Biology subjects

Zhu, S.

Publications and source records attributed to Zhu, S..

At least 163 records · Page 9Linked to original sources

bioRxiv2021

Efficient ancestry and mutation simulation with msprime 1.0

Stochastic simulation is a key tool in population genetics, since the models involved are often analytically intractable and simulation is usually the only way of obtaining ground-truth data to evaluate inferences. Because of this necessity, a large number of specialised simulation programs have been developed, each filling a particular niche, but with largely overlapping functionality and a substantial duplication of effort. Here, we introduce msprime version 1.0, which efficiently implements ancestry and mutation simulations based on the succinct tree sequence data structure and tskit library. We summarise msprimes many features, and show that its performance is excellent, often many times faster and more memory efficient than specialised alternatives. These high-performance features have been thoroughly tested and validated, and built using a collaborative, open source development model, which reduces duplication of effort and promotes software quality via community engagement.

genetics↗

bioRxiv2021

Cryo-ET of a human GBP coatomer governing cell-autonomous innate immunity to infection

All living organisms deploy cell-autonomous defenses to combat infection. In plants and animals, these activities generate large supramolecular complexes that recruit immune proteins for protection. Here, we solve the native structure of a massive antimicrobial complex generated by polymerization of 30,000 human guanylate-binding proteins (GBPs) over the entire surface of virulent bacteria. Construction of this giant nanomachine takes [~]1-3 minutes, remains stable for hours, and acts as a cytokine and cell death signaling platform atop the coated bacterium. Cryo-ET of this "coatomer" revealed thousands of human GBP1 molecules undergo [~]260 [A] insertion into the bacterial outer membrane, triggering lipopolysaccharide release that activates co-assembled caspase-4. Together, our results provide a quasi-atomic view of how the GBP coatomer mobilizes cytosolic immunity to combat infection in humans. One-Sentence SummaryThousands of GBPs coat cytosolic bacteria to engineer an antimicrobial signaling platform inside human cells.

immunology↗

bioRxiv2021

Single-shot rAAV5-based Vaccine Provides Long-term Protective Immunity against SARS-CoV-2 and Its Variants

The COVID-19 pandemic and the SARS-CoV-2 with its variants have posed unprecedented challenges worldwide. Existing vaccines have limited effectiveness against the SARS-CoV-2 variants. Therefore, novel vaccines to match current mutated viral lineages with long-term protective immunity are urgently in demand. In the current study, we for the first time designed a recombinant Adeno-Associated Virus 5 (rAAV5)-based vaccine named as rAAV-COVID-19 vaccine (Covacinplus) by using RBD-plus of spike protein with both the single-stranded and the self-complementary AAV5 delivering vectors (ssAAV5 and scAAAV5), which provides excellent protection from SARS-CoV-2 infection. A single dose vaccination induced the strong immune response against SARS-CoV-2. The induced neutralizing antibodies (NAs) titers were maintained at a high peak level of over 1:1024 even after more than one year of injection and accompanied with functional T-cells responses in mice. Importantly, both ssAAV- and scAAV-based RBD-plus vaccines exhibited high levels of serum NAs against current circulating variants including variants Alpha, Beta, Gamma and Delta. SARS-CoV-2 virus challenge test showed that ssAAV5-RBD-plus vaccine protected both young and old age mice from SARS-CoV-2 infection in the upper and the lower respiratory tracts. Moreover, whole genome sequencing demonstrated that AAV vector DNA sequences were not found in the genome of the vaccinated mice after one year vaccination, demonstrating excellent safety of the vaccine. Taken together, this study suggests that rAAV5-based vaccine is powerful against SARS-CoV-2 and its variants with long-term protective immunity and excellent safety, which has great potential for development into prophylactic vaccination in human to end this global pandemic.

immunology↗

bioRxiv2021

An ultrapotent neutralizing bispecific antibody with broad spectrum against SARS-CoV-2 variants

In spite of the successful development of effective countermeasures against Covid-19, variants have and will continue to emerge that could compromise the efficacy of currently approved neutralizing antibodies and vaccines. Consequently, novel and more efficacious agents are urgently needed. We have developed a bispecific antibody, 2022, consisting of two antibodies, 2F8 and VHH18. 2F8 was isolated from our proprietary fully synthetic human IDEAL (Intelligently Designed and Engineered Antibody Library)-VH/VL library and VHH18 is a single domain antibody isolated from IDEAL-nanobody library. 2022 was constructed by attaching VHH18 to the C-terminal of Fc of 2F8. 2022 binds two non-overlapping epitopes simultaneously on the RBD of the SARS-CoV-2 spike protein and blocks the binding of RBD to human angiotensin-converting enzyme 2 (ACE2). 2022 potently neutralizes SARS-CoV-2 and all of the variants tested in both pseudovirus and live virus assays, including variants carrying mutations known to resist neutralizing antibodies approved under EUA and that reduce the protection efficiency of current effective vaccines. The half-maximum inhibitory concentration (IC50) of 2022 is 270 pM, 30 pM, 20 pM, and 1 pM, for wild-type, alpha, beta, and delta pseudovirus, respectively. In the live virus assay, 2022 has an IC50 of 26.4 pM, 13.3 pM, and 88.6 pM, for wild-type, beta, and delta live virus, respectively. In a mouse model of SARS-CoV-2, 2022 showed strong prophylactic and therapeutic effects. A single administration of 2022 intranasal (i.n.) or intraperitoneal (i.p.) 24 hours before virus challenge completely protected all mice from bodyweight loss, as compared with up to 20% loss of bodyweight in placebo treated mice. In addition, the lung viral titers were undetectable (FRNT assay) in all mice treated with 2022 either prophylactically or therapeutically, as compared with around 1x105 pfu/g lung tissue in placebo treated mice. In summary, bispecific antibody 2022 showed potent binding and neutralizing activity across a variety of SARS-CoV-2 variants and could be an attractive weapon to combat the ongoing waves of the COVID-19 pandemic propagated mainly by variants, especially, the much more contagious delta variant.

microbiology↗

bioRxiv2021

MetaBinner: a high-performance and stand-alone ensemble binning method to recover individual genomes from complex microbial communities

Binning is an essential procedure during metagenomic data analysis. However, the available individual binning methods usually do not simultaneously fully use different features or biological information. Furthermore, it is challenging to integrate multiple binning results efficiently and effectively. Therefore, we developed an ensemble binner, MetaBinner, which generates component results with multiple types of features and utilizes single-copy gene (SCG) information for k-means initialization. It then utilizes a two-step ensemble strategy based on SCGs to integrate the component results. Extensive experimental results over three large-scale simulated datasets and one real-world dataset demonstrate that MetaBinner outperforms other state-of-the-art individual binners and ensemble binners. MetaBinner is freely available at https://github.com/ziyewang/MetaBinner.

bioinformatics↗

bioRxiv2021

Critical Assessment of Metagenome Interpretation - the second round of challenges

Evaluating metagenomic software is key for optimizing metagenome interpretation and focus of the community-driven initiative for the Critical Assessment of Metagenome Interpretation (CAMI). In its second challenge, CAMI engaged the community to assess their methods on realistic and complex metagenomic datasets with long and short reads, created from [~]1,700 novel and known microbial genomes, as well as [~]600 novel plasmids and viruses. Altogether 5,002 results by 76 program versions were analyzed, representing a 22x increase in results. Substantial improvements were seen in metagenome assembly, some due to using long-read data. The presence of related strains still was challenging for assembly and genome binning, as was assembly quality for the latter. Taxon profilers demonstrated a marked maturation, with taxon profilers and binners excelling at higher bacterial taxonomic ranks, but underperforming for viruses and archaea. Assessment of clinical pathogen detection techniques revealed a need to improve reproducibility. Analysis of program runtimes and memory usage identified highly efficient programs, including some top performers with other metrics. The CAMI II results identify current challenges, but also guide researchers in selecting methods for specific analyses.

bioinformatics↗

bioRxiv2021

Optogenetic suppression of corticogeniculate feedback in anesthetized ferrets is overridden by visual stimulation

The feedforward projection from the retina shapes the spatial receptive field properties of neurons in the dorsal lateral geniculate nucleus of the thalamus (LGN). Corticogeniculate feedback from the visual cortex appears to exert a more subtle, modulatory influence on LGN responses. Studies involving manipulations of corticogeniculate feedback have yielded inconsistent findings, but the reasons for these inconsistencies are not known. To examine the functional contributions of corticogeniculate feedback, and to resolve past inconsistencies, we examined the effects of selective optogenetic suppression of corticogeniculate neurons in anesthetized ferrets. In particular, we examined the responses of LGN and V1 neurons during optogenetic suppression of corticogeniculate feedback in the presence and absence of visual stimulation and across conditions in which the frequency of LED illumination varied. Optogenetic suppression of corticogeniculate feedback decreased activity among LGN neurons in the absence of visual stimulation, dispelling the notion that anesthesia causes a floor effect. In contrast, suppressing corticogeniculate feedback did not affect the visual responses of LGN neurons, suggesting that feedforward visual stimulus drive overrides weak corticogeniculate influence. Optogenetic effects on LGN and V1 neuronal responses depended on the frequency of LED illumination, with higher frequency illumination inducing slow oscillations in V1, dis-inhibiting V1 neurons locally, and producing more suppression among LGN neurons. These results demonstrate that corticogeniculate influence depends on stimulation parameters including visual stimulus conditions and frequency of inactivation. Furthermore, weak corticogeniculate influence is overridden by strong feedforward visual stimulus drive - this attribute is the most likely source of inconsistencies in past studies. Significance StatementAlthough corticogeniculate synapses onto thalamic neurons far outnumber those coming from the retina, the function of corticogeniculate feedback in vision has remained a stubborn puzzle. Prior studies of corticogeniculate feedback have yielded inconsistent findings, but the source for these inconsistencies is unknown. We utilized selective optogenetic suppression of corticogeniculate feedback to examine its effects on thalamic neuronal responses and to resolve sources of prior inconsistencies. We found that suppression of corticogeniculate feedback reduced thalamic responses, but only in the absence of visual input. This suggests that the major source of inconsistencies across prior studies is the fact that weak corticogeniculate influence is overcome by strong feedforward visual stimulus drive.

neuroscience↗

bioRxiv2021

Artemisinin-based hybrids produce intracellular proteasome inhibitors that overcome resistance in Plasmodium falciparum

Artemisinin resistant Plasmodium falciparum (Pf) is spreading despite combination chemotherapy (ACT). Here we report the design of artezomibs, single-molecule hybrids of an artemisinin and a Pf-selective proteasome inhibitor. Artezomibs exert a novel mode of action inside the malaria parasites. The artemisinin component covalently modifies parasite proteins, which become substrates of the Pf proteasome. The proteasomal degradation products that bear the proteasome inhibitor component of the hybrid then inhibit Pf proteasomes, including those with mutations that reduce binding affinity of the proteasome inhibitor component on its own. We demonstrated that artezomibs circumvent both artemisinin resistance conferred by Kelch13 polymorphism and resistance to the proteasome inhibitor associated with mutations in Pf proteasomes. This mode of action may enable the use of a single molecule with one pharmacokinetic profile to prevent the emergence of resistance.

microbiology↗

bioRxiv2021

Competition between distinct ApoE alleles and mCRP for the endothelial receptor CD31 differentially regulates neurovascular inflammation and Alzheimer's disease pathology

BACKGROUNDC-reactive protein (CRP) in peripheral inflammation is associated with increased Alzheimers disease (AD) risk in Apolipoprotein E4 (ApoE4), but not ApoE3 or E2, humans. It remains unknown whether peripheral monomeric CRP (mCRP) induces AD pathogenesis through some receptor of blood-facing endothelia in the brain in an ApoE genotype dependent fashion. METHODSWe used human samples, ApoE knock-in and deficient mouse models, and primary brain endothelia. Different ApoE mice were intraperitoneally (i.p.) injected with mCRP. The characterizations by immunostaining, proximity ligation assay (PLA) and siRNA were conducted to identify the receptor for mCRP. Brain microvessel and endothelia were isolated for RNA sequencing to explore the molecular pathway. RESULTSWe demonstrate that CD31 (PECAM-1), a blood-facing endothelial receptor in brain, is a competitive target of both mCRP and ApoE protein. ApoE2 competes more strongly with mCRP for CD31 than ApoE4 does, and expressing ApoE4 or knocking out ApoE gene results in higher levels of mCRP-CD31 binding, leading to a decrease of CD31 expression but an increase in CD31 phosphorylation, along with greater cerebrovascular damage and AD pathology. This competitive binding mediates differential endothelial molecular responses depending on ApoE genotype, increasing cerebrovascular inflammation and mitochondria impairment in ApoE4 mice, while inducing vasculogenesis and protective changes in the presence of ApoE2. CONCLUSIONSOur study reveals a novel and dynamic endothelial ApoE-mCRP-CD31 pathway for AD pathogenesis during chronic inflammation and provides some insight into the opposing ApoE4-neurodegenerative and ApoE2-neuroprotective effects in AD. Clinical PerspectiveO_ST_ABSWHAT IS NEW?C_ST_ABSO_LICD31 is a competitive target of both mCRP and ApoE in brain endothelia in an ApoE-allele dependent pattern C_LIO_LImCRP increases CD31 phosphorylation in the brain endothelia and damages cerebrovasculature in ApoE4 carriers and AD brains C_LIO_LImCRP expression results in neuroprotective or neurodegenerative pathway activation in an ApoE-dependent manner C_LI WHAT ARE THE CLINICAL IMPLICATIONS?O_LIAlthough ApoE4 is a major genetic risk factor of AD, some ApoE4 carriers do not develop AD by the age of 90. C_LIO_LIElderly people often experience peripheral inflammatory attacks and develop chronic low-grade inflammation, which results in the formation and release of mCRP. Because CRP is routine clinical laboratory test, clinicians can use blood CRP level to predict AD risk in ApoE4 carriers. C_LIO_LIEvidence of Apoe4 genotype and chronic low-grade inflammation stages marked by elevated CRP levels should be targeted in personalized treatment and clinical trials for AD. C_LI

neuroscience↗

bioRxiv2021

Tmem138, a photoreceptor connecting cilium (CC) protein, is required for rhodopsin transport across the cilium and outer segment (OS) biogenesis

Photoreceptor connecting cilium (CC) is structurally analogous to the transition zone (TZ) of primary cilia and gates the molecular trafficking between the inner and the outer segment (OS). Retinal dystrophies with underlying CC defects are manifested in a broad array of syndromic conditions known as ciliopathies as well as non-syndromic retinal degenerations. Despite extensive studies, protein trafficking across the photoreceptor CC is largely unknown. Here we genetically inactivated mouse Tmem138, a gene encoding a ciliary membrane protein localized to the ciliary TZ and linked to Joubert syndrome (JBTS). Germline deletion of Tmem138 abolished OS morphogenesis followed by rapid photoreceptor degeneration. Tmem138 was found localized to the photoreceptor CC and, accordingly, the molecular compartments of the CC and axoneme of the mutant photoreceptors were altered despite ciliogenesis proceeding normally at the early stage of photoreceptor development. To gain further insights into Tmem138 function in OS biogenesis, we focused on trafficking of rhodopsin, the most abundant protein of the OS. Mislocalization of rhodopsin was readily observed as early as P5 in the mutant photoreceptors prior to growth of the OS. Ablation of Tmem138 in mature rods recapitulated the molecular changes in the germline mutants, causing well-formed outer segment discs to disintegrate accompanied by mislocalization of rhodopsin in the cell body. Furthermore, Tmem138 interacted with rhodopsin, and two additional CC compartment proteins Ahi1 and Tmem231, which were both altered in the mutant photoreceptors. Taken together, these results suggest that Tmem138 has a distinct role in gating the transport of rhodopsin and likely other OS bound proteins through formation of CC transport complex(es).

cell biology↗

bioRxiv2021

Metagenome-genome-wide association studies reveal human genetic impact on the oral microbiome

The oral microbiota contains billions of microbial cells, which could contribute to diseases in a number of body sites. Challenged by eating, drinking and dental hygiene on a daily basis, the oral microbiota is regarded as highly dynamic. Here, we report significant human genomic associations with the oral metagenome from more than 1,915 individuals, for both the tongue dorsum and saliva. We identified five genetic loci associated with oral microbiota at study-wide significance (p < 3.16 x 10-11). Four of the five associations were well replicated in an independent cohort of 1,439 individuals: rs1196764 at APPL2 with Prevotella jejuni, Oribacterium uSGB 3339 and Solobacterium uSGB 315; rs3775944 at the serum uric acid transporter SLC2A9 with Oribacterium uSGB 1215, Oribacterium uSGB 489 and Lachnoanaerobaculum umeaense; rs4911713 near OR11H1 with species F0422 uSGB 392; and rs36186689 at LOC105371703 with Eggerthia. Further analyses confirmed 84% (386/455 for tongue dorsum) and 85% (391/466 for saliva) of genetic-microbiota associations including 6 genome-wide significant associations mutually validated between the two niches. Human genetics accounted for at least 10% of oral microbiome compositions between individuals. Machine learning models showed that polygenetic risk score dominated over oral microbiome in predicting predisposing risk of dental diseases such as dental calculus and gingival bleeding. These findings indicate that human genetic differences are one explanation for a stable or recurrent oral microbiome in each individual.

genetics↗

bioRxiv2021

Genetic characteristics of human papillomavirus type 16, 18, 52 and 58 in southern China

Persistent infections of high-risk human papillomaviruses (HPVs) are the leading cause of cervical cancers. We collected cervical exfoliated cell samples from females in Changsha city, Hunan Province and obtained 358 viral genomes of four major HPV types, including HPV 16 (n=82), 18 (n=35), 52 (n=121) and 58 (n=100). The lineage/sublineage distribution of the four HPVs confirmed previous epidemiological reports, with the predominant prevailing sublineage as A4 (50%), A1 (37%) and A3 (13%) for HPV16, A1 (83%) for HPV18, B2 (86%) for HPV52 and A1 (65%), A3 (19%) and A2 (12%) for HPV58. We also identified two potentially novel HPV18 sublineages, i.e. A6 and A7. Virus mutation analysis further revealed the presence of HPV16 and HPV58 strains associated with potentially high oncogenicity. These findings expanded our knowledge on the HPV genetic diversity in China, providing valuable evidence to facilitate HPV DNA screening, vaccine effectiveness evaluation and control strategy development.

genomics↗

bioRxiv2021

Eye movements reveal spatiotemporal dynamics of active sensing and planning in navigation

Goal-oriented navigation is widely understood to depend upon internal maps. Although this may be the case in many settings, humans tend to rely on vision in complex, unfamiliar environments. To study the nature of gaze during visually-guided navigation, we tasked humans to navigate to transiently visible goals in virtual mazes of varying levels of difficulty, observing that they took near-optimal trajectories in all arenas. By analyzing participants eye movements, we gained insights into how they performed visually-informed planning. The spatial distribution of gaze revealed that environmental complexity mediated a striking trade-off in the extent to which attention was directed towards two complimentary aspects of the world model: the reward location and task-relevant transitions. The temporal evolution of gaze revealed rapid, sequential prospection of the future path, evocative of neural replay. These findings suggest that the spatiotemporal characteristics of gaze during navigation are significantly shaped by the unique cognitive computations underlying real-world, sequential decision making.

neuroscience↗

bioRxiv2021

Genome-wide CRISPR activation screen identifies novel receptors for SARS-CoV-2 entry

The ongoing pandemic of coronavirus disease 2019 (COVID-19) caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has been endangering worldwide public health and economy. SARS-CoV-2 infects a variety of tissues where the known receptor ACE2 is low or almost absent, suggesting the existence of alternative pathways for virus entry. Here, we performed a genome-wide barcoded-CRISPRa screen to identify novel host factors that enable SARS-CoV-2 infection. In addition to known host proteins, i.e. ACE2, TMPRSS2 and NRP1, we identified multiple host components, among which LDLRAD3, TMEM30A and CLEC4G were confirmed as functional receptors for SARS-CoV-2. All these membrane proteins bind directly to spikes N-terminal domain (NTD). Their essential and physiological roles have all been confirmed in either neuron or liver cells. In particular, LDLRAD3 and CLEC4G mediate SARS-CoV-2 entry and infection in a fashion independent of ACE2. The identification of the novel receptors and entry mechanisms could advance our understanding of the multiorgan tropism of SARS-CoV-2, and may shed light on the development of the therapeutic countermeasures against COVID-19.

microbiology↗

bioRxiv2021

Impact of blood storage conditions on the transcript profile of plasma cell-free RNA

BACKGROUNDPlasma cell-free RNA (cfRNA) are potential biomarkers for disease prediction and diagnosis. However, pre-analysis factors, such as the delay in blood processing and storage may lead to unreliable results, though no study has systematically evaluated the effect of blood storage conditions on the whole transcriptome of plasma cfRNA yet. METHODSWe collected peripheral blood samples from four healthy subjects and allowed them to stand at room temperature or 4{square} for different time periods (0h, 2h, 6h and 24h) prior to plasma separation. Then, plasma cfRNA stability was evaluated by measuring expression changes of cell-free mRNA, lncRNA and miRNA using high throughput sequencing-based profiling. Finally, their paired leukocyte RNA data were integrated to depict the effect of leukocytes on plasma cfRNA during storage. RESULTSPlasma mRNA and lncRNA presented high correlations (Pearson R2 [≥] 0.8) and fewer variations when blood was stored at 4{square} for 6 hours or stored at RT for 2 hours. miRNA was more stable, with minimal R2 of 0.86 at 4{square} for at least 24 hours or at RT for 6 hours. Correlations of plasma RNA and leukocyte RNA increased with the incubation time, and the relative proportion of neutrophils in plasma grown from 14.3% to 61.2% at RT (P = 0.004), indicating leukocyte RNA contamination. Besides, the tissue enriched genes in plasma were down-regulated with the extension of storage time. CONCLUSIONSOur results characterized the effects of short-term storage of blood samples on plasma cfRNA, which will facilitate further researches or clinical applications to avoid bias resulting from sample processing.

molecular biology↗

bioRxiv2021

fmr1 mutation interacts with sensory experience to alter the early development of behavior and sensory coding in zebrafish

While Autism Spectrum Disorders (ASDs) are developmental in origin little is known about how they affect the early development of behavior and sensory coding, or how this is modulated by the sensory environment. The most common inherited form of autism is Fragile X syndrome, caused by a mutation in FMR1. Here we show that zebrafish fmr1-/- mutant larvae raised in a naturalistic visual environment display early deficits in hunting behavior, tectal map development, tectal network properties and decoding of spatial stimuli. However when given a choice they preferred an environment with reduced visual stimulation, and rearing them in this environment improved these metrics. Older fmr1-/- fish showed differences in social behavior, spending more time observing a conspecific, but responding more slowly to social cues. Together these results help reveal how fmr1-/- changes the early development of vertebrate brain function, and how manipulating the environment could potentially help reduce these changes.

neuroscience↗

bioRxiv2021

Auxin-triggered changes in the Arabidopsis root tip (phospho)proteome reveal novel root growth regulators

Auxin plays a dual role in growth regulation and, depending on the tissue and concentration of the hormone, it can either promote or inhibit division and expansion processes in plants. Recent studies revealed that, beyond transcriptional reprogramming, alternative auxin-controlled mechanisms regulate root growth. Here, we explored the impact of different auxin concentrations on the root tip proteome and phosphoproteome, generating a unique resource. From the phosphoproteome data we pinpointed (novel) growth regulators, such as the RALF34-THE1 module. Our results together with previously published studies suggest that auxin, H+-ATPases, cell wall modifications and cell wall sensing receptor-like kinases are tightly embedded in a pathway regulating cell elongation. Furthermore, our study assigned a novel role to MKK2 as a regulator of primary root growth and a (potential) regulator of auxin biosynthesis and signalling, and suggests the importance of the MKK2 Thr31 phosphorylation site for growth regulation in the Arabidopsis root tip. ONE SENTENCE SUMMARYAn auxin-triggered Arabidopsis root tip (phospho)proteome reveals novel root growth regulators

plant biology↗

bioRxiv2021

Learning association for single-cell transcriptomics by integrating profiling of gene expression and alternative polyadenylation

Single-cell RNA-sequencing (scRNA-seq) has enabled transcriptome-wide profiling of gene expressions in individual cells. A myriad of computational methods have been proposed to learn cell-cell similarities and/or cluster cells, however, high variability and dropout rate inherent in scRNA-seq confounds reliable quantification of cell-cell associations based on the gene expression profile alone. Lately bioinformatics studies have emerged to capture key transcriptome information on alternative polyadenylation (APA) from standard scRNA-seq and revealed APA dynamics among cell types, suggesting the possibility of discerning cell identities with the APA profile. Complementary information at both layers of APA isoforms and genes creates great potential to develop cost-efficient approaches to dissect cell types based on multiple modalities derived from existing scRNA-seq data without changing experimental technologies. We proposed a toolkit called scLAPA for learning association for single-cell transcriptomics by combing single-cell profiling of gene expression and alternative polyadenylation derived from the same scRNA-seq data. We compared scLAPA with seven similarity metrics and five clustering methods using diverse scRNA-seq datasets. Comparative results showed that scLAPA is more effective and robust for learning cell-cell similarities and clustering cell types than competing methods. Moreover, with scLAPA we found two hidden subpopulations of peripheral blood mononuclear cells that were undetectable using the gene expression data alone. As a comprehensive toolkit, scLAPA provides a unique strategy to learn cell-cell associations, improve cell type clustering and discover novel cell types by augmentation of gene expression profiles with polyadenylation information, which can be incorporated in most existing scRNA-seq pipelines. scLAPA is available at https://github.com/BMILAB/scLAPA.

bioinformatics↗