Biology subjects

Hu, J.

Publications and source records attributed to Hu, J..

At least 469 records · Page 26Linked to original sources

bioRxiv2020

Radiosynthesis, in vitro and preliminary in vivo evaluation of the novel glutamine derived PET tracers fluorobiphenylglutamine

IntroductionGlucose has been deemed the driving force of tumor growth for decades. However, research has shown that several tumors metabolically shift towards glutaminolysis. The development of radiolabeled glutamine derivatives could be a useful molecular imaging tool for visualizing these tumors. We elaborated on the glutamine-derived PET tracers by developing two novel probes, namely [18F]fluorophenylglutamine and [18F]fluorobiphenylglutamine Materials and methodsBoth tracers were labelled with fluorine-18 using our recently reported ruthenium-based direct aromatic fluorination method. Their affinity was evaluated with a [3H]glutamine inhibition experiment in a human PC-3 and a rat F98 cell line. The imaging potential of [18F]fluorophenylglutamine and [18F]fluorobiphenylglutamine was tested using a mouse PC-3 and a rat F98 tumor model. ResultsThe radiosynthesis of both tracers was successful with overall non-decay corrected yields of 18.46 {+/-} 4.18 % (n=10) ([18F]fluorophenylglutamine) and 8.05 {+/-} 3.25 % (n=5) ([18F]fluorobiphenylglutamine). In vitro inhibition experiments showed a moderate and low affinity of fluorophenylglutamine and fluorobiphenylglutamine, respectively, towards the human ASCT-2 transporter. Both compounds had a low affinity towards the rat ASCT-2 transporter. These results were endorsed by the in vivo experiments with low uptake of both tracers in the F98 rat xenograft, low uptake of [18F]FBPG in the mice PC-3 xenograft and a moderate uptake of [18F]FPG in the PC-3 tumors. ConclusionWe investigated the imaging potential of two novel PET radiotracers [18F]FPG and [18F]FBPG. [18F]FPG is the first example of a glutamine radiotracer derivatized with a phenyl group which enables the exploration of further derivatization of the phenyl group to increase the affinity and imaging qualities. We hypothesize that increasing the affinity of [18F]FPG by optimizing the substituents of the arene ring can result in a high-quality glutamine-based PET radiotracer. Advances in Knowledge and Implications for patient careWe hereby report novel glutamine-based PET-tracers. These tracers are tagged on the arene group with fluorine-18, hereby preventing in vivo defluorination, which can occur with alkyl labelled tracers (e.g. (2S,4R)4-[18F]fluoroglutamine). [18F]FPG shows clear tumor uptake in vivo, has no in vivo defluorination and has a straightforward production. We believe this tracer is a good starting point for the development of a high-quality tracer which is useful for the clinical visualization of the glutamine transport. O_FIG_DISPLAY_L [Figure 1] M_FIG_DISPLAY C_FIG_DISPLAY

cancer biology↗

bioRxiv2020

Comparison of the two up-to-date sequencing technologies for genome assembly: HiFi reads of Pacbio Sequel II system and ultralong reads of Oxford Nanopore

The availability of reference genomes has revolutionized the study of biology. Multiple competing technologies have been developed to improve the quality and robustness of genome assemblies during the last decade. The two widely-used long read sequencing providers - Pacbio (PB) and Oxford Nanopore Technologies (ONT) - have recently updated their platforms: PB enable high throughput HiFi reads with base-level resolution with >99% and ONT generated reads as long as 2 Mb. We applied the two up-to-date platforms to one single rice individual, and then compared the two assemblies to investigate the advantages and limitations of each. The results showed that ONT ultralong reads delivered higher contiguity producing a total of 18 contigs of which 10 were assembled into a single chromosome compared to that of 394 contigs and three chromosome-level contigs for the PB assembly. The ONT ultralong reads also prevented assembly errors caused by long repetitive regions for which we observed a total 44 genes of false redundancies and 10 genes of false losses in the PB assembly leading to over/under-estimations of the gene families in those long repetitive regions. We also noted that the PB HiFi reads generated assemblies with considerably less errors at the level of single nucleotide and small InDels than that of the ONT assembly which generated an average 1.06 errors per Kb assembly and finally engendered 1,475 incorrect gene annotations via altered or truncated protein predictions.

genomics↗

bioRxiv2020

Iterative transfer learning with neural network for clustering and cell type classification in single-cell RNA-seq analysis

An important step in single-cell RNA-seq (scRNA-seq) analysis is to cluster cells into different populations or types. Here we describe ItClust, an Iterative Transfer learning algorithm with neural network for scRNA-seq Clustering. ItClust learns cell type knowledge from well-annotated source data, but also leverages information in the target data to make it less dependent on the source data quality. Through extensive evaluations using datasets from different species and tissues generated with diverse scRNA-seq protocols, we show that ItClust significantly improves clustering and cell type classification accuracy compared to popular unsupervised clustering and supervised cell type classification algorithms.

genomics↗

bioRxiv2020

Microbial trend analysis for common dynamic trend, group comparison and classification in longitudinal microbiome study

MotivationThe human microbiome is inherently dynamic and its dynamic nature plays a critical role in maintaining health and driving disease. With an increasing number of longitudinal microbiome studies, scientists are eager to learn the comprehensive characterization of microbial dynamics and their implications to the health and disease-related phenotypes. However, due to the challenging structure of longitudinal microbiome data, few analytic methods are available to characterize the microbial dynamics over time. ResultsWe propose a microbial trend analysis (MTA) framework for the high-dimensional and phylogenetically-based longitudinal microbiome data. In particular, MTA can perform three tasks: 1) capture the common microbial dynamic trends for a group of subjects on the community level and identify the dominant taxa; 2) examine whether or not the microbial overall dynamic trends are significantly different in groups; 3) classify an individual subject based on its longitudinal microbial profiling. Our extensive simulations demonstrate that the proposed MTA framework is robust and powerful in hypothesis testing, taxon identification, and subject classification. Our real data analyses further illustrate the utility of MTA through a longitudinal study in mice. ConclusionsThe proposed MTA framework is an attractive and effective tool in investigating dynamic microbial pattern from longitudinal microbiome studies.

bioinformatics↗

bioRxiv2020

Transmission dynamics of 2019 novel coronavirus (2019-nCoV)

RationaleSeveral studies have estimated basic production number of novel coronavirus pneumonia (NCP). However, the time-varying transmission dynamics of NCP during the outbreak remain unclear. ObjectivesWe aimed to estimate the basic and time-varying transmission dynamics of NCP across China, and compared them with SARS. MethodsData on NCP cases by February 7, 2020 were collected from epidemiological investigations or official websites. Data on severe acute respiratory syndrome (SARS) cases in Guangdong Province, Beijing and Hong Kong during 2002-2003 were also obtained. We estimated the doubling time, basic reproduction number (R0) and time-varying reproduction number (Rt) of NCP and SARS. Measurements and main resultsAs of February 7, 2020, 34,598 NCP cases were identified in China, and daily confirmed cases decreased after February 4. The doubling time of NCP nationwide was 2.4 days which was shorter than that of SARS in Guangdong (14.3 days), Hong Kong (5.7 days) and Beijing (12.4 days). The R0 of NCP cases nationwide and in Wuhan were 4.5 and 4.4 respectively, which were higher than R0 of SARS in Guangdong (R0=2.3), Hongkong (R0=2.3), and Beijing (R0=2.6). The Rt for NCP continuously decreased especially after January 16 nationwide and in Wuhan. The R0 for secondary NCP cases in Guangdong was 0.6, and the Rt values were less than 1 during the epidemic. ConclusionsNCP may have a higher transmissibility than SARS, and the efforts of containing the outbreak are effective. However, the efforts are needed to persist in for reducing time-varying reproduction number below one. At a Glance CommentaryO_ST_ABSScientific Knowledge on the SubjectC_ST_ABSSince December 29, 2019, pneumonia infection with 2019-nCoV, now named as Novel Coronavirus Pneumonia (NCP), occurred in Wuhan, Hubei Province, China. The disease has rapidly spread from Wuhan to other areas. As a novel virus, the time-varying transmission dynamics of NCP remain unclear, and it is also important to compare it with SARS. What This Study Adds to the FieldWe compared the transmission dynamics of NCP with SARS, and found that NCP has a higher transmissibility than SARS. Time-varying production number indicates that rigorous control measures taken by governments are effective across China, and persistent efforts are needed to be taken for reducing instantaneous reproduction number below one.

systems biology↗

bioRxiv2020

Bcl6 Preserves the Suppressive Function of Regulatory T Cells during Tumorigenesis

During tumorigenesis, tumor infiltrating regulatory T (Treg) cells restrict the function of effector T cells in tumor microenvironment and thereby promoting tumor growth. The anti-tumor activity of effector T cells can be therapeutically unleashed, and is now being exploited for the treatment of various types of human cancers. However, the immune suppressive function of Treg cells remains a major hurdle to broader effectiveness of tumor immunotherapy. In this article, we reported that the deletion of Bcl6 specifically in Treg cells led to stunted tumor growth, which was caused by impaired Treg cell responses. Notably, Bcl6 is essential in maintaining the lineage stability of Treg cells in tumor microenvironment. Meanwhile, we found that the absence of follicular regulatory T (Tfr) cells, which is a result of Bcl6 deletion in Foxp3+ cells, was dispensable for tumor control. Importantly, the increased Bcl6 expression in Treg cells is associated with poor prognosis of human colorectal cancer and lymph node metastasis of skin melanoma. Furthermore, Bcl6 deletion in Treg cells exhibits synergistic effects with immune checkpoint blockade therapy. Collectively, these results indicate that Bcl6 actively participates in regulating Treg cell immune responses during tumorigenesis and can be exploited as a therapeutic target of anti-tumor immunity.

immunology↗

bioRxiv2020

Compartmentalization of enhanced biomolecular interactions for high-throughput drug screening in test tubes

Modification-dependent and -independent biomolecular interactions (BIs), including protein-protein, protein-DNA/RNA and protein-lipid, play crucial roles in all cellular processes. Dysregulation of BIs or malfunction of the associated enzymes results in various diseases, thus they are attractive targets for therapies. High-throughput screening (HTS) can greatly facilitate the discovery of drugs for these targets. Here we describe a HTS drug discovery method, called compartmentalization of enhanced biomolecular interactions in test tubes (CEBIT). CEBIT uses selective recruitment of biomolecules into phase separated compartments harboring their cognate binding partners as readouts. CEBIT were tailored to detect various BIs and associated modifying enzymes. Using CEBIT-based HTS assays, we successfully identified known inhibitors of the p53/MDM2 interaction and of SUV39H1 from a compound library. CEBIT is simple and versatile, and is likely to become a powerful tool for drug discovery and basic biomedical research.

biophysics↗

bioRxiv2020

scTPA: A web tool for single-cell transcriptome analysis of pathway activation signatures

The most fundamental challenge in current single-cell RNA-seq data analysis is functional interpretation and annotation of cell clusters. The biological pathways in distinct cell types have different activation patterns, which facilitates understanding cell functions in single-cell transcriptomics. However, no effective web tool has been implemented for single-cell transcriptomic data analysis based on prior biological pathway knowledge. Here, we introduce scTPA (http://sctpa.bio-data.cn/sctpa), which is a web-based platform providing pathway-based analysis of single-cell RNA-seq data in human and mouse. scTPA incorporates four widely-used gene set enrichment methods to estimate the pathway activation scores of single cells based on a collection of available biological pathways with different functional and taxonomic classifications. The clustering analysis and cell-type-specific activation pathway identification were provided for the functional interpretation of cell types from pathway-oriented perspective. An intuitive interface allows users to conveniently visualize and download single-cell pathway signatures. Together, scTPA is a comprehensive tool to identify pathway activation signatures for dissecting single cell heterogeneity.

genomics↗

bioRxiv2020

CRISPR-mediated activation of biosynthetic gene clusters for bioactive molecule discovery in filamentous fungi

Accessing the full biosynthetic potential encoded in the genomes of fungi is limited by the low expression of most biosynthetic gene clusters (BGCs) under common laboratory culture conditions. CRISPR-mediated transcriptional activation (CRISPRa) of fungal BGC could accelerate genomics-driven bioactive secondary metabolite discovery. In this work, we established the first CRISPRa system for filamentous fungi. First, we constructed a CRISPR/dLbCas12a-VPR-based system and demonstrated the activation of a fluorescent reporter in Aspergillus nidulans. Then, we targeted the native nonribosomal peptide synthetase-like (NRPS-like) gene micA in both chromosomal and episomal contexts, achieving increased production of the compound microperfuranone. Finally, multi-gene CRISPRa led to the discovery of the mic cluster product as dehydromicroperfuranone. Additionally, we demonstrated the utility of the variant dLbCas12aD156R-VPR for CRISPRa at room temperature culture conditions. Different aspects that influence the efficiency of CRISPRa in fungi were investigated, providing a framework for the further development of fungal artificial transcription factors based on CRISPR/Cas.Competing Interest StatementThe authors have declared no competing interest.View Full Text

synthetic biology↗

bioRxiv2020

PD-L1 upregulation by IFN-α/γ-mediated Stat1 suppresses anti-HBV T cell response

Programmed death ligand 1 (PD-L1) has been recently shown to be a major obstacle to antiviral immunity by binding to its receptor programmed death 1 (PD-1) on specific IFN-{gamma} producing T cells in chronic hepatitis B. Currently, IFN- is widely used to treat hepatitis B virus(HBV) infection, but its antiviral effect vary greatly and the mechanism is not totally clear. We found that IFN-/{gamma} induced a marked increase of PD-L1 expression in hepatocytes. Signal and activators of transcription (Stat1) was then identified as a major transcription factor involved in IFN-/{gamma}-mediated PD-L1 elevation both in vitro and in mice. Blockage of the PD-L1/PD-1 interaction by a specific mAb greatly enhanced HBV-specific T cell activity by the gp96 adjuvanted therapeutic vaccine, and promoted HBV clearance in HBV transgenic mice. Our results demonstrate the IFN-/{gamma}-Stat1-PD-L1 axis plays an important role in mediating T cell hyporesponsiveness and inactivating liver-infiltrating T cells in the hepatic microenvironment. These data raise further potential interest in enhancing the anti-HBV efficacy of IFN- and therapeutic vaccines.

molecular biology↗

bioRxiv2019

A Stochastic Model for Actin Waves in Eukaryotic Cells

A stochastic model of spontaneous actin wave formation in eukaryotic cells that includes positive feedback between the actin network and filament nucleating factors on the membrane is developed and analyzed. Simulation results show that the model can produce a variety of actin network behavior depending on the conditions. Actin spots of diameter about 0.5 {micro}m can be formed and persist for tens of seconds at low actin concentrations, and these spots may either shrink and die or grow and develop into fully-developed propagating waves. The model correctly captures the vertical profile of actin waves along line scans through wave fronts, as well as the separation between the region enclosed by circular actin waves and the external area. Our results show how the complicated actin behavior depends on the amounts and state of various membrane molecules. Author SummaryLocomotion of eukaryotic cells is a complex process that involves the spatio-temporal control and integration of a number of sub-processes, including the transduction of chemical or mechanical signals from the environment, local and global modification of the cytoskeleton, and translation of the intra- and extracellular signals into a mechanical response. In view of the complexity of the processes, understanding how force generation and mechanical interactions with the surroundings are controlled in space and time to produce cell-level movement is a major challenge. Recent experimental work has shown that a variety of actin waves propagate within cells, both under normal conditions and during re-building of the cytoskeleton following its disruption. Controlled disruption and re-building of the actin network has led to new insights into the key components involved in actin waves, and here we develop a stochastic model that can qualitatively and quantitatively describe the dynamical behavior of such waves.

systems biology↗

bioRxiv2019

The Tung Tree (Vernicia Fordii) Genome Provides A Resource for Understanding Genome Evolution and Oil Improvement

Tung tree (Vernicia fordii) is an economically important woody oil plant that produces tung oil containing a high proportion of eleostearic acid ([~]80%). Here we report a high-quality, chromosome-scale tung tree genome sequence of 1.12 Gb with 28,422 predicted genes and over 73% repeat sequences. Tung tree genome was assembled by combining Illumina short reads, PacBio single-molecule real-time long reads and Hi-C sequencing data. Insertion time analysis revealed that the repeat-driven tung tree genome expansion might be due to long standing long terminal repeat (LTR) retrotransposon bursts and lack of efficient DNA deletion mechanisms. An electronic fluorescent pictographic (eFP) browser was generated based on genomic and RNA-seq data from 17 various tissues and developmental stages. We identified 88 nucleotide-binding site (NBS)-encoding resistance genes, of which 17 genes may help the tung tree resist the Fusarium wilt shortly after infection. A total of 651 oil-related genes were identified and 88 of them were predicted to be directly involved in tung oil biosynthesis. The fewer phosphoenolpyruvate carboxykinase (PEPC) genes, and synergistic effects between transcription factors and oil biosynthesis-related genes may contribute to high oil content in tung seeds. The tung tree genome should provide valuable resources for molecular breeding and genetic improvement.

bioinformatics↗

bioRxiv2019

Characterization of the Key Region and Phosphorylation Sites of EcaICE1 for its Molecular Interaction with EcaHOS1 Protein from Eucalyptus camaldulensis

ICE1 (inducer of CBF expression 1), a MYC-like bHLH transcriptional activator, plays an important role in plant under cold stress via regulating transcriptional expression of downstream cold-responsive genes. Ubiquitination-proteasome pathway mediated by high expression of osmotically responsive gene1 (HOS1) can effectively induce the degradation of ICE1 and decrease the expression of expression of CBFs and their downstream genes under cold stress response in Arabidopsis, but the knowledge about ubiquitination regulation of ICE1 by HOS1 is still limited in woody plants. In this study, a complete open reading frame of EcaICE1 gene and EcaHOS1 was amplified from the tissue culture seedlings of Eucalyptus camaldulensis. Yeast two-hybrid (Y2H) and BiFC assay results showed that EcaICE1 can interact with EcaHOS1 protein in the nucleus, and further Y2H assay demonstrated that the 126-185 amino acid region at the N-terminus of EcaICE1 protein was indispensable for its interaction with EcaHOS1 protein. Moreover, we found that the amino acids at positions 143, 145, 158 and 184 within the key interaction region were the potential phosphorylation sites of EcaICE1 based on bioinformatics analysis, and that only the substitution of Serine (Ser) 158 by Alanine (Ala) blocked the protein-protein interactions between EcaICE1 and EcaHOS1 by Y2H and {beta}-galactosidase assays using site-direct mutagenesis. In summary, we first reported that EcaICE1 could interact with EcaHOS1 protein in Eucalyptus, and identified Ser 158 of EcaICE1 as the key phosphorylation site for its interaction with EcaHOS1 protein.

plant biology↗

bioRxiv2019

Characterization of transcription activation domain of EcaICE1 and its interaction with EcaSIZ1 in Eucalyptus camaldulensis

SUMOylation in plants is associated with biotic and abiotic stress responses, flowering and other aspects of development, and ICE1 protein SUMOylation by SUMO E3 ligase SIZ1 plays important roles in plant cold tolerance. Here, we reported the subcellular localization of EcaICE1 and its interaction with EcaSIZ1 in Eucalyptus camaldulensis. The genes EcaICE1 and EcaSIZ1 were isolated by homologous cloning. The subcellular localization analysis showed that EcaICE1 was located in nucleus. Bimolecular fluorescence complementation (BiFC) analysis revealed that EcaICE1 could interact with EcaSIZ1 in the nucleus of Nicotiana benthamiana leaves. Moreover, yeast two-hybrid assay confirmed that the amino acid region from position 84 to 126 in EcaICE1 was critical for the strong transactivation activity of EcaICE1 and that the C terminal region from position 361 to 557 in EcaICE1 was the key region for its interaction with EcaSIZ1 by using different truncated lengths of non-transactivation activity of EcaICE1 as the bait protein. Collectively, our results showed that EcaICE1 may have a SUMOylation pathway similar to Arabidopsis thaliana.

plant biology↗

bioRxiv2019

The microRNA476a/RFL module regulates adventitious root formation through a mitochondria-dependent pathway in Populus

Adventitious root (AR) formation at the base of stem cuttings determines the efficiency of clonal propagation for woody plants. Many endogenous and environmental factors influence AR formation. However, our knowledge about the regulation of AR development by mitochondrial metabolism in plants is very limited. Here we identified Populus-specific miR476a as a novel regulator of wound-induced adventitious rooting via orchestrating mitochondrial homeostasis in poplar. MiR476a exhibited inducible expression during AR formation and directly targets several Restorer of Fertility like (RFL) genes encoding mitochondrion-localized pentatricopeptide repeat proteins. Genetic modification of miR476-RFL expression revealed the miR476/RFL-mediated dynamic regulation of mitochondrial homeostasis on AR formation in transgenic poplar. Furthermore, mitochondrial perturbation via exogenous chemical inhibitor validated that the miR476a/RFL-directed AR formation depended on mitochondrial regulation though modulating the auxin pathway. Our results established a miRNA-directed mitochondrion-auxin signaling cascade required for AR development, providing novel insights into the understanding of mitochondrial regulation on plant developmental plasticity.

plant biology↗

bioRxiv2019

The rice PPR756 coordinates with MORFs for multiple RNA editing in mitochondria

In land plants, the pentatricopeptide repeat (PPR) proteins form a large family involved in post-transcriptional processing of RNA in mitochondria and chloroplasts, which is critical for plant development and evolutionary adaption. Although studies showed a number of PPR proteins generally influence the editing of organellar genes, few of them were characterized in detail in rice. Here, we report a PLS-E subclass PPR protein in rice, PPR756, loss of function of which led to the abolishment of RNA editing events among three mitochondrial genes including atp6, ccmC, and nad7. Their defective C-to-U transformation then resulted in improper amino acid retention which could cause abortive pollen development. Furthermore, PPR756 could bind to the three target genes directly and interact with three OsMORFs (multiple organellar RNA editing factors): OsMORF1, OsMORF8-1, and OsMORF8-2. The knock-out plants of PPR756 exhibited retarded growth and greener leaves during the early vegetative stages, along with sterile pollen and lower seed setting at the reproductive stage. These results established a role for PPR756 in rice development, participating in RNA editing of three various transcripts and cooperating with OsMORFs via an editosome manner in rice.

developmental biology↗

bioRxiv2019

The substrate-dependent stereoselectivity of the multicopper oxidase (MCO)-catalyzed oxidative coupling in the biosynthesis of the bisnaphthopyrone viriditoxin

VdtB, the multiple-copper oxidase (MCO) from the bisnaphthopyrone (M)-viriditoxin biosynthetic pathway in Paecilomyces variotii, was shown to catalyze regioselective 6,6'-coupling of semi-viriditoxin (1). The stereoselectivity of the oxidative coupling reaction for the production of the atropisomer (M)-viriditoxin, however, was controlled by VdtD, a non-catalytic dirigent protein from the pathway. In this work, VdtB either alone or together with VdtD were investigated for its stereoselective control upon coupling of other monomeric naphthopyrone derivatives from the pathway with different minor structural variations in terms of presence/absence of O-methylation at C7-position and C3-C4{Delta} 2 double bond on the pyrone ring, and the different side-chain modifications. We showed that VdtB could favour either M- or P-form coupling in a substrate-dependent manner. For some substrates, VdtB could catalyze oxidative coupling in an enantiomerically selective manner. The efficiency of the VdtD in exerting stereoselective control of the oxidative coupling reaction also varies between substrates. The results point to a model whereby VdtB and VdtD form a VdtB-ligand-VdtD complex in which the stereochemical outcome of the coupling reaction depends on how the substrate interacts with both proteins, based on the substrate structure. Our findings contributed to a more comprehensive understanding of dirigent protein-mediated MCO-catalyzed stereoselective oxidative coupling reactions in fungi.

biochemistry↗

bioRxiv2019

Deep convolutional neural network and attention mechanism based pan-specific model for interpretable MHC-I peptide binding prediction

Accurate prediction of peptide binding affinity to the major histocompatibility complex (MHC) proteins has the potential to design better therapeutic vaccines. Previous work has shown that pan-specific prediction algorithms can achieve better prediction performance than other approaches. However, most of the top algorithms are neural networks based black box models. Here, we propose DeepAttentionPan, an improved pan-specific model, based on convolutional neural networks and attention mechanisms for more flexible, stable and interpretable MHC-I binding prediction. With the attention mechanism, our ensemble model consisting of 20 trained networks achieves high and more stabilized prediction performance. Extensive tests on IEDBs weekly benchmark dataset show that our method achieves state-of-the-art prediction performance on 21 test allele datasets. Analysis of the peptide positional attention weights learned by our model demonstrates its capability to capture critical binding positions of the peptides, which leads to mechanistic understanding of MHC-peptide binding with high alignment with experimentally verified results. Furthermore, we show that with transfer learning, our pan model can be fine-tuned for alleles with few samples to achieve additional performance improvement. DeepAttentionPan is freely available as an open source software at https://github.com/jjin49/DeepAttentionPan.\n\nAuthor summaryHuman leukocyte antigen (HLA) proteins are classes of proteins that are responsible for immune system regulation in humans. The peptides are short chains of amino acids. HLA class I group present peptides from inside the cell to the cell surface for scrutiny by T cell receptors. For instance, if the cell is infected by a virus, the HLA system will bind to the peptides derived from viral proteins and bring them to the surface of the cell so that the cell can be destroyed by the immune system. Since the HLA genes exhibit extensive polymorphism, there are many HLA alleles binding to different peptides. And this diversity represents challenges in predicting binders for different HLA alleles, which are important in vaccine designs and characterization of immune responses. Before computational algorithms are used to predict the binding relationships of HLA-peptide pairs, scientists need to conduct costly biological experiments to do preliminary screening among a number of peptides and need to use mutant experiments to identify key peptide positions that contribute to the binding. While previous computational methods have been proposed to predict the binding affinity, identifying the binding anchors is not well addressed. Here we developed a deep neural network models with the attention mechanism to learn the binding relationships automatically in an end-to-end way. Our models are able to identify the important binding positions of the peptide sequence by learning the positional importance distribution, which used to be studied a lot only through costly experimental methods. Our model thus not only improves the performance of binding affinity prediction but also allows us to gain biological insight of binding motifs of different alleles via interpreting the learned deep neural network models.

bioinformatics↗