Supervised dimensionality reduction for exploration of single-cell data by HSS-LDA

Integrative, high-resolution analysis of single cells across experimental conditions with PARAFAC2

Effective tools for exploration and analysis are needed to extract insights from large-scale single-cell measurement data. However, current techniques for handling single-cell studies performed across experimental conditions (e.g., samples, perturbations, or patients) require restrictive assumptions, lack flexibility, or do not adequately deconvolute condition-to-condition variation from cell-to-cell variation. Here, we report that the tensor decomposition method PARAFAC2 (Pf2) enables the dimensionality reduction of single-cell data across conditions. We demonstrate these benefits across two distinct contexts of single-cell RNA-sequencing (scRNA-seq) experiments of peripheral immune cells: pharmacologic drug perturbations and systemic lupus erythematosus (SLE) patient samples. By isolating relevant gene modules across cells and conditions, Pf2 enables straightforward associations of gene variation patterns across specific patients or perturbations while connecting each coordinated change to certain cells without pre-defining cell types. The theoretical grounding of Pf2 suggests a unified framework for many modeling tasks associated with single-cell data. Thus, Pf2 provides an intuitive universal dimensionality reduction approach for multi-sample single-cell studies across diverse biological contexts.

Multi-omic profiling reveals the endogenous and neoplastic responses to immunotherapies in cutaneous T cell lymphoma

Cutaneous T cell lymphomas (CTCLs) are skin cancers with poor survival rates and limited treatments. While immunotherapies have shown some efficacy, the immunological consequences of administering immune-activating agents to CTCL patients have not been systematically characterized. We apply a suite of high-dimensional technologies to investigate the local, cellular, and systemic responses in CTCL patients receiving either mono- or combination anti-PD-1 plus interferon-gamma (IFN-γ) therapy. Neoplastic T cells display no evidence of activation after immunotherapy. IFN-γ induces muted endogenous immunological responses, while anti-PD-1 elicits broader changes, including increased abundance of CLA+CD39+ T cells. We develop an unbiased multi-omic profiling approach enabling discovery of immune modules stratifying patients. We identify an enrichment of activated regulatory CLA+CD39+ T cells in non-responders and activated cytotoxic CLA+CD39+ T cells in leukemic patients. Our results provide insights into the effects of immunotherapy in CTCL patients and a generalizable framework for multi-omic analysis of clinical trials.

Improving reduced-order models through nonlinear decoding of projection-dependent outputs

A fundamental hindrance to building data-driven reduced-order models (ROMs) is the poor topological quality of a low-dimensional data projection. This includes behavior such as overlapping, twisting, or large curvatures or uneven data density that can generate nonuniqueness and steep gradients in quantities of interest (QoIs). Here, we employ an encoder-decoder neural network architecture for dimensionality reduction. We find that nonlinear decoding of projection-dependent QoIs, when embedded in a dimensionality reduction technique, promotes improved low-dimensional representations of complex multiscale and multiphysics datasets. When data projection (encoding) is affected by forcing accurate nonlinear reconstruction of the QoIs (decoding), we minimize nonuniqueness and gradients in representing QoIs on a projection. This in turn leads to enhanced predictive accuracy of a ROM. Our findings are relevant to a variety of disciplines that develop data-driven ROMs of dynamical systems such as reacting flows, plasma physics, atmospheric physics, or computational neuroscience.

Unravelling human hematopoietic progenitor cell diversity through association with intrinsic regulatory factors

Hematopoietic stem and progenitor cell (HSPC) transplantation is an essential therapy for hematological conditions, but finer definitions of human HSPC subsets with associated function could enable better tuning of grafts and more routine, lower-risk application. To deeply phenotype HSPCs, following a screen of 328 antigens, we quantified 41 surface proteins and functional regulators on millions of CD34+ and CD34- cells, spanning four primary human hematopoietic tissues: bone marrow, mobilized peripheral blood, cord blood, and fetal liver. We propose more granular definitions of HSPC subsets and provide new, detailed differentiation trajectories of erythroid and myeloid lineages. These aspects of our revised human hematopoietic model were validated with corresponding epigenetic analysis and in vitro clonal differentiation assays. Overall, we demonstrate the utility of using molecular regulators as surrogates for cellular identity and functional potential, providing a framework for description, prospective isolation, and cross-tissue comparison of HSPCs in humans.

High-dimensional profiling of pediatric immune responses to solid organ transplantation

Solid organ transplant remains a life-saving therapy for children with end-stage heart, lung, liver, or kidney disease; however, ∼33% of allograft recipients experience acute rejection within the first year after transplant. Our ability to detect early rejection is hampered by an incomplete understanding of the immune changes associated with allograft health, particularly in the pediatric population. We performed detailed, multilineage, single-cell analysis of the peripheral blood immune composition in pediatric solid organ transplant recipients, with high-dimensional mass cytometry. Supervised and unsupervised analysis methods to study cell-type proportions indicate that the allograft type strongly influences the post-transplant immune profile. Further, when organ-specific differences are considered, graft health is associated with changes in the proportion of distinct T cell subpopulations. Together, these data form the basis for mechanistic studies into the pathobiology of rejection and allow for the development of new immunosuppressive agents with greater specificity.

Prior anti-CTLA-4 therapy impacts molecular characteristics associated with anti-PD-1 response in advanced melanoma

Immune checkpoint inhibitors (ICIs), including CTLA-4- and PD-1-blocking antibodies, can have profound effects on tumor immune cell infiltration that have not been consistent in biopsy series reported to date. Here, we analyze seven molecular datasets of samples from patients with advanced melanoma (N = 514) treated with ICI agents to investigate clinical, genomic, and transcriptomic features of anti-PD-1 response in cutaneous melanoma. We find that prior anti-CTLA-4 therapy is associated with differences in genomic, individual gene, and gene signatures in anti-PD-1 responders. Anti-CTLA-4-experienced melanoma tumors that respond to PD-1 blockade exhibit increased tumor mutational burden, inflammatory signatures, and altered cell cycle processes compared with anti-CTLA-4-naive tumors or anti-CTLA-4-experienced, anti-PD-1-nonresponsive melanoma tumors. We report a harmonized, aggregate resource and suggest that prior CTLA-4 blockade therapy is associated with marked differences in the tumor microenvironment that impact the predictive features of PD-1 blockade therapy response.

Magnitude and kinetics of the human immune cell response associated with severe dengue progression by single-cell proteomics

Approximately 5 million dengue virus-infected patients progress to a potentially life-threatening severe dengue (SD) infection annually. To identify the immune features and temporal dynamics underlying SD progression, we performed deep immune profiling by mass cytometry of PBMCs collected longitudinally from SD progressors (SDp) and uncomplicated dengue (D) patients. While D is characterized by early activation of innate immune responses, in SDp there is rapid expansion and activation of IgG-secreting plasma cells and memory and regulatory T cells. Concurrently, SDp, particularly children, demonstrate increased proinflammatory NK cells, inadequate expansion of CD16⁺ monocytes, and high expression of the FcγR CD64 on myeloid cells, yet a signature of diminished antigen presentation. Syndrome-specific determinants include suppressed dendritic cell abundance in shock/hemorrhage versus enriched plasma cell expansion in organ impairment. This study reveals uncoordinated immune responses in SDp and provides insights into SD pathogenesis in humans with potential implications for prediction and treatment.

Data science through the lens of systems immunology

Glass, a post-doctoral researcher, and Amouzgar, a PhD student, in Bendall lab proposed a supervised dimensionality reduction method to explore and analyze single-cell data. Their Patterns paper highlights the advantages of supervised learning in single-cell datasets with class labels. They talk about the essential role of data science in this project and in their lives.

A supervised take on dimensionality reduction via hybrid subset selection

Amouzgar et al. present HSS-LDA, a supervised dimensionality reduction approach for single-cell data that outperforms existing unsupervised techniques. They couple hybrid subset selection to linear discriminant analysis and identify interpretable linear combinations of predictors that best separate predefined biological groups.