Biology subjects

Hwang, W.

Publications and source records attributed to Hwang, W..

At least 19 recordsLinked to original sources

bioRxiv2026

Network analysis of correlated conformational dynamics across the evolution of influenza A virus NS1

Protein evolution requires sequence changes to alter molecular properties without disrupting the dynamic organization necessary for function. How residue-level networks of correlated motions accommodate such mutations remains poorly understood. Here, we investigate the evolution of correlated conformational dynamics in nonstructural protein 1 (NS1) variants of influenza A viruses that emerged between 1918 and 2004. From molecular dynamics (MD) simulations we built residue networks in which edge weights reflect generalized correlations of conformational motion. The network of correlated dynamics is characterized using centrality analysis. Overall network properties are well-conserved across NS1 variants, indicating evolutionary robustness of the dynamical network. Residues ranked highly across multiple centrality measures form a predominantly hydrophobic cluster in the protein core. These residues exhibited low conformational entropy and high sequence conservation, consistent with rigid, coordinated motions. In contrast, residues exhibiting large differences in betweenness and closeness centrality ranks occupied flexible, solvent-exposed regions surrounding the core network. These residues also showed lower evolutionary conservation. These findings reveal a core-periphery structure in which a rigid hydrophobic network enables long-range dynamic connectivity while more adaptable peripheral regions accommodate evolutionary sequence variation. This architecture provides a mechanistic framework for long-range epistatic interactions without compromising the underlying structural framework.

biophysics↗

bioRxiv2026

Spatially pooling photon information enables photon-efficient quantitative imaging

Quantitative fluorescence imaging techniques such as fluorescence lifetime imaging microscopy and hyperspectral imaging infer molecular contrast from photons distributed across spatial pixels and temporal or spectral channels. In the few-photon regime, however, conventional pixel-wise analysis discards the spatial relationships imposed across neighboring pixels by the microscope point-spread function (PSF). Here we show that this spatially distributed information can be recovered without prior knowledge of emitter positions, spatial support or component assignments. We introduce SPOOL (Spatially Pooled Optical Observation Likelihood), a training-free Poisson inverse framework that jointly recovers source-space amplitudes and quantitative contrast by combining the PSF with temporal-decay or spectral-response dictionaries. For an isolated source, the attainable precision gain is governed by a dimensionless optical quantity: the PSF width expressed in detector pixels. The predicted gain therefore scales with optical sampling rather than with the physical origin of the contrast. The model predicts that lifetime-precision gain scales approximately linearly with the number of pixels spanning the PSF full width at half maximum, a scaling reproduced by Monte Carlo simulations. At one detected photon per foreground pixel, the reconstruction reduces lifetime dispersion sixfold in fluorescent-bead experiments and decreases the lifetime root-mean-square error relative to a high-photon reference from 1.19 to 0.45 ns in dual-labeled cells. The same framework transfers unchanged to hyperspectral imaging, recovering spectral contrast from generic emission bands without prior fluorophore spectra.

biophysics↗

bioRxiv2026

PSF-Driven Spatio-Temporal Blending in Fluorescence Lifetime Imaging Microscopy and Its Mitigation via Mean-Shift Super-Resolution-Based Masking.

Fluorescence Lifetime Imaging Microscopy (FLIM) enables quantitative mapping of molecular environments in living systems with high biochemical specificity. However, spatial overlap dictated by the diffraction-limited point spread function (PSF) causes a mixing of temporal signals: photons from neighboring emitters collected within the same pixel yield composite decay profiles, generating apparent intermediate lifetimes that can be mistaken for variations in the local molecular environment. We introduce a workflow that applies Mean-Shift Super-Resolution (MSSR) to raw intensity data to generate intensity-derived spatial masks prior to phasor-based lifetime analysis. The method is computationally efficient and preserves decay kinetics because it operates on intensity-derived spatial information rather than modifying temporal data. In U2OS cells labeled with spectrally-overlapping fluorophores, phasor analysis reveals an intermediate lifetime population localized at PSF-overlap interfaces, consistent with optical mixing rather than intrinsic lifetime heterogeneity. MSSR-derived masking suppressed this mixed population while preserving stable phasor cluster centers -i.e. the distribution of similar phasor coordinates in the phasor plane- for each fluorophore. Simulations of strictly monoexponential fluorescence decay emitters further show that blended lifetime decay profiles are present at separations up to 4{sigma} and becomes maximal near [~]1.6{sigma}, indicating that conventional spatial resolution criteria can underestimate lifetime cross-talk. Application of this workflow to three-component FLIM showed also a reduced overlap of pixel distributions in phasor plots while maintaining distinct lifetime signatures. Overall, MSSR-based spatial refinement provides an accessible strategy to improve the spatial resolution while maintaining accuracy of FLIM measurements.

biophysics↗

bioRxiv2025

Atomistic TCR-ligand interactions instruct memory T-cell differentiation

Memory CD8 T cells provide durable protection against recurrent infection and cancer, but how distinct memory fates are specified remains unclear. Here, we define the functional landscape of 242 murine CD8 TCR{beta} clonotypes specific for influenza NP366-374/H-2D ligand (pMHC) by integrating single-cell transcriptomics and TCR sequencing with force-dependent biophysics, structural analysis, and in vivo investigation of memory development. Fate essentially tracks with TCR mechanotransduction: TCM-associated clonotypes preferentially engage pMHC through TCR{beta}, whereas TEM-associated clonotypes preferentially engage through TCR. Clonotypes with balanced, bipolar signaling generate both memory subsets, optimize expansion, and display broad heterosubtypic crossreactivity but with exhaustion susceptibility. In contrast, diverse TCM clonotypes are largely clonal singlets with considerable mutant-epitope recognition in aggregate and stemness preservation, revealing complementary strategies for immunodefense coverage. Atomistically, subtle pMHC contact differences and load transmission through the TCR holoreceptor appear to tune phosphorylation, memory differentiation, and functional durability, features applicable to adoptive T-cell immunotherapies.

immunology↗

bioRxiv2025

RE3DB: A multi-omics phylogenomics platform for rice E3 ubiquitin ligases identifies novel regulators of pollen germination

The ubiquitin proteasome system (UPS) shapes rice development and stress responses through E3 ubiquitin ligases, yet roles of E3 ligases in male gametophyte function remain undiscovered and plant-focused functional informatics resource is absent. We developed the Rice E3 Ubiquitin Ligase Database (RE3DB; https://re3db.khu.ac.kr/) and used it to nominate F-box ligases involved in rice (Oryza sativa) pollen germination. RE3DB combines 15 published datasets comprising 1,602 rice E3-ligase genes; RNA-seq across 25 tissues and stress conditions; protein-to-transcript ratios from 14 tissues; six high-confidence protein-protein interaction resources; and family level phylogenies. Integrated modules provide (i) classification/annotation, (ii) E3-substrate prediction from interaction evidence, co-expression and coordinated proteolytic turnover, and (iii) functional-redundancy assessment using phylogenetic heatmaps. Applying RE3DB, we found extensive reproductive enrichment: 284 F-box genes are preferentially expressed in mature anthers. Comparative transcriptomics across nine male-sterile rice mutants identified 111 F-box genes significantly downregulated in at least one mutant (log2 fold change < -1), resolving distinct co-regulation clusters. Integration with a literature-curated regulatory network positions previously uncharacterized F-box genes within modules governing pollen-tube initiation and elongation, consistent with regulation by the RUPO receptor-like kinase and the OsMADS62/63/68 transcription-factor complex. RE3DB is the first comprehensive multi-omics web platform dedicated to functional characterization of rice E3 ligases. By unifying classification, interaction and expression-based inference, and redundancy analysis, it accelerates hypothesis-driven discovery and guides targeted mutagenesis in molecular plant science. The prioritized F-box cohort provides testable candidates for UPS-mediated control of pollen germination and tube growth, and the framework is readily extensible to other crops, holding significant potential for agricultural biotechnology.

plant biology↗

bioRxiv2025

Multimodal gene and targeted drug therapy for chronic myelogenous leukemia: Computational target analysis and therapeutic validation

Developing an efficient and safe therapy necessitates a mechanistic understanding of the complex underlying pathology and manipulation of the multiple pathways at the molecular and genetic level. Network-based simulation of chronic myeloid leukemia (CML), a relatively well-understood cancer model, revealed the dynamics of simultaneously expressing pro-apoptotic BIM and silencing pro-survival MCL-1 in combination with the BCR-ABL-targeted tyrosine kinase inhibitor dasatinib. Viral/nonviral chimeric nanoparticles (ChNPs) composed of a BIM-expressing adeno-associated virus (AAV) core and a degradable polymeric shell that encapsulates MCL-1 siRNA (BIM/MCL-1 ChNPs) synergistically and selectively killed BCR-ABL+ CML cells in combination with dasatinib. In a mouse CML model, the BIM/MCL-1 ChNPs and dasatinib combination therapy suppressed proliferation of BCR-ABL+ hematopoietic cells and prevented leukemic infiltration of organs. The synergistic anti-leukemic effect was further pronounced in an acute phase model of the disease. This study investigated a strategy of developing a versatile and tunable multimodal therapy assisted by a computational toolset that analyzes the molecular foundation of a disease and predicts therapeutic response. The interdisciplinary approach developed and validated in this study can be used in discovering new therapies for cancer and other diseases. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=182 SRC="FIGDIR/small/656072v1_ufig1.gif" ALT="Figure 1"> View larger version (43K): org.highwire.dtl.DTLVardef@1e3632borg.highwire.dtl.DTLVardef@1839f12org.highwire.dtl.DTLVardef@4060c0org.highwire.dtl.DTLVardef@8ff0de_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOGraphical abstractC_FLOATNO summarizing the study design and approach. A schematic representation of the integrated in silico and in vivo pipeline utilized in the study. The workflow begins with in silico simulations, including Boolean network modeling and protein-protein interaction (PPI) network analyses, leading to target discovery, optimized nanoparticle design, and validation in pathological contexts. This was followed by therapeutic efficacy assessments of BIM/MCL-1 ChNPs and their combination with dasatinib in BCR-ABL+ leukemia models in vitro and in vivo. C_FIG

cancer biology↗

bioRxiv2025

Utilising multi-modal data-driven network analysis to identify monotherapy and combinational therapy targets in SOX2-dependent squamous cell lung cancer

Drug discovery requires a deep understanding of disease mechanisms, making the integration and analysis of diverse, multi-modal data types essential. These data types, spanning omics types, disease- associated, and biological pathway information, often originate from disparate sources and must be combined to uncover critical insights. We have developed iPANDDA (in-silico Pipeline for Agnostic Network-based Drug Discovery Analysis), a computational pipeline that integrates multi-modal data through a network-based approach to predict candidate drug target proteins for specific diseases. We applied iPANDDA to lung squamous cell carcinoma (LUSC), a subtype of non-small cell lung cancer (NSCLC) that accounts for approximately 25% of all lung cancer cases globally. Despite advances in cancer therapeutics, targeted therapies specifically approved for LUSC remain lacking, exacerbated by the shortage of robust models for studying LUSC carcinogenesis and therapeutic responses. The SOX2 gene, amplified in about 50% of LUSC patients, plays a critical role in driving and maintaining the cancer phenotype. Using iPANDDA, we identified relevant therapeutic targets for SOX2-dependent LUSC. Selected candidate drug targets were validated in vitro using cell-based models. We conducted target inhibition studies in both SOX2-dependent and non-SOX2-dependent cell lines, evaluating the effects of inhibition and knockout through cell viability assays. Our findings confirmed key monotherapy and combination therapy targets for SOX2-focused models. Specifically, we validated the AKT and mTOR complexes as promising therapeutic targets for LUSC. Additionally, we identified potential pathways for developing novel combination therapies targeting SOX2-dependent LUSC. iPANDDA offers a robust approach to refining and focusing therapeutic strategies for diseases with unmet clinical needs.

bioinformatics↗

bioRxiv2024

Load-based divergence in the dynamic allostery of two TCRs recognizing the same pMHC

Increasing evidence suggests that mechanical load on the {beta} T cell receptor (TCR) is crucial for recognizing the antigenic peptide-loaded major histocompatibility complex (pMHC) molecule. Our recent all-atom molecular dynamics (MD) simulations revealed that the inter-domain motion of the TCR is responsible for the load-induced catch bond behavior of the TCR-pMHC complex and peptide discrimination. To further examine the generality of the mechanism, we perform all-atom MD simulations of the B7 TCR under different conditions for comparison with our previous simulations of the A6 TCR. The two TCRs recognize the same pMHC and have similar interfaces with pMHC in crystal structures. We find that the B7 TCR-pMHC interface stabilizes under [~]15-pN load using a conserved dynamic allostery mechanism that involves the asymmetric motion of the TCR chassis. However, despite forming comparable contacts with pMHC as A6 in the crystal structure, B7 has fewer high-occupancy contacts with pMHC and exhibits higher mechanical compliance during the simulation. These results indicate that the dynamic allostery common to the TCR{beta} chassis can amplify slight differences in interfacial contacts into distinctive mechanical responses and nuanced biological outcomes.

biophysics↗

bioRxiv2024

Evolutionary rewiring of the dynamic network underpinning allosteric epistasis in NS1 of influenza A virus

Viral proteins frequently mutate to evade or antagonize host innate immune responses, yet the impact of these mutations on the molecular energy landscape remains unclear. Epistasis, the intramolecular communications between mutations, often renders the combined mutational effects unpredictable. Nonstructural protein 1 (NS1) is a major virulence factor of the influenza A virus (IAV) that activates host PI3K by binding to its p85{beta} subunit. Here, we present the deep analysis for the impact of evolutionary mutations in NS1 that emerged between the 1918 pandemic IAV strain and its descendant PR8 strain. Our analysis reveal how the mutations rewired inter-residue communications which underlies long-range allosteric and epistatic networks in NS1. Our findings show that PR8 NS1 binds to p85{beta} with approximately 10-fold greater affinity than 1918 NS1 due to allosteric mutational effects. Notably, these mutations also exhibited long-range epistatic effects. NMR chemical shift perturbation and methyl-axis order parameter analyses revealed that the mutations induced long-range structural and dynamic changes in PR8 NS1, enhancing its affinity to p85{beta}. Complementary MD simulations and graph-based network analysis uncover how these mutations rewire dynamic residue interaction networks, which underlies the long-range epistasis and allosteric effects on p85{beta}-binding affinity. Significantly, we find that conformational dynamics of residues with high betweenness centrality play a crucial role in communications between network communities and are highly conserved across influenza A virus evolution. These findings advance our mechanistic understanding of the allosteric and epistatic communications between distant residues and provides insight into their role in the molecular evolution of NS1.

biophysics↗

bioRxiv2024

Antibody responses in Klebsiella pneumoniae bloodstream infection: a cohort study

BackgroundKlebsiella pneumonia (Kpn) is the fourth leading cause of infection-related deaths globally, yet little is known about human antibody responses to invasive Kpn. In this study, we sought to determine whether the O-specific polysaccharide (OPS) antigen, a vaccine candidate, is immunogenic in humans with Kpn bloodstream infection (BSI). We also sought to define the cross-reactivity of human antibody responses among structurally related Kpn OPS subtypes and to assess the impact of capsule production on OPS-targeted antibody binding and function. MethodsWe measured plasma antibody responses to OPS (and MrkA, a fimbrial protein) in a cohort of patients with Kpn BSI and compared these with controls, including a cohort of healthy individuals and a cohort of individuals with Enterococcus BSI. We performed flow cytometry to measure the impact of Kpn capsule production on whole cell antibody binding and complement deposition, utilizing patient isolates with variable levels of capsule production and isogenic capsule-deficient strains derived from these isolates. FindingsWe enrolled 69 patients with Kpn BSI. Common OPS serotypes accounted for 57/69 (83%) of infections. OPS was highly immunogenic in patients with Kpn BSI, and peak OPS-IgG antibody responses in patients were 10 to 30-fold higher than antibody levels detected in healthy controls, depending on the serotype. There was significant cross-reactivity among structurally similar OPS subtypes, including the O1v1/O1v2, O2v1/O2v2 and O3/O3b subtypes. Physiological amounts of capsule produced by both hyperencapsulated and non-hyperencapsulated Kpn significantly inhibited OPS-targeted antibody binding and function. InterpretationOPS was highly immunogenic in patients with Kpn BSI, supporting its potential as a candidate vaccine antigen. The strong cross-reactivity observed between similar OPS subtypes in humans with Kpn BSI suggests that it may not be necessary to include all subtypes in an OPS-based vaccine. However, these observations are tempered by the fact that capsule production, even in non-highly encapsulated strains, has the potential to interfere with OPS antibody binding. This may limit the effectiveness of vaccines that exclusively target OPS. FundingNational Institute of Allergy and Infectious Diseases at the National Institutes of Health. Research in ContextO_ST_ABSEvidence before this studyC_ST_ABSDespite the potential of O-specific polysaccharide (OPS) as a vaccine antigen against Klebsiella pneumoniae (Kpn), the immunogenicity of OPS in humans remains largely unstudied, creating a significant knowledge gap with regard to vaccine development. A search of PubMed for publications up to March 18, 2024, using the terms "Klebsiella pneumoniae" and "O-specific polysaccharide" or "O-antigen" or "lipopolysaccharide" revealed no prior studies addressing OPS antibody responses in humans with Kpn bloodstream infections (BSI). One prior study1 evaluated antibody response to a single lipopolysaccharide (which contains one subtype of OPS) in humans with invasive Kpn infection; however, in this study OPS typing of the infecting strains and target antigen were not described. Added value of this studyOur investigation into OPS immunogenicity in a human cohort marks a significant advance. Analyzing plasma antibody responses in 69 patients with Kpn BSI, we found OPS to be broadly immunogenic across all the types and subtypes examined, and there was significant cross-reactivity among structurally related OPS antigens. We also demonstrated that Kpn capsule production inhibit OPS antibody binding and the activation of complement on the bacterial surface, even in classical Kpn strains expressing lower levels of capsule. Implications of all the available evidenceWhile the immunogenicity and broad cross-reactivity of OPS in humans with Kpn BSI suggests it is a promising vaccine candidate, the obstruction of OPS antibody binding and engagement by physiologic levels of Kpn capsule underscores the potential limitations of an exclusively OPS-antigen based vaccine for Kpn. Our study provides insights for the strategic development of vaccines aimed at combating Kpn infections, an important antimicrobial resistant pathogen.

microbiology↗

bioRxiv2023

Asymmetric framework motion of TCR alpha beta controls load-dependent peptide discrimination

Mechanical force is critical for the interaction between an {beta}T cell receptor (TCR) and a peptide-bound major histocompatibility complex (pMHC) molecule to initiate productive T-cell activation. However, the underlying mechanism remains unclear. We use all-atom molecular dynamics simulations to examine the A6 TCR bound to HLA-A*02:01 presenting agonist or antagonist peptides under different extensions to simulate the effects of applied load on the complex, elucidating their divergent biological responses. We found that TCR and {beta} chains move asymmetrically, which impacts the interface with pMHC, in particular the peptide-sensing CDR3 loops. For the wild-type agonist, the complex stabilizes in a load-dependent manner while antagonists destabilize it. Simulations of the C{beta} FG-loop deletion, which reduces the catch bond response, and simulations with in silico mutant peptides further support the observed behaviors. The present results highlight the combined role of interdomain motion, fluctuating forces, and interfacial contacts in determining the mechanical response and fine peptide discrimination by a TCR, thereby resolving the conundrum of nearly identical crystal structures of TCR{beta}-pMHC agonist and antagonist complexes.

biophysics↗

bioRxiv2023

Inadequate structural constraint on Fab approach rather than paratope elicitation limits HIV-1 MPER vaccine utility

Broadly neutralizing antibodies (bnAbs) against HIV-1 target conserved epitopes, thereby inhibiting viral entry. Yet surprisingly, those recognizing linear epitopes in the HIV-1 gp41 membrane proximal external region (MPER) are elicited neither by peptide nor protein scaffold vaccines. Here, we observe that while Abs generated by MPER/liposome vaccines may exhibit human bnAb-like paratopes, B-cell programming without constraints imposed by the gp160 ectodomain selects Abs unable to access the MPER within its native "crawlspace". During natural infection, the flexible hinge of IgG3 partially mitigates steric occlusion of less pliable IgG1 subclass Abs with identical MPER specificity, until affinity maturation refines entry mechanisms. The IgG3 subclass maintains B-cell competitiveness, exploiting bivalent ligation resulting from greater intramolecular Fab arm length, offsetting weak antibody affinity. These findings suggest future immunization strategies.

immunology↗

bioRxiv2022

Pre-T cell receptor Self-MHC Sampling Restricts Thymocyte Dedifferentiation

Summary paragraphProgramming T lymphocytes to distinguish self from non-self is a vital, multi-step process arising in the thymus1-4. Signalling through the pre-T cell receptor (preTCR), a CD3-associated heterodimer comprising an invariant pT chain and a clone-specific {beta} chain, constitutes a critical early checkpoint in thymocyte development within the {beta} T-cell lineage5, 6. Recent work demonstrates that preTCRs arrayed on double negative (DN) thymocytes, like {beta} TCRs appearing on double positive (DP) thymocytes, ligate peptides bound to MHC molecules (pMHC) on thymic stroma but via a different molecular docking strategy7-10. Here we show the consequences of those distinctive interactions for thymocyte progression, using synchronized fetal thymic progenitor cultures differing in the presence or absence of pMHC on support stroma, determining single cell transcriptomes at key thymocyte developmental transitions. Although MHC negative stroma fosters {beta} T lymphocyte differentiation, the absence of pMHC-preTCR interplay leads to deviant thymocyte transcriptional programming associated with de-differentiation. Highly proliferative DN and DP subsets with antecedent characteristics of T cell lymphoblastic and myeloid malignancies emerge. Thus, at least in vitro, beyond fostering {beta} chain repertoire broadening for subsequent {beta} TCR utilization, preTCR-pMHC interaction limits cellular plasticity to facilitate normal thymocyte differentiation and proliferation that, if absent, introduces significant developmental vulnerabilities.

immunology↗

bioRxiv2022

dialogi: Utilising NLP with chemical and disease similarities to drive the identification of Drug-Induced Liver Injury literature

Drug-Induced Liver Injury (DILI), despite its low occurrence rate, can cause severe side effects or even lead to death. Thus, it is one of the leading causes for terminating the development of new, and restricting the use of already-circulating, drugs. Moreover, its multifactorial nature, combined with a clinical presentation that often mimics other liver diseases, complicate the identification of DILI-related literature, which remains the main medium for sourcing results from the clinical practice and experimental studies. In this work- contributing to the Literature AI for DILI Challenge of the Critical Assessment of Massive Data Analysis (CAMDA) 2021- we present an automated pipeline for distinguishing between DILI-positive and negative papers. We used Natural Language Processing (NLP) to filter out the uninformative parts of a text, and identify and extract mentions of chemicals and diseases. We combined that information with small-molecule and disease embeddings, which are capable of capturing chemical and disease similarities, to improve classification performance. The former are directly sourced from the Chemical Checker (CC). For the latter, we collected data that encode different aspects of disease similarity from the National Library of Medicines (NLM) Medical Subject Headings (MeSH) thesaurus and the Comparative Toxicogenomics Database (CTD). Following a similar procedure as the one used in the CC, vector representations for diseases were learnt and evaluated. Two Neural Network (NN) classifiers were developed: one that only accepts texts as input (baseline model) and an augmented classifier that also utilises chemical and disease embeddings (extended model). We trained, validated, and tested the models through a Nested Cross-Validation (NCV) scheme with 10 outer and 5 inner folds. During this, the baseline and extended models performed virtually identically, with macro F1-scores of 95.04 {+/-} 0.61% and 94.80 {+/-} 0.41%, respectively. Upon validation on an external, withheld, dataset, representing imbalanced data, the extended model achieved an F1-score of 91.14 {+/-} 1.62%, outperforming its baseline counterpart, which got a lower score of 88.30 {+/-} 2.44%. We make further comparisons between the classifiers and discuss future improvements and directions, including utilising chemical and disease embeddings for visualisation and exploratory analysis of the DILI-positive literature.

bioinformatics↗

bioRxiv2022

Intercellular CRISPR Screens Enhance the Discovery of Cancer Immunotherapy Targets

Cancer immunotherapy works through the interplay between immune and cancer cells. Particularly, interactions between cytotoxic T lymphocytes (CTLs) and cancer cells, such as PDCD1 (PD-1) and CD274 (PD-L1), are crucial for removing cancer cells. However, immune checkpoint inhibitors targeting these interactions are effective only to a subset of patients, requiring the development of novel immunotherapy drugs with novel targets. Genome-wide clustered regularly interspaced short palindromic repeats (CRISPR) screening in either cancer or immune cells has been used to discover regulators of immune cell function as immunotherapeutic targets. However, the method has two main limitations. First, performing CRISPR screens in one cell type alone makes it difficult to identify essential intercellular interactions due to the focus on single genes instead of interactions. Second, pooled screening is associated with high noise levels. Therefore, we propose intercellular CRISPR screens, which perform genome-wide CRISPR screening in every interacting cell type to discover intercellular interactions as immunotherapeutic targets. Intercellular CRISPR screens use two individual genome-wide CRISPR screens one each in immune and cancer cells to evaluate intercellular interactions that are crucial for killing cancer cells. We used two publicly available genome-wide CRISPR screening datasets obtained while triple-negative breast cancer (TNBC) cells and CTLs were interacting. We analyzed 4825 interactions between 1391 ligands and receptors on TNBC cells and CTLs to assess the effects of intercellular interactions on CTL function by incorporating both CRISPR datasets and the expression levels of ligands and receptors. Our results showed that intercellular CRISPR screens discovered targets of approved drugs, a few of which were not identifiable using single datasets. To quantitatively evaluate the methods performance, we used data for cytokines and costimulatory molecules because they constitute the majority of immunotherapeutic targets. Combining both CRISPR datasets improved the F1 score of discovering these genes relative to using single CRISPR datasets by more than twice. Our results indicate that intercellular CRISPR screens can identify novel immune-oncology targets that were not obtained using individual CRISPR screens. The pipeline can be extended to other cancer and immune cell types, such as natural killer cells, to identify important intercellular interactions as potential immunotherapeutic targets.

immunology↗

bioRxiv2022

DILIc: An AI-based classifier to search for Drug-Induced Liver Injury literature

Drug-Induced Liver Injury (DILI) is a class of Adverse Drug Reactions (ADR) which causes problems in both clinical and research settings. It is the most frequent cause of acute liver failure in the majority of western countries and is a major cause of attrition of novel drug candidates. Manual trawling of literature for is the main route of deriving information on DILI from research studies. This makes it an inefficient process prone to human error. Therefore, an automatized AI model capable of retrieving DILI-related papers from the huge ocean of literature could be invaluable for the drug discovery community. In this project, we built an artificial intelligence (AI) model combining the power of Natural Language Processing (NLP) and Machine Learning (ML) to address this problem. This model uses NLP to filter out meaningless text (e.g. stopwords) and uses customized functions to extract relevant keywords as singleton, pair, triplet and so on. These keywords are processed by apriori pattern mining algorithm to extract relevant patterns which are used to estimate initial weightings for a ML classifier. Along with pattern importance and frequency, an FDA-approved drug list mentioning DILI adds extra confidence in classification. The combined power of these methods build a DILI classifier (DILIC) with 94.91% cross-validation and 94.14% external validation accuracy. To make DILIC as accessible as possible, including to researchers without coding experience, an R Shiny App capable of classifing single or multiple entries for DILI is developed to enhance ease of user experience and made available at https://researchmind.co.uk/diliclassifier/).

bioinformatics↗

bioRxiv2021

Cryptic diversity within the Poecilochirus carabi mite species complex phoretic on Nicrophorus burying beetles: phylogeny, biogeography, and host specificity

Coevolution is often considered a major driver of speciation, but evidence for this claim is not always found because diversity might be cryptic. When morphological divergence is low, molecular data are needed to uncover diversity. A taxon for which this holds true are the mites, which are known for their extensive and often cryptic diversity. We studied mites of the genus Poecilochirus that are phoretic on burying beetles (Silphidae: Nicrophorus). Poecilochirus taxonomy is poorly understood. Most studies on this genus focus on the evolutionary ecology of Poecilochirus carabi sensu lato, a complex of at least two biological species. Based on molecular data of 230 specimens from 43 locations worldwide, we identified 24 genetic clusters that may represent species. We estimate that these mites began to diversify during the Paleogene, when the clade containing P. subterraneus branched off and the remaining mites diverged into two further clades. One clade resembles P. monospinosus and P. austroasiaticus. The other clade contains 17 genetic clusters resembling P. carabi s.l.. Among these are P. carabi sensu stricto, P. necrophori, and potentially many additional cryptic species. Our analyses suggest that these clades were formed in the miocene by large-scale geographic separation. Diversification also seems to have happened on a smaller scale, potentially due to adaptation to specific hosts or local abiotic conditions, causing some clusters to specialize on certain beetle species. Our results suggest that biodiversity in this genus was generated by multiple interacting forces shaping the tangled webs of life.

zoology↗

bioRxiv2020

A Robust Genome-Wide Association Study Uncovers Signature Genetic Alterations among Pseudomonas aeruginosa Cystic Fibrosis Isolates

Pseudomonas aeruginosa (PA) is an opportunistic pathogen that causes diverse human infections such as chronic airway infection in cystic fibrosis (CF) patients. Although many sequenced genomes are available, a comprehensive comparison between genomes of CF versus non-CF PA isolates remains yet to be conducted. In order to gain a deeper understanding into the PA adaptation in the CF airway, we performed a Genome-Wide Association Study (GWAS) using a total of 1,001 PA genomes. Genetic variations uniquely identified among CF isolates were categorized into (i) alterations in protein-coding regions either large- or small-scale and (ii) polymorphic variations in intergenic regions. We introduced each CF-specific genetic alteration into the genome of PAO1, a prototype PA strain and experimentally validated their outcomes. Loci readily mutated among CF isolates include genes encoding a probable sulphatase and a probable TonB-dependent receptor (PA2332~PA2336), L-cysteine transporter (YecS, PA0313) and a probable transcriptional regulator (PA5438). A promoter region of heme/hemoglobin uptake outer membrane receptor (PhuR, PA4710) was similarly identified as meaningfully different between the CF and non-CF isolate groups. Our analysis, the first of its kind, highlights how PA evolves its genome to persist and survive within the context of chronic CF infection.

microbiology↗