases were associated with morphological WM dysconnectivity, which was correlated with clinical variables and able to diagnose and differentiate the diseases. Altogether, our findings indicate that morphological WM networks provide a reliable and meaningful means to explore WM architecture in health and disease.

neuroscience↗

bioRxiv2023

Yeast Ixr1 mediates the DNA replication stress response through it HMGB DNA binding domains and interaction with checkpoint Mrc1

BackgroundHigh mobility group box (HMGB) family protein Ixr1 has been shown to be involved in DNA damage repair, however, its role and mechanism remain largely unclear. MethodsGenes of S. cerevisiae were deleted or tagged with myc, GFP, or mcherry using the lithium acetate method. Sensitivity of strains to hydroxyurea (HU), methyl methanesulfonate (MMS), camptothe-cin (CPT), 4-nitroquinoline N-oxide (4-NQ), or Zeocin was tested. Distribution of GFP or mcherry fusion proteins was visualized with laser scanning confocal microscopy. RNA-seq was used to determine differential gene expression between mutant and control strains. ResultsIxr1 deletion (ixr1{Delta}) mutant strain was sensitive to HU. Additionally, phosphorylation of effector of DNA damage checkpoint kinase Rad53 was lower in ixr1{Delta} than WT. Deletion of DNA damage checkpoint mediators ixr1{Delta} Rad9{Delta} was more sensitive to HU than ixr1{Delta} or Rad9{Delta}, and ixr1{Delta} mrc1{Delta} had similar sensitivity to HU as mrc1{Delta} but stronger than ixr1{Delta}. Deletion of ribonucleotide reductase inhibitors sml1{Delta} or crt10{Delta} didnt reduce the sensitivity of ixr1{Delta} induced by HU. Repli-cation fork nuclease exo1{Delta} ixr1{Delta} or helicase sgs1{Delta} ixr1{Delta} double deletions were more sensitive to HU than single deletion. In addition, laser scanning confocal microscopy imaging indicated that in response to HU, Ixr1 may be in the same pathway as Mrc1, possibly downstream. Gene Ontol-ogy enrichment analysis of differentially expressed genes (DEGs) between ixr1{Delta} and wildtype, untreated and treated with HU, confirmed that Ixr1 plays an important role in regulating the transcription of genes related to DNA replication or DNA damage repair. We also found that, re-gardless of HU exposure, Ixr1 localized to the nucleus and may bind DNA through its two HMG-boxes. ConclusionIxr1 participates in the DNA replication stress response through a DNA damage checkpoint pathway mediated by Mrc1, and regulates expression of genes related to DNA damage repair.

genetics↗

bioRxiv2023

Developmental Alterations in Brain Network Asymmetry in 3- to 9-Month Infants with Congenital Sensorineural Hearing Loss

Auditory exposure plays crucial roles in shaping healthy brain development and generating lateralization of functional network organization. However, little is known about whether and how an initial lack of auditory exposure in early infancy may disrupt development of functional network lateralization. We addressed this issue by recruiting 55 infants with congenital sensorineural hearing loss (SNHL) and 60 typically developing (TD) controls. Resting-state fNIRS imaging data were acquired to construct hemispheric cerebral networks, and graph theory was applied to quantify the topological characteristics of hemispheric networks. The infants with SNHL exhibited efficient small-world characteristic within each hemispheric network, however, the lateralization of functional network efficiency was substantially disrupted. Compared with TD infants with significantly increased network efficiency lateralized toward left hemisphere with age, the SNHL infants did not exhibit the emergence and development of such cerebral lateralization. Furthermore, the increased leftward asymmetry in nodal efficiency with age was found in TD but not in SNHL infants. Interestingly, the degree of hearing loss had no correlation with lateralization strength in the SNHL group. These results suggest that SNHL infants exhibited disrupted development of cortical lateralization in functional network organization, and highlight the importance of auditory stimulation-promoted multisensory functional integration in early infancy.

neuroscience↗

bioRxiv2023

Flexible Analysis of Spatial Transcriptomics Data (FAST): A Deconvolution Approach

MotivationSpatial transcriptomics is a state-of-art technique that allows researchers to study gene expression patterns in tissues over the spatial domain. As a result of technical limitations, the majority of spatial transcriptomics techniques provide bulk data for each sequencing spot. Consequently, in order to obtain high-resolution spatial transcriptomics data, performing deconvolution becomes essential. Deconvolution enables the determination of the proportions of different cell types along with the corresponding gene expression levels for each cell type within each spot. Most existing deconvolution methods rely on reference data (e.g., single-cell data), which may not be available in real applications. Current reference-free methods encounter limitations due to their dependence on distribution assumptions, reliance on marker genes, or the absence of leveraging histology and spatial information. Consequently, there is a critical demand for the development of highly adaptable, robust, and user-friendly reference-free deconvolution methods capable of unifying or leveraging case-specific information in the analysis of spatial transcriptomics data. ResultsWe propose a novel reference-free method based on regularized non-negative matrix factorization (NMF), named Flexible Analysis of Spatial Transcriptomics (FAST), that can effectively incorporate gene expression data, spatial coordinates, and histology information into a unified deconvolution framework. Compared to existing methods, FAST imposes fewer distribution assumptions, utilizes the spatial structure information of tissues, and encourages interpretable factorization results. These features enable greater flexibility and accuracy, making FAST an effective tool for deciphering the complex cell-type composition of tissues and advancing our understanding of various biological processes and diseases. Extensive simulation studies have shown that FAST outperforms other existing reference-free methods. In real data applications, FAST is able to uncover the underlying tissue structures and identify the corresponding marker genes.

bioinformatics↗

bioRxiv2023

Protein-protein interaction prediction via structure-based deep learning

Protein-protein interactions (PPIs) play an essential role in life activities. Many machine learning algorithms based on protein sequence information have been developed to predict PPIs. However, these models have difficulty dealing with various sequence lengths and suffer from low generalization and prediction accuracy. In this study, we proposed a novel end-to-end deep learning framework, RSPPI, combining Residual Neural Network (ResNet) and Spatial Pyramid Pooling (SPP), to predict PPIs based on the protein sequence physicochemistry properties and spatial structural information. In the RSPPI model, ResNet was employed to extract the structural and physicochemical information from the protein 3D structure and primary sequence; the SPP layer was used to transform feature maps to a single vector and avoid the fixed-length requirement. The RSPPI model possessed excellent cross-species performance and outperformed several state-of-the-art methods based either on protein sequence or gene ontology in most evaluation metrics. The RSPPI model provides a novel strategic direction to develop an AI PPI prediction algorithm.

bioinformatics↗