Biology subjects

Richardson, S.

Publications and source records attributed to Richardson, S..

At least 19 recordsLinked to original sources

bioRxiv2026

Structural Basis for Activation of HRI by the DELE1 C-terminal domain

Haem-Regulated Inhibitor (HRI, aka EIF2AK1) is one of four stress-sensing kinases that phosphorylate eIF2 as part of the integrated stress response (ISR). HRI was first characterized as a haem-sensing kinase where it is inhibited when bound to haem. Recent studies have revealed another role for HRI in sensing mitochondrial stress via the mitochondrial protein DELE1. Upon stress, DELE1 is cleaved, and its C-terminal fragment (DELE1CTD) is released from the mitochondria to the cytoplasm, where it interacts with HRI, to trigger the ISR. This pathway is critical for mitochondrial quality control and neuronal health. Here, we perform biophysical analysis to demonstrate that purified recombinant DELE1CTD binds with high affinity to the N-terminal region of HRI, competing with and displacing haem to activate HRI and pointing to a shared binding site. Using Hydrogen Deuterium Exchange Mass Spectrometry (HDX-MS), we map the binding footprint of DELE1CTD on HRI and confirm that it overlaps with the haem-binding site. These findings support a simple activation mechanism: DELE1CTD activates HRI by excluding inhibitory haem from its binding site.

biochemistry↗

bioRxiv2026

Cell phenotypes in the biomedical literature: a systematic analysis and text mining corpus

The variety of cell phenotypes identified by single-cell technologies is rapidly expanding, yet this knowledge is dispersed across the scientific literature and incompletely represented in structured resources. We present the CellLink corpus, a manually annotated collection of over 22,000 mentions of human and mouse cell populations in recent journal articles, distinguishing specific cell phenotypes, heterogeneous cell populations, and vague cell populations, and linking to Cell Ontology (CL) terms as either exact or related matches, covering nearly half of the terms in the current CL. A systematic analysis reveals lineage-specific patterns in how authors utilize anatomical context, molecular signatures, functional roles, developmental stage, and other attributes in cell naming. We show that fine-tuning transformer-based models on CellLink yields strong performance for named entity recognition, while embedding-based approaches support zero-shot entity linking and distinguishing exact from related matches. We further demonstrate the utility of CellLink to expand and refine the chondrocyte branch of CL.

bioinformatics↗

bioRxiv2025

Post-infection pig and ferret antisera show similar antigenic profiles for human influenza A(H1N1pdm09) viruses

BackgroundMonitoring antigenic drift in human influenza A viruses is essential for vaccine strain selection and ensuring protection against circulating strains. Antigenic drift is traditionally assessed using ferret antisera, which provide monospecific responses, and human vaccinee sera, which reflect exposure to multiple antigens. In this study we evaluated the pig as an alternative source of antisera to study antigenic drift compared to immune responses in ferrets and humans. We included seasonal influenza A(H1N1pdm09) human viruses that had shown different antigenic characteristics when using ferret or human antisera. MethodsPairs of pigs were inoculated with six human A(H1N1)pdm09 viruses circulating between 2019 and 2023, a period of marked antigenic drift. Pig and ferret antisera were analysed by hemagglutination inhibition (HI) and virus neutralization (VN) assays. ResultsPigs were successfully infected with all strains, shedding virus and producing antibody responses, confirming their susceptibility to human influenza A viruses. Antigenic reactivity of pig antisera was qualitatively comparable to ferret antisera in both HI and VN assays, although maximum homologous antibody titres were significantly higher in ferrets. The antisera raised against viruses in circulation in 2019 and before, exempified by A/Guangdong-Maonan/SWL1536/2019, clade 5a.1, were clearly differentiated by both ferret and pig antisera from those in clade 5a.2 and its derivatives that became predominant. ConclusionsFerrets and pigs showed comparable responses and both distinguished clade 5a.1 from clade 5a.2. However, neither model recognised antigenically drifted variants from 2019-2022, including subclades 5a.2-C, 5a.2a-C.1/C.1.9, and .5a.2a.1-C.1.1/D, which were distinguishable using human post-vaccination antisera.

immunology↗

bioRxiv2025

Stochastic gates for covariate selection in population pharmacokinetics modelling

Covariate selection in population pharmacokinetics modelling is essential for understanding interindividual variability in drug response and optimizing dosing. Traditional stepwise covariate modelling is often time-consuming, compared to the new machine learning alternatives. This study investigates the use of Neural Networks with Stochastic Gates for automated covariate selection, aiming to efficiently identify relevant covariates while penalizing excessive covariate inclusion. On various synthetic datasets the approach demonstrated robustness in detecting important covariates, overcoming challenges such as high correlations, low covariate frequencies, high interindividual variability and complex covariate dependencies. In real clinical data from a monalizumab study, the method successfully identified covariates that matched those found by experts. However, for tixagevimab/cilgavimab, it identified a superset of covariates, indicating a potential need for further pruning. This machine learning-based method enhances the covariate pre-selection process in population pharmacokinetics model development, offering significant time savings and improving efficiency even under challenging scenarios. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=89 SRC="FIGDIR/small/656586v1_ufig1.gif" ALT="Figure 1"> View larger version (22K): org.highwire.dtl.DTLVardef@1f5180borg.highwire.dtl.DTLVardef@1fd97e7org.highwire.dtl.DTLVardef@1ffc3d0org.highwire.dtl.DTLVardef@90c207_HPS_FORMAT_FIGEXP M_FIG C_FIG Study HighlightsO_ST_ABSWHAT IS THE CURRENT KNOWLEDGE ON THE TOPIC?C_ST_ABSCovariate selection in population pharmacokinetics modelling is a crucial step in drug development and dosing determination. Traditionally, this process is conducted in a stepwise manner, which can be very time-consuming. Recently, fast machine learning (ML)-based methods for covariate selection have emerged in the literature, offering more efficient alternatives. WHAT QUESTION DID THIS STUDY ADDRESS?Can we use Neural Networks with Stochastic Gates, incorporating explicit penalization on the number of covariates, to select superior set of covariates compared to prior ML-based methods? WHAT DOES THIS STUDY ADD TO OUR KNOWLEDGE?Neural Networks with Stochastic Gates provide a reliable approach in accurately detecting covariates because they can effectively eliminate irrelevant covariates even in cases with high inter-covariates correlation. HOW MIGHT THIS CHANGE DRUG DISCOVERY, DEVELOPMENT, AND/OR THERAPEUTICS?This approach accelerates pharmacokinetic model development by significantly reducing the time required for covariate evaluation. Additionally, it allows for the screening of a wider set of covariates, potentially leading to fewer falsely missed covariates and better-quality models, a task that would be infeasible with traditional stepwise covariate modelling.

pharmacology and toxicology↗

bioRxiv2025

Repurposing Romidepsin for Osteosarcoma: Screening FDA-Approved Oncology Drugs with Three-Dimensional Osteosarcoma Spheroids

Osteosarcoma is the most common primary malignant bone tumor and predominantly affects children, adolescents, and young adults. It is the third most common cause of cancer-related deaths among 9-24-year-olds. Despite aggressive chemotherapeutic and surgical therapies, the survival rate is only 25% for patients with detectable lung metastases at diagnosis and only 70% in patients that present without detectable lung metastases. The poor prognosis is due to growth of metastases irrespective of whether they are initially large enough to detect clinically. It is therefore necessary to develop new methods to target the growth of lung micrometastases. An NCI panel of FDA-approved oncology drugs was therefore screened using three highly metastatic human osteosarcoma cell lines. To more closely approximate in vivo micro-metastases, the screen used a 3D multicellular in vitro osteosarcoma spheroid (sarcosphere) model. Among 13 hits from the initial screen, we identified the histone deacetylase inhibitor (HDI) romidepsin as the most promising inhibitor in secondary screens based on sarcosphere viability. Romidepsin potency was evident with and without standard-of-care chemotherapeutics (MAP: Methotrexate, Adriamycin, Cisplatin) at drug concentrations that are clinically achievable and did not affect non-transformed cells. By those criteria, romidepsin also substantially outperformed the other three FDA-approved HDIs and eight HDIs in clinical trials. Importantly, sarcospheres derived from 30-50% of human and canine patient samples were also sensitive to romidepsin with ED50s 10- to 700-fold less than the Cmax in human patients. Based on these 3-D screening approaches, romidepsin is a promising drug to repurpose for osteosarcoma. Significance StatementOur unbiased sarcosphere-based drug screen identified romidepsin as a promising candidate to repurpose for canine and human patients with metastatic osteosarcoma. This screening strategy allowed us to identify romidepsin-sensitive and -resistant patients. Sarcosphere-based screening may therefore be useful to stratify patients most likely to respond clinically to romidepsin or other drugs.

cancer biology↗

bioRxiv2025

Polymerase mutations underlie adaptation of H5N1 influenza virus to dairy cattle and other mammals.

In early 2024, an unprecedented outbreak of H5N1 high pathogenicity avian influenza was detected in dairy cattle in the USA1. As of mid-2025 the epidemic is ongoing, resulting in spillbacks into poultry, wild birds and other mammals including humans2. Here, we present molecular and virological evidence that the cattle B3.13 genotype H5N1 viruses rapidly accumulated adaptations in polymerase genes that enabled better replication in bovine cells and tissues, as well as cells of other mammalian species including humans and pigs. We find evidence of several mammalian adaptations gained early in the evolution of these viruses in cattle including PB2 M631L, which is found in all cattle sequences, and PA K497R, which is found in the majority. Structurally, PB2 M631L maps to the polymerase-ANP32 interface, an essential host factor for viral genome replication. We show that this mutation adapts the polymerase to better interact with bovine ANP32 proteins, particularly ANP32A, and thereby enhances virus replication in bovine mammary systems and primary human airway cultures. Importantly, we show that ongoing evolution during 2024 and 2025 in the PB2 gene, including E627K and a convergently arising D740N substitution, further increase polymerase activity and virus replication in a range of mammalian cells. Thus, the continued circulation of H5N1 in dairy cattle not only allows virus adaption improving replicative ability in cattle, but also increases the risk of zoonotic spillover.

microbiology↗

bioRxiv2024

The Haemagglutinin Gene of Bovine Origin H5N1 Influenza Viruses Currently Retains an Avian Influenza Virus phenotype.

Clade 2.3.4.4b H5N1 high pathogenicity avian influenza virus (HPAIV) has caused a panzootic affecting all continents except Australia, expanding its host range to several mammalian species. In March 2024, H5N1 HPAIV was first detected in dairy cattle and goats in the United States. Since then, over 230 dairy farms across 14 states have tested positive, with zoonotic infections reported among dairy workers. This raises concerns about the virus undergoing evolutionary changes in cattle that could enhance its zoonotic potential. The Influenza glycoprotein haemagglutinin (HA) facilitates entry into host cells through receptor binding and pH-induced fusion with cellular membranes. Adaptive changes in HA modulate virus-host cell interactions. This study compared the HA genes of cattle and goat H5N1 viruses with the dominant avian-origin clade 2.3.4.4b H5N1 in the United Kingdom, focusing on receptor binding, pH fusion, and thermostability. All the tested H5N1 viruses showed binding exclusively to avian-like receptors, with a pH fusion of 5.9, outside the pH range associated with efficient human airborne transmissibility (pH 5.0 to 5.5). We further investigated the impact of emerging HA substitutions seen in the ongoing cattle outbreaks, but saw little phenotypic difference, with continued exclusive binding to avian-like receptor analogues and pHs of fusion above 5.8. This suggests that the HA genes from the cattle and goat outbreaks do not pose an enhanced threat compared to circulating avian viruses. However, given the rapid evolution of H5 viruses, continuous monitoring and updated risk assessments remain essential to understanding virus zoonotic and pandemic risks.

microbiology↗

bioRxiv2024

Discovery and Characterization of a Pan-betacoronavirus S2-binding antibody

SUMMARY/ABSTRACTThree coronaviruses have spilled over from animal reservoirs into the human population and caused deadly epidemics or pandemics. The continued emergence of coronaviruses highlights the need for pan-coronavirus interventions for effective pandemic preparedness. Here, using LIBRA-seq, we report a panel of 50 coronavirus antibodies isolated from human B cells. Of these antibodies, 54043-5 was shown to bind the S2 subunit of spike proteins from alpha-, beta-, and deltacoronaviruses. A cryo-EM structure of 54043-5 bound to the pre-fusion S2 subunit of the SARS-CoV-2 spike defined an epitope at the apex of S2 that is highly conserved among betacoronaviruses. Although non-neutralizing, 54043-5 induced Fc-dependent antiviral responses, including ADCC and ADCP. In murine SARS-CoV-2 challenge studies, protection against disease was observed after introduction of Leu234Ala, Leu235Ala, and Pro329Gly (LALA-PG) substitutions in the Fc region of 54043-5. Together, these data provide new insights into the protective mechanisms of non-neutralizing antibodies and define a broadly conserved epitope within the S2 subunit.

molecular biology↗

bioRxiv2022

Comparative genetic study of the colony structure and colony spatial distribution between the higher termite Amitermes parvulus and the lower, subterranean termite Reticulitermes flavipes in an urban environment

In insects, ecological competition has often resulted in phenotypic changes and modifications to foraging areas. In termites - and social insects as a whole - colonies cannot easily escape competition through the relocation of their colony. In these species, the outcomes of inter and intra-specific competition are influenced by different life history traits, such as colony size, breeding system (number and types of reproductives), food preference, tunneling patterns, nest site selection, and antagonism between colonies. Here, we investigated variation in breeding system and spatial distribution among colonies of a higher termite Amitermes parvulus and a subterranean termite Reticulitermes flavipes within an urban landscape. We first developed microsatellite markers as a tool to study these life history traits in A. parvulus. Second, we assessed competitive exclusion or tolerance of A. parvulus and R. flavipes colonies by determining their fine-scale distribution using monitoring stations on a grid site, and their large-scale distribution across an urban landscape. Third, we investigated the breeding system of A. parvulus colonies. We showed that the numerous colonies of R. flavipes inhabiting a restricted area contrast with the few, but spatially expansive colonies of A. parvulus, suggesting these species face different degrees of intra-specific competition. We showed that colonies of A. parvulus frequently merged together, and all of them were headed by inbred neotenic reproductives, two characteristics rarely observed in higher termites. Overall, our study revealed drastic differences in colony structure, breeding systems and foraging ranges between the two species. These differences may reflect differences in food preference and food availability between the two species allowing their co-existence within the same urban environment.

evolutionary biology↗

bioRxiv2022

SARS-CoV-2 BA.4 infection triggers more cross-reactive neutralizing antibodies than BA.1

SARS-CoV-2 variants of concern (VOCs) differentially trigger neutralizing antibodies with variable cross-neutralizing capacity. Here we show that unlike SARS-CoV-2 Omicron BA.1, which triggered neutralizing antibodies with limited cross-reactivity, BA.4/5 infection triggers highly cross-reactive neutralizing antibodies. Cross-reactivity was observed both in the absence of prior vaccination and also in breakthrough infections following vaccination. This suggests that next-generation vaccines incorporating BA.4, which is spreading globally, might result in enhanced neutralization breadth.

immunology↗

bioRxiv2022

Fc effector activity and neutralization against SARS-CoV-2 BA.4 is compromised in convalescent sera, regardless of the infecting variant

The SARS-CoV-2 Omicron BA.1 variant, which exhibits high level neutralization resistance, has since evolved into several sub-lineages including BA.4 and BA.5, which have dominated the fifth wave of infection in South Africa. Here we assessed the sensitivity of BA.4 to neutralization and antibody dependent cellular cytotoxicity (ADCC) in convalescent donors infected with four previous variants of SARS-CoV-2, as well as in post-vaccination breakthrough infections (BTIs) caused by Delta or BA.1. We confirm that BA.4 shows high level resistance to neutralization, regardless of the infecting variant. However, breakthrough infections, which trigger potent neutralization, retained activity against BA.4, albeit at reduced titers. Fold reduction of neutralization in BTIs was lower than that seen in unvaccinated convalescent donors, suggesting maturation of neutralizing responses to become more resilient against VOCs in hybrid immunity. BA.4 sensitivity to ADCC was reduced but remained detectable in both convalescent donors and in BTIs. Overall, the high neutralization resistance of BA.4, even to antibodies from BA.1 infections, provides an immunological mechanism for the rapid spread of BA.4 immediately after a BA.1-dominated wave. Furthermore, although ADCC activity against BA.4 was reduced, residual activity may nonetheless contribute to the protection from disease.

immunology↗

bioRxiv2022

Medulloblastoma Group 3 and 4 Tumors Comprise a Clinically and Biologically Significant Expression Continuum Reflecting Human Cerebellar Development

Medulloblastoma is currently sub-classified into distinct DNA methylation subgroups/subtypes with particular clinico-molecular features. Using RNA-seq in large well annotated cohorts of medulloblastoma we show that transcriptionally Group3 and Group4 medulloblastomas exist not as discrete types but as intermediates on a bipolar continuum between archetypal Group3 and Group4 entities. Continuum position is prognostic, reflects propensity for specific DNA copy-number changes, key switches in isoform/enhancer usage and RNA-editing. Examining scRNA-seq profiles we show intra-tumoral transcriptional heterogeneity along the continuum is limited in a subtype-dependent manner. By integrating with a human scRNA-seq reference atlas we show this continuum is mirrored by an equivalent continuum of transcriptional cell types in early fetal cerebellar development. We identify unique developmental niches for all four major subgroups and link each to a common developmental antecedent. Our findings show a transcriptional continuum arising from oncogenic disruption of highly specific fetal cerebellar cell types, linked to almost every aspect of Group3/Group4 molecular biology and clinico-pathology.

cancer biology↗

bioRxiv2022

Gene expression profiling reveals B cells are highly educated by the pancreatic environment during autoimmune diabetes

B cells play an important role in driving the development of type 1 diabetes, however, it remains unclear how they contribute to local beta-cell destruction during disease progression. Using gene expression profiling of B cell subsets in the pancreas and pancreatic lymph nodes, we reveal that B cells are highly modified by the inflamed pancreatic tissue and can be distinguished by their transcriptional profile from those in the lymph node. We identified both a discrete and a core shared gene expression profile in islet CD19+CD138- and CD19+CD138+ B cell subsets, the latter known to have enriched autoreactivity during diabetes development. Upon localisation to pancreatic islets, CD138+ B cells overexpressed genes associated with adhesion molecules and growth factors compared to CD138- B cells. Their shared signature displayed gene expression changes related to the differentiation of antibody-secreting cells and gene regulatory networks associated with interferon signalling pathways, pro-inflammatory cytokines and toll-like receptor activation. Finally, abundant TLR7 expression was detected in islet B cells, and was enhanced specifically in CD138+ B cells. Our study, therefore, provides a detailed transcriptional analysis of islet B cells identifying specific gene signatures and interaction networks that point towards a functional role for B cells in driving autoimmune diabetes.

immunology↗

bioRxiv2020

EPISPOT: an epigenome-driven approach for detecting and interpreting hotspots in molecular QTL studies

We present EPISPOT, a fully joint framework which exploits large panels of epigenetic annotations as variant-level information to enhance molecular quantitative trait locus (QTL) mapping. Thanks to a purpose-built Bayesian inferential algorithm, EPISPOT accommodates functional information for both cis and trans actions, including QTL hotspot effects. It effectively couples simultaneous QTL analysis of thousands of genetic variants and molecular traits, and hypothesis-free selection of biologically interpretable annotations which directly contribute to the QTL effects. This unified, epigenome-aided learning boosts statistical power and sheds light on the regulatory basis of the uncovered hits; EPISPOT therefore marks an essential step towards improving the challenging detection and functional interpretation of trans-acting genetic variants and hotspots. We illustrate the advantages of EPISPOT in simulations emulating real-data conditions and in a monocyte expression QTL study, which confirms known hotspots and finds other signals, as well as plausible mechanisms of action. In particular, by highlighting the role of monocyte DNase-I sensitivity sites from > 150 epigenetic annotations, we clarify the mediation effects and cell-type specificity of major hotspots close to the lysozyme gene. Our approach forgoes the daunting and underpowered task of one-annotation-at-a-time enrichment analyses for prioritising cis and trans QTL hits and is tailored to any transcriptomic, proteomic or metabolomic QTL problem. By enabling principled epigenome-driven QTL mapping transcriptome-wide, EPISPOT helps progress towards a better functional understanding of genetic regulation.

genetics↗

bioRxiv2020

TLR7 ligation augments haematopoiesis in Rps14 (uS11) deficiency via paradoxical suppression of inflammatory signalling and enhanced differentiation

Myelodysplastic syndrome (MDS) is a haematological malignancy characterised by blood cytopenias and predisposition to acute myeloid leukaemia (AML). Therapies for MDS are lacking, particularly those that impact the early stages of disease. We developed a model of MDS using zebrafish using knockout of Rps14, the primary mediator of the anaemia associated with del (5q) MDS. These mutant animals display dose- and age-dependent abnormalities in haematopoiesis, culminating in bone marrow failure with dysplastic features. We utilized rps14 knockdown to undertake an in vivo small molecule screen to identify compounds that ameliorate the MDS phenotype, identifying imiquimod, an agonist of TLR7 and TLR8. Imiquimod alleviates anaemia by promoting haematopoietic stem and progenitor cell expansion and erythroid differentiation, the mechanism of which is dependent on TLR7 ligation. TLR7 activation in this setting paradoxically promoted an anti-inflammatory gene signature suggesting crosstalk between pro-inflammatory pathways endogenous to Rps14 loss and TLR7 pathway activation. Finally, we show that in highly purified human bone marrow samples from anaemic patients, imiquimod leads to an increase in erythroid output from myelo-erythroid progenitors and common myeloid progenitors. Our findings have both specific implications for the development of targeted therapeutics for del (5q) MDS and wider significance identifying a potential role for TLR7 ligation in modifying anaemia.Competing Interest StatementThe authors have declared no competing interest.View Full Text

cancer biology↗

bioRxiv2018

Sparse variable and covariance selection for high-dimensional seemingly unrelated Bayesian regression

MotivationOur work is motivated by the search for metabolite Quantitative Trait Loci (QTL) in a cohort of more than 5,000 people. There are 158 metabolites measured by NMR spectroscopy measured in the 31-year follow-up of the Northern Finland Birth Cohort 1966 (NFBC66). These metabolites, as with many multivariate phenotypes produced by high-throughput biomarker technology, exhibit strong correlation structures. Existing approaches for combining such data with genetic variants for multivariate QTL analysis generally ignore phenotypic correlations or make restrictive assumptions about the associations between phenotypes and genetic loci. ResultsWe present a computationally efficient Bayesian Seemingly Unrelated Regressions (SUR) model for high-dimensional data, with cell-sparse variable selection and sparse graphical structure for covariance selection. Cell-sparsity allows different phenotype responses to be associated with different genetic predictors and the graphical structure is used to represent the conditional dependencies between phenotype variables. To achieve feasible computation of the large model space, we exploit a factorisation of the covariance matrix. Applying the model to the NFBC66 data with 9,000 directly-genotyped Single Nucleotide Polymorphisms, we are able to simultaneously estimate genotype-phenotype associations and the residual dependence structure amongst the metabolites at the same time. Availability and implementationThe R package BayesSUR with full documentation is available at https://cran.r-project.org/web/packages/BayesSUR/ Contactalex.lewin@lshtm.ac.uk

bioinformatics↗

bioRxiv2018

The Delirium and Population Health Informatics Cohort study protocol: ascertaining the determinants and outcomes from delirium in a whole population.

BackgroundDelirium affects 25% of older inpatients and is associated with long-term cognitive impairment and future dementia. However, no population studies have systematically ascertained cognitive function before, cognitive deficits during, and cognitive impairment after delirium. Therefore, there is a need to address the following question: does delirium, and its features (including severity, duration, and presumed aetiologies), predict long-term cognitive impairment, independent of cognitive impairment at baseline?\n\nMethodsThe Delirium and Population Health Informatics Cohort (DELPHIC) study is an observational population-based cohort study based in the London Borough of Camden. It is recruiting 2000 individuals aged [≥]70 years and prospectively following them for two years, including daily ascertainment of all inpatient episodes for delirium. Daily inpatient assessments include the Memorial Delirium Assessment Scale, the Observational Scale for Level of Arousal, and the Hierarchical Assessment of Balance and Mobility. Data on delirium aetiology is also collected. The primary outcome is the change in the modified Telephone Interview for Cognitive Status at two years.\n\nDiscussionDELPHIC is the first population sample to assess older persons before, during and after hospitalisation. The cumulative incidence of delirium in the general population aged [≥]70 will be described. DELPHIC offers the opportunity to quantify the impact of delirium on cognitive and functional outcomes. Overall, DELPHIC will provide a real-time public health observatory whereby information from primary, secondary, intermediate and social care can be integrated to understand how acute illness is linked to health and social care outcomes.

epidemiology↗

bioRxiv2017

Robust expression variability testing reveals heterogeneous T cell responses

Cell-to-cell transcriptional variability in otherwise homogeneous cell populations plays a crucial role in tissue function and development. Single-cell RNA sequencing can characterise this variability in a transcriptome-wide manner. However, technical variation and the confounding between variability and mean expression estimates hinders meaningful comparison of expression variability between cell populations. To address this problem, we introduce a novel analysis approach that extends the BASiCS statistical framework to derive a residual measure of variability that is not confounded by mean expression. Moreover, we introduce a new and robust procedure for quantifying technical noise in experiments where technical spike-in molecules are not available. We illustrate how our method provides biological insight into the dynamics of cell-to-cell expression variability, highlighting a synchronisation of the translational machinery in immune cells upon activation. Additionally, our approach identifies new patterns of variability across CD4+ T cell differentiation.

bioinformatics↗