Fig. 1: Longitudinal single-cell workflow and landscape of PBMC gene expression.

A. Schematic representation of the data generation workflow in the study. B. Disease severity definitions for categorizing each sample by clinical parameters at the time of sample collection. A number of samples under each disease severity is indicated. C. Longitudinal disease course for 108 patients (after QC). Dots represent the collection day and disease severity of each of the 286 samples (after QC). Gray lines: the trajectory of individuals with stable disease. Colored lines: individuals with variable severity, colored by the severity of the at-presentation sample. D. Uniform manifold approximation and projection (UMAP) reduced-dimensionality representation of 346,680 high-quality post-QC single cells in gene expression space, annotated with major cluster labels using supervised clustering by the Reference Component Analysis 2 (RCA2) algorithm. E. Same as D, colored by disease severity of each sample. F-H. UMAP plots of T, NK cells (F), B cells (G), and myeloid cells (H), with cells colored by subtype based on unsupervised clustering of single-cell transcriptomes. I. T-cell clonality index (fraction of T cells derived from the 5 most abundant TCR clones; beyond Day 8), estimated using the single-cell immune profiling assay. p-value: Kruskal–Wallis test. Source data are provided as a Source Data file. Box-and-whisker plots show the median (center line), 25th, and 75th percentile (lower and upper boundary), with 1.5x inter-quartile range indicated by whiskers and outliers shown as individual data points.