Biology subjects

Huang, S.

Publications and source records attributed to Huang, S..

At least 325 records · Page 18Linked to original sources

bioRxiv2021

In vitro models of the human esophagus reveal ancestrally diverse response to injury

European Americans (EA) are more susceptible to esophageal tissue damage and inflammation when exposed to gastric acid and bile acid reflux and have a higher incidence of esophageal adenocarcinoma when compared to African Americans (AA). Population studies have implicated specific genes for these differences; however, the underlying cause for these differences is not well understood. We describe a robust long-term culture system to grow primary human esophagus in vitro, use single cell RNA sequencing to compare primary human biopsies to their in vitro counterparts, identify known and new molecular markers of basal cell types, and demonstrate that in vivo cellular heterogeneity is maintained in vitro. We further developed an ancestrally diverse biobank and a high-content, image based, screening assay to interrogate bile-acid injury response. These results demonstrated that AA esophageal cells responded significantly differently than EA-derived cells, mirroring clinical findings, having important implications for addressing disparities in early drug development pipelines.

cell biology↗

bioRxiv2021

A high-continuity and annotated tomato reference genome

Genetic and functional genomics studies require a high-quality genome assembly. Tomato (Solanum lycopersicum), an important horticultural crop, is an ideal model species for the study of fruit development. Here, we assembled an updated reference genome of S. lycopersicum cv. Heinz 1706 that was 799.09 Mb in length, containing 34,384 predicted protein-coding genes and 65.66% repetitive sequences. By comparing the genomes of S. lycopersicum and S. pimpinellifolium LA2093, we found a large number of genomic fragments probably associated with human selection, which may have had crucial roles in the domestication of tomato. Our results offer opportunities for understanding the evolution of the tomato genome and will facilitate the study of genetic mechanisms in tomato biology. Information for the assembled genome SLT1.0 was deposited both into the Genome Warehouse (GWH) database (https://bigd.big.ac.cn/gwh/) in the BIG Data Center under Accession Number GWHBAUD00000000.

genomics↗

bioRxiv2021

The Taxus genome provides insights into paclitaxel biosynthesis

The ancient gymnosperm genus Taxus is the exclusive source of the anticancer drug paclitaxel, yet no reference genome sequences are available for comprehensively elucidating the paclitaxel biosynthesis pathway. We have completed a chromosome-level genome of Taxus chinensis var. mairei with a total length of 10.23 Gb. Taxus shared an ancestral whole-genome duplication with the coniferophyte lineage and underwent distinct transposon evolution. We discovered a unique physical and functional grouping of CYP725As (cytochrome P450) in the Taxus genome for paclitaxel biosynthesis. We also identified a gene cluster in the taxadiene biosynthesis, which was mainly formed by gene duplications. This study will facilitate the elucidation of paclitaxel biosynthesis and unleash the biotechnological potential of Taxus. One Sentence SummaryA chromosome-level genome assembly of Taxus chinensis var. mairei uncovers its unique genome evolution process and genetic architectures for the paclitaxel biosynthesis pathway.

genomics↗

bioRxiv2021

Identification of heat responsive genes in pea stipules and anthers through transcriptional profiling

Field pea (Pisum sativum L.), a cool-season legume crop, is known for poor heat tolerance. Our previous work identified PR11-2 and PR11-90 as heat tolerant and susceptible lines in a recombinant inbred population. CDC Amarillo, a Canadian elite pea variety, was considered as another heat tolerant variety based on its similar field performance as PR11-2. This study aimed to characterize the differential transcription. Plants of these three varieties were stressed for 3h at 38{degrees}C prior to self-pollination, and RNAs from heat stressed anthers and stipules on the same flowering node were extracted and sequenced via the Illumina NovaSeq platform for the characterization of heat responsive genes. In silico results were further validated by qPCR assay. Differentially expressed genes (DEGs) were identified at log2 fold change, the three varieties shared 588 DEGs which were up-regulated and 220 genes which were down-regulated in anthers when subjected to heat treatment. In stipules, 879 DEGs (463/416 upregulation/downregulation) were consistent among varieties. The above heat-induced genes of the two plant organs were related to several biological processes i.e., response to heat, protein folding and DNA templated transcription. Ten gene ontology (GO) terms were over-represented in the consistently down-regulated DEGs of the two organs, and these terms were mainly related to cell wall macromolecule metabolism, lipid transport, lipid localization, and lipid metabolic processes. GO enrichment analysis on distinct DEGs of individual pea varieties suggested that heat affected biological processes were dynamic, and variety distinct responses provide insight into molecular mechanisms of heat-tolerance response. Several biological processes, e.g., cellular response to DNA damage stimulus in stipule, electron transport chain in anther that were only observed in heat induced PR11-2 and CDC Amarillo, and their relevance to field pea heat tolerance is worth further validation.

plant biology↗

bioRxiv2021

Isolation, screening, degradation characteristics of a quinclorac-degrading bacteria D and its potential of bioremediation for rice field environment polluted by quinclorac

Quinclorac (QNC) is a highly selective, hormonal, and low-toxic herbicide with a long duration. And the growth and development of subsequent crops are easily affected by QNC accumulated in the soil. In this paper, a QNC-degrading strain D was isolated and screened from the rice paddy soil. Through morphology, physiological and biochemical tests and 16Sr DNA gene analysis, strain D was identified as Cellulosimicrobium cellulans sp. And the QNC degradation characteristics of strain D were studied. Under the optimal culture conditions, the QNC-degrading rate was 45.9% after culturing for 21 days. The QNC-degrading efficiency of strain D in the field was evaluated by a simulated pot experiment. The results show that strain D can promote the growth of rice and QNC-degrading effectively. This research could provide a new bacterial species for microbial degradation of QNC and lay a theoretical foundation for further research on QNC remediation. ImportanceAt present, some QNC-degrading bacteria have been isolated from different environments, but there are no reports of Cellulosimicrobium cellulans sp. bacterial that could degrade QNC. In this study, a new QNC-degradation strain was selected from the paddy soil. The degradation characteristics of strain D were studied in detail. The results shown that strain D had a satisfactory quinclorac-degrading efficiency. Two degradation products of QNC by strain D were identified by HPLC-Q-TOF/MS: 3-pyridylacetic acid (138.0548 m/z) and 3-ethylpyridine (108.0805 m/z), which have not been reported before. The strain D had a potential ability of quinclorac-degrading effectively in the quinclorac-polluted paddy field environment.

microbiology↗

bioRxiv2021

OGUs enable effective, phylogeny-aware analysis of even shallow metagenome community structures

We introduce Operational Genomic Unit (OGU), a metagenome analysis strategy that directly exploits sequence alignment hits to individual reference genomes as the minimum unit for assessing the diversity of microbial communities and their relevance to environmental factors. This approach is independent from taxonomic classification, granting the possibility of maximal resolution of community composition, and organizes features into an accurate hierarchy using a phylogenomic tree. The outputs are suitable for contemporary analytical protocols for community ecology, differential abundance and supervised learning while supporting phylogenetic methods, such as UniFrac and phylofactorization, that are seldomly applied to shotgun metagenomics despite being prevalent in 16S rRNA gene amplicon studies. As demonstrated in one synthetic and two real-world case studies, the OGU method produces biologically meaningful patterns from microbiome datasets. Such patterns further remain detectable at very low metagenomic sequencing depths. Compared with taxonomic unit-based analyses implemented in currently adopted metagenomics tools, and the analysis of 16S rRNA gene amplicon sequence variants, this method shows superiority in informing biologically relevant insights, including stronger correlation with body environment and host sex on the Human Microbiome Project dataset, and more accurate prediction of human age by the gut microbiomes in the Finnish population. We provide Woltka, a bioinformatics tool to implement this method, with full integration with the QIIME 2 package and the Qiita web platform, to facilitate OGU adoption in future metagenomics studies. ImportanceShotgun metagenomics is a powerful, yet computationally challenging, technique compared to 16S rRNA gene amplicon sequencing for decoding the composition and structure of microbial communities. However, current analyses of metagenomic data are primarily based on taxonomic classification, which is limited in feature resolution compared to 16S rRNA amplicon sequence variant analysis. To solve these challenges, we introduce Operational Genomic Units (OGUs), which are the individual reference genomes derived from sequence alignment results, without further assigning them taxonomy. The OGU method advances current read-based metagenomics in two dimensions: (i) providing maximal resolution of community composition while (ii) permitting use of phylogeny-aware tools. Our analysis of real-world datasets shows several advantages over currently adopted metagenomic analysis methods and the finest-grained 16S rRNA analysis methods in predicting biological traits. We thus propose the adoption of OGU as standard practice in metagenomic studies.

bioinformatics↗

bioRxiv2021

Characterization of a glycolipid glycosyltransferase with broad substrate specificity from the marine bacterium Candidatus Pelagibacter sp. HTCC7211

In the marine environment, phosphorus availability significantly affects the lipid composition in many cosmopolitan marine heterotrophic bacteria, including members of the SAR11 clade and the Roseobacter clade. Under phosphorus stress conditions, non-phosphorus sugar-containing glycoglycerolipids are substitutes for phospholipids in these bacteria. Although these glycoglycerolipids play an important role as surrogates for phospholipids under phosphate deprivation, glycoglycerolipid synthases in marine microbes are poorly studied. In the present study, we biochemically characterized a glycolipid glycosyltransferase (GTcp) from the marine bacterium Candidatus Pelagibacter sp. HTCC7211, a member of the SAR11 clade. Our results showed that GTcp is able to act as a multifunctional enzyme by synthesizing different glycoglycerolipids with UDP-glucose, UDP-galactose, or UDP-glucuronic acid as sugar donors and diacylglycerol as the acceptor. Analyses of enzyme kinetic parameters demonstrated that Mg2+ notably changes the enzymes affinity for UDP-glucose, which improves its catalytic efficiency. Homology modelling and mutational analyses revealed binding sites for the sugar donor and the diacylglycerol lipid acceptor, which provided insights into the retaining mechanism of GTcp with its GT-B fold. A phylogenetic analysis showed that GTcp and its homologs form a group in the GT4 glycosyltransferase family. These results not only provide new insights into the glycoglycerolipid synthesis mechanism in lipid remodelling, but also describe an efficient enzymatic tool for future synthesis of bioactive molecules. ImportanceThe bilayer formed by membrane lipids serves as the containment unit for living microbial cells. In the marine environment, it has been firmly established that phytoplankton and heterotrophic bacteria can substitute phospholipids with non-phosphorus sugar-containing glycoglycerolipids in response to phosphorus limitation. However, little is known about how these glycoglycerolipids are synthesized. Here, we determined the biochemical characteristics of a glycolipid glycosyltransferase (GTcp) from the marine bacterium Candidatus Pelagibacter sp. HTCC7211. GTcp and its homologs form a group in the GT4 glycosyltransferase family, and can synthesize neutral glycolipids (MGlc-DAG and MGal-DAG) and an acidic glycolipid (MGlcA-DAG). We also uncover the key residues for DAG-binding through molecular docking, site-direct mutagenesis and subsequent enzyme activity assays. Our data provide new insights into the glycoglycerolipid synthesis mechanism in lipid remodelling.

microbiology↗

bioRxiv2021

Serotonin and dopamine modulate aging in response to food perception and availability

An organisms ability to perceive and respond to changes in its environment is crucial for its health and survival. Here we reveal how the most well-studied longevity intervention, dietary restriction (DR), acts in-part through a cell non-autonomous signaling pathway that is inhibited by the perception of attractive smells. Using an intestinal reporter for a key gene induced by DR but suppressed by attractive smells, we identify three compounds that block food perception in C. elegans, thereby increasing longevity as DR mimetics. These compounds clearly implicate serotonin and dopamine in limiting lifespan in response to food perception. We further identify an enteric neuron in this pathway that signals through the serotonin receptor 5-HT1A/ser-4 and dopamine receptor DRD2/dop-3. Aspects of this pathway are conserved in D. melanogaster and mammalian cells. Thus, blocking food perception through antagonism of serotonin or dopamine receptors is a plausible approach to mimic the benefits of dietary restriction.

molecular biology↗

bioRxiv2021

Mechanical bistability enabled by ectodermal compression facilitates Drosophila mesoderm invagination

Apical constriction driven by non-muscle myosin II ("myosin") provides a well-conserved mechanism to mediate epithelial folding. It remains unclear how contractile forces near the apical surface of a cell sheet drive out-of-the-plane bending of the sheet and whether myosin contractility is required throughout folding. By optogenetic-mediated acute inhibition of myosin, we find that during Drosophila mesoderm invagination, myosin contractility is critical to prevent tissue relaxation during the early, "priming" stage of folding but is dispensable for the actual folding step after the tissue passes through a stereotyped transitional configuration. The binary response suggests that the mesoderm is mechanically bistable during gastrulation. Combined modeling analysis and experimental measurements suggest that the observed mechanical bistability may arise from apicobasal shrinkage of the surrounding ectoderm, which promotes mesoderm invagination by facilitating a buckling transition. Our results suggest that Drosophila mesoderm invagination requires a joint action of local myosin contractility and mechanical bistability of the epithelium to trigger epithelial buckling.

developmental biology↗

bioRxiv2021

What makes our lungs look red?

What makes our lungs look red? This seems to be a naIve and trivial question. Indeed, all the textbooks tell us that what makes our body red is the presence of blood, or more precisely the red blood cells (RBC). Here we provide some experimental evidence as well arguments to prove that this belief is wrong, or only partially true. In fact, we identified an important population of red cells located outside of the blood vessels as highly compacted clusters, in the connection tissues of the lungs of several species of animals including rat and human being. These red cells possessed each a nucleus, expressing several biomarkers of different cell types such as PF4, vWF, SCF-1R, CD200R, TGF-b, etc... Interestingly, being morphologically heterogeneous, they react collectively to certain stimuli. For example, they aggregated on collagen fibers forming clusters of cells resembling that observed in vivo. The red cells may have some features of Hematopoietic Stem Cells, since they were capable of differentiating into other cell types such as alveolar macrophages. In nasal polyps, these cells formed vessel-like structures, confined within a CD31-positive tubes. Upon exposure to toxic chemicals, they formed dense networks, suggesting a possible role in coagulation. Furthermore, the number of these red cells was greatly increased in the lungs of diseased donors, especially in the lungs of CF patients. Instead of being secreted as what happens in normal red cells, vWF proteins were tethered on the cytoplasmic membrane of the red cells isolated from the lungs of CF donors, which may explain at least partially the fibrotic nature of the CF lungs. Taken together, we conclude that what makes the lungs look red is not the red blood cells, rather a distinct population of red cells, we call them Red Soma Cells (RSC). We believe that the discovery and characterization of this important population of cells will have profound theoretic as well as therapeutic implications.

cell biology↗

bioRxiv2021

The level of synovial human VEGFA, IL-8 and MIP-1α correlate with truncation of lubricin glycans in osteoarthritis

Osteoarthrithis (OA) is an endemic disease due to the increase of the worlds elderly population. Previously thought to be a consequence of an imbalance between cartilage degradation and biosynthesis, it is now recognized as a disease also involving inflammation, hence influencing the level of inflammatory cytokines, growth factors and chemokines. Lubricin is a mucin type molecule where its OA induced glycosylation truncation propels a deteriorating lubrication of the articular cartilage. The objective of this study was to explore the OA driven truncation of O-linked glycosylation of synovial lubricin and its cross talk with systemic and local (synovial fluid, SF) inflammation. We compared the systemic level of cytokines/chemokine in OA patients and controls plasma with their local level in SF using a 44 plex screen. The level of 27 cytokines and chemokines was consistently measured in both plasma and SF. The data showed that the levels of cytokines and chemokines in OA plasma display limited correlation to their counterpart in SF. The level of synovial IL-8 and MIP-1 and VEGFA in OA patients, but not their plasma level, where the only cytokines that displayed a significant correlation to the observed lubricin O-linked glycosylation truncation. These cytokines were also shown to be upregulated exposing fibroblast like synoviocytes from healthy and OA patients to recombinant lubricin with truncated glycans mainly consisting of Tn-antigens, while lubricin with sialylated and non-sialylated T anigens did not have any effect. The data suggest that truncated glycans of lubricin, as found in OA, promotes the synovial cytokine production and exerebate the local synovial inflammation.

pathology↗

bioRxiv2021

In vivo co-registered hybrid-contrast imaging by successive photoacoustic tomography and magnetic resonance imaging

Magnetic resonance imaging (MRI) and photoacoustic tomography (PAT) are two advanced imaging modalities that offer two distinct image contrasts: MRI has a multi-parameter contrast mechanism that provides excellent anatomical soft tissue contrast, whereas PAT is capable of mapping tissue physiological metabolism and exogenous contrast agents with optical specificity. Attempts have been made to integrate these two modalities, but rigid and reliable registration of the images for in vivo imaging is still challenging. In this paper, we present a complete hardware-software solution for the successive acquisition and co-registration of PAT and MRI images in in vivo animal studies. Based on commercial PAT and MRI scanners, our solution includes a 3D-printed dual-modality animal imaging bed, a 3-D spatial image co-registration algorithm with bi-model markers, and a robust modality switching protocol for in vivo imaging studies. Using the proposed solution, we successfully demonstrated co-registered hybrid-contrast PAT-MRI imaging that simultaneously display multi-scale anatomical, functional and molecular characteristics on healthy and cancerous living mice. Week-long longitudinal dual-modality imaging of tumor development reveals information on size, border, vascular pattern, blood oxygenation, and molecular probe metabolism of the tumor micro-environment at the same time. Additionally, by incorporating soft-tissue information in the co-registered MRI image, we further show that PAT image quality could be enhanced by MRI-guided light fluence correction. The proposed methodology holds the promise for a wide range of pre-clinical research applications that benefit from the PAT-MRI dual-modality image contrast.

bioengineering↗

bioRxiv2021

Inter-tissue convergence of gene expression and loss of cellular identity during ageing

Developmental trajectories of gene expression may reverse in their direction during ageing, a phenomenon previously linked to cellular identity loss. Our analysis of cerebral cortex, lung, liver and muscle transcriptomes of 16 mice, covering development and ageing intervals, revealed widespread but tissue-specific ageing-associated expression reversals. Cumulatively, these reversals create a unique phenomenon: mammalian tissue transcriptomes diverge from each other during postnatal development, but during ageing, they tend to converge towards similar expression levels, a process we term Divergence followed by Convergence, or DiCo. We found that DiCo was most prevalent among tissue-specific genes and associated with loss of tissue identity, which is confirmed using data from independent mouse and human datasets. Further, using publicly available single-cell transcriptome data, we showed that DiCo could be driven both by alterations in tissue cell type composition and also by cell-autonomous expression changes within particular cell types.

genomics↗

bioRxiv2021

Immune elicitation by sensing the conserved signature from phytocytokines and microbes via the Arabidopsis MIK2 receptor

Sessile plants encode a large number of small peptides and cell surface-resident receptor kinases, most of which have unknown functions. Here, we report that the Arabidopsis receptor kinase MALE DISCOVERER 1-INTERACTING RECEPTOR-LIKE KINASE 2 (MIK2) recognizes the conserved signature motif of SERINE RICH ENDOGENOUS PEPTIDEs (SCOOPs) from plants as well as proteins present in fungal Fusarium spp. and bacterial Comamonadaceae, and elicits potent immune responses. SCOOP signature peptides trigger diverse immune and physiological responses in a MIK2-dependent manner with a sub-nanomolar sensitivity and directly bind to the extracellular leucine rich-repeat domain of MIK2 in vivo and in vitro, indicating that MIK2 is the receptor of SCOOP peptides. Perception of SCOOP peptides induces the association of MIK2 and the coreceptors SOMATIC EMBRYOGENESIS RECEPTOR KINASE 3 (SERK3) and SERK4 and relays the signaling through the cytosolic receptor-like kinases BOTRYTIS-INDUCED KINASE 1 (BIK1) and AVRPPHB SUSCEPTIBLE1 (PBS1)-LIKE 1 (PBL1). Our study identified a unique plant receptor that bears a dual recognition capability sensing the conserved peptide motif from phytocytokines and microbial proteins via a convergent signaling relay to ensure a robust immune response.

plant biology↗

bioRxiv2021

Differential chemoarchitecture of Purkinje neurons contributes to intrinsic firing properties

Purkinje cells (PCs) are central to cerebellar information coding and appreciation for the diversity of their firing patterns and molecular profiles is growing. Heterogeneous subpopulations of PCs have been identified that display differences in intrinsic firing properties without clear mechanistic insight into what underlies the divergence in firing parameters. Although long used as a general PC marker, we report that the calcium binding protein parvalbumin labels a subpopulation of PCs with a conserved distribution pattern across the animals examined. We trained a convolutional neural network to recognize the parvalbumin-positive subtype and create maps of whole cerebellar distribution and find that PCs within these areas have differences in spontaneous firing that can be modified by altering calcium buffer content. These subtypes also show differential responses to potassium and calcium channel blockade, suggesting a mechanistic role for variability in PC intrinsic firing through differences in ion channel composition. It is proposed that ion channels drive the diversity in PC intrinsic firing phenotype and parvalbumin calcium buffering provides capacity for the highest firing rates observed. These findings open new avenues for detailed classification of PC subtypes.

neuroscience↗

bioRxiv2021

Large-scale genetic association and single cell accessible chromatin mapping defines cell type-specific mechanisms of type 1 diabetes risk

Translating genome-wide association studies (GWAS) of complex disease into mechanistic insight requires a comprehensive understanding of risk variant effects on disease-relevant cell types. To uncover cell type-specific mechanisms of type 1 diabetes (T1D) risk, we combined genetic association mapping and single cell epigenomics. We performed the largest to-date GWAS of T1D in 489,679 samples imputed into 59.2M variants, which identified 74 novel association signals including several large-effect rare variants. Fine-mapping of 141 total signals substantially improved resolution of causal variant credible sets, which primarily mapped to non-coding sequence. To annotate cell type-specific regulatory mechanisms of T1D risk variants, we mapped 448,142 candidate cis-regulatory elements (cCREs) in pancreas and peripheral blood mononuclear cell types using snATAC-seq of 131,554 nuclei. T1D risk variants were enriched in cCREs active in CD4+ T cells as well as several additional cell types including pancreatic exocrine acinar and ductal cells. High-probability T1D risk variants at multiple signals mapped to exocrine-specific cCREs including novel loci near CEL, GP2 and CFTR. At the CFTR locus, the likely causal variant rs7795896 mapped in a ductal-specific distal cCRE which regulated CFTR and the risk allele reduced transcription factor binding, enhancer activity and CFTR expression in ductal cells. These findings support a role for the exocrine pancreas in T1D pathogenesis and highlight the power of combining large-scale GWAS and single cell epigenomics to provide insight into the cellular origins of complex disease.

genetics↗

bioRxiv2021

Bottom-up inputs are required for the establishment of top-down connectivity onto cortical layer 1 neurogliaform cells

Higher order feedback projections to sensory cortical areas converge on layer 1 (L1), the primary site for integration of top-down information via the apical dendrites of pyramidal neurons and L1 GABAergic interneurons. Here, we investigated the contribution of early thalamic inputs onto L1 interneurons for the establishment of top-down inputs in the primary visual cortex. We find that bottom-up thalamic inputs predominate during early L1 development and preferentially target neurogliaform cells. We find that these projections are critical for the subsequent strengthening of feedback inputs from the anterior cingulate cortex. Enucleation or selective removal of thalamic afferents blocked this phenomenon. Notably, while early activation of anterior cingulate afferents resulted in a premature strengthening of these top-down inputs to neurogliaform cells, this was also dependent on thalamic inputs. Our results demonstrate that the proper establishment of top-down feedback inputs critically depends on bottom-up inputs from the thalamus during early postnatal development.

neuroscience↗

bioRxiv2021

Semi-supervised Calibration of Risk with Noisy Event Times (SCORNET) Using Electronic Health Record Data

Leveraging large-scale electronic health record (EHR) data to estimate survival curves for clinical events can enable more powerful risk estimation and comparative effectiveness research. However, use of EHR data is hindered by a lack of direct event times observations. Occurrence times of relevant diagnostic codes or target disease mentions in clinical notes are at best a good approximation of the true disease onset time. On the other hand, extracting precise information on the exact event time requires laborious manual chart review and is sometimes altogether infeasible due to a lack of detailed documentation. Current status labels - binary indicators of phenotype status during follow up - are significantly more efficient and feasible to compile, enabling more precise survival curve estimation given limited resources. Existing survival analysis methods using current status labels focus almost entirely on supervised estimation, and naive incorporation of unlabeled data into these methods may lead to biased results. In this paper we propose Semi-supervised Calibration of Risk with Noisy Event Times (SCORNET), which yields a consistent and efficient survival curve estimator by leveraging a small size of current status labels and a large size of imperfect surrogate features. In addition to providing theoretical justification of SCORNET, we demonstrate in both simulation and real-world EHR settings that SCORNET achieves efficiency akin to the parametric Weibull regression model, while also exhibiting non-parametric flexibility and relatively low empirical bias in a variety of generative settings.

bioinformatics↗