Journal: Nature Communications

Article Title: An amphioxus neurula stage cell atlas supports a complex scenario for the emergence of vertebrate head mesoderm

doi: 10.1038/s41467-024-48774-4

Figure Lengend Snippet: a drawings of Mediterranean amphioxus developmental stages from the egg to the larva (with one open gill slit) stage. The developmental time (hours post fertilization, hpf) is given for embryos raised at 19 °C, and we highlight the neurula stage presented in this study (21 hpf). b Two-dimensional projection of cell clusters (metacells) using a force-directed layout based on the co-clustering graphs for individual cells (see Methods ). Metacells are color-coded by cell type. c Normalized fold change expression of top variable genes (rows) per metacell (columns, grouped by cell type). For each metacell, we selected up to 30 markers with a minimum fold change ≥ 2. Selected gene names from known markers, used to annotate each cell type, are indicated to the right of the heatmap. Genes in bold case are shown in panel ( d ). d Pie charts depicting the fraction of cells mapped to each cell type among the cell transcriptomes and the cell counting experiment (top); and the 3D reconstruction with assignment of nuclei to each germ layer (bottom). A transverse section is shown on the left, and dorsal views with anterior to the top on the right (full, without epidermis nuclei, without epidermis and neural cells nuclei. e Expression profile of previously unknown marker genes for specific cell types (neural, endoderm, anterior epidermis, presumptive cerebral vesicle, notochord, and tailbud) analyzed by in situ hybridization (ISH, top, with anterior to the left and dorsal to the top in side views, n = 10 embryos) and corresponding two-dimensional expression maps (bottom, based on the same layout as panel b ). Gene expression is shown as density maps representing UMI counts (per 10,000 UMIs) in each cell.

Article Snippet: Two-dimensional projection of the metacells were created using a force-directed layout based on the metacell co-clustering graph ( mcell_mc2d_force_knn function).

Techniques: Expressing, Cell Counting, Marker, In Situ Hybridization, Gene Expression

Journal: Nature Communications

Article Title: An amphioxus neurula stage cell atlas supports a complex scenario for the emergence of vertebrate head mesoderm

doi: 10.1038/s41467-024-48774-4

Figure Lengend Snippet: a 2D projection of endodermal metacells on a side view scheme of an amphioxus neurula stage embryo with anterior to the left and dorsal to the top. b Gene expression as normalized fold changes of selected gene markers and corresponding ISH ( n = 10 embryos). Gene expression is shown as density maps representing UMI counts (per 10,000 UMIs) in each cell, with cells positioned in the vicinity of their corresponding metacells. c Multiple in situ hybridization for Dmbx , Fgfrl and Brachyury2 ( n = 5 embryos). Lateral views are shown on the left, with anterior to the left and dorsal to the top. Optical cross-sections at the level of the dotted line are also shown. The labeling intensity is stronger on one side due to the thickness of the embryo that was imaged on one side after mounting. The three genes are coexpressed in the anteriormost region of the dorsal mesendoderm (metacell 1). Normalized fold change expression of each marker is shown on the right. Scale bar: 25 µm.

Article Snippet: Two-dimensional projection of the metacells were created using a force-directed layout based on the metacell co-clustering graph ( mcell_mc2d_force_knn function).

Techniques: Gene Expression, In Situ Hybridization, Labeling, Expressing, Marker

Journal: Nature Communications

Article Title: An amphioxus neurula stage cell atlas supports a complex scenario for the emergence of vertebrate head mesoderm

doi: 10.1038/s41467-024-48774-4

Figure Lengend Snippet: a 2D projection of somitic metacells on a side view scheme of an amphioxus neurula stage embryo, with anterior to the left and dorsal to the top. b Gene expression as normalized fold changes of selected gene markers and corresponding ISH ( n = 10 embryos). c Multiple in situ hybridization for Ripply+Alx+Gata1/2/3 and Ripply+Tcf21/Msc+Alx ( n = 5 embryos). Lateral views are shown on the left, with anterior to the left and dorsal to the top. Optical cross-sections at the level of the dotted line are also shown. The labeling intensity is stronger on one side due to the thickness of the embryo. Ripply , Alx and Gata1/2/3 are coexpressed in the non-muscular part of formed somites (metacells 2 and 11) whereas Ripply , Tcf21/Msc and Alx are coexpressed in the first somite pair (metacells 1 and 3). Scale bar: 25 µm.

Article Snippet: Two-dimensional projection of the metacells were created using a force-directed layout based on the metacell co-clustering graph ( mcell_mc2d_force_knn function).

Techniques: Gene Expression, In Situ Hybridization, Labeling

Journal: Cell

Article Title: Stepwise emergence of the neuronal gene expression program in early animal evolution

doi: 10.1016/j.cell.2023.08.027

Figure Lengend Snippet: A multi-species placozoan whole-body cell atlas (A) Consensus phylogenetic tree obtained with Bayesian inference under the CAT + GTR + Г4 mixture model on the Metazoa-only 209-markers concatenated aminoacid matrix recoded into 4 categories (SR4). Bayesian posterior probabilities are indicated as supports in key nodes. The cladogram to the right depicts the phylogenetic relationships among placozoans, highlighting the four species here studied. (B) Summary of the statistical support for alternative phylogenetic positions of Placozoa in the different datasets analyzed: (1) only metazoans (63 species) versus metazoans and choanoflagellates as outgroup (81 species); (2) high-information markers (filtered for tree-likeness score with MARE −d 2 parameter) markers filtered for compositional homogeneity (denoted as CH; markers failing the compositional heterogeneity based on simulated alignments using the LG + Γ4 model in p4, at p > 0.01); and (3) original aminoacid multiple sequence alignments versus recoded alignments with three different schemes (SR4, SR6, and Dayhoff6). (C) 2D projection of metacells for each species sampled in this study and pie charts indicating the relative proportion of cells in each broad cell type category, based on a force-directed layout of the metacell co-clustering graph (see ). Right, a broad cell type clustering tree of all four species obtained using the UPGMA average algorithm on Log-Det distance matrices, based on binary ortholog activity in each cell type (fold change ≥ 2). (D) Normalized expression of top variable genes (rows, fold change ≥ 2 with a maximum of 15 genes per metacell) across metacells (columns). Broad cell types are color-coded in the x axis and red squares highlight the peptidergic progenitor metacells. (E) Fluorescent HCR-ISH of Trichoplax sp. H2 specimens showing the expression of an upper epithelia-like marker (calpain-9, top) and the expression of a marker gene for the unknown cell type (β-secretase, bottom). Images correspond to the maximum projection of 183 and 70 optical sections, respectively. The dotted lines indicate the sections used for the extended orthogonal views (45 slices). Arrowheads in the orthogonal views indicate the upper part of the animals. Insets in the bottom image show the detail of cells localized in the rim of the animal. Cells highlighted in the insets were imaged at higher magnification in the portions indicated with a square. Expression of the marker genes is shown to the left of each panel. Scale bars are 50 μm for the general views and 5 μm for the insets. See also Figures S1 , , and .

Article Snippet: Two-dimensional projections of the metacells were produced using a force-directed layout based on the metacell co-clustering graph ( mcell_mc2d_force_knn function in Metacell ), using K = 20 nearest neighbors and a maximum confusion degree = 5.

Techniques: Sequencing, Activity Assay, Expressing, Marker

Journal: Cell

Article Title: Stepwise emergence of the neuronal gene expression program in early animal evolution

doi: 10.1016/j.cell.2023.08.027

Figure Lengend Snippet: scRNA-seq summary statistics, related to Figure 1 (A) Distribution of total RNA molecules per cell in each placozoan sampled. (B) Clicktag (CT) sample demultiplexing statistics for an example experiment mixing T. adhaerens H1 and C. collaboinventa H23. Top left: distribution of relative sizes (in UMI/cell) of each cell when their transcriptome is mapped to each of the multiplexed species ( T. adhaerens H1 and C. collaboinventa H23). Top right: UMIs/cell of each cell, classified according to whether its UMI counts are higher in one species or the other, intermediate (doublets), or non-cells (empty droplets). Middle left: fraction of normalized CT counts associated with the most common pair of CT barcodes for each cell, classifying cells in two categories: (1) determined cells, where the first and second most abundant CT barcodes are concordant (from the same sample in the experimental design) but the first and third ones are discordant (from different samples), which represent bona fide cells from a single species; or (2) whether the first and second most abundant CTs are discordant (from different samples), which represent possible doublets. Middle right: distribution of CT counts/cell, classified according to whether its CT counts are concordant for one species or the other (determined cells in the left histogram), intra-species doublets (the discordant first and second barcodes come from different samples of the same species), inter-species doublets (the discordant first and second barcodes come from samples of different species), or unclassified (low CT counts). Bottom left: single-cell uniform manifold approximation and projection (UMAP) projection based on normalized CT counts. We removed cells belonging to Louvain clusters with a high fraction of cells classified as doublets in either the cross-species UMI- or CT-based doublet detection procedures (clusters highlighted in blue). Bottom right, heatmap showing the normalized CT counts per single cell (each sample was labeled with two different barcodes, e.g., BC53 + BC54). (C) Summary of the doublet calls for the five CT datasets. Notice the consistency between cross-species UMI- and CT-based doublet calls (which in addition allow us to identify intra-species doublets). (D) Metacell confusion matrices that represent metacell pairwise similarities derived from the K-nn graph connectivity between all cells in each pair of metacells. Colors indicate the broad cell type classification of metacells. (E) Cell type sample composition. (F) Metacell summary statistics. Barplots indicate the number of cells per metacell. Boxplots indicate the number of transcripts/UMIs per single cell grouped into metacells. Colors indicate the broad cell type classification of metacells.

Article Snippet: Two-dimensional projections of the metacells were produced using a force-directed layout based on the metacell co-clustering graph ( mcell_mc2d_force_knn function in Metacell ), using K = 20 nearest neighbors and a maximum confusion degree = 5.

Techniques: Labeling, Derivative Assay

Journal: Cell

Article Title: Stepwise emergence of the neuronal gene expression program in early animal evolution

doi: 10.1016/j.cell.2023.08.027

Figure Lengend Snippet: Cell type comparisons across Placozoa, related to Figures 1 and (A) Schematic representation of the main steps in the ICC algorithm applied to metacells. (B) Distribution of ICC-derived expression conservation (EC) scores for each pair of species and for paralog versus ortholog gene pairs. (C) Heatmaps indicating the EC-weighted Pearson correlation between cell types across placozoans. (D) Same as (C) but showing SAMap scores. (E) Force-directed network of cell type similarity across species, using the weighted Fruchterman-Reingold algorithm. Nodes represent cell types (larger nodes correspond to placozoans, smaller ones correspond to other species), and edges represent pairwise similarities as weighted Pearson correlation coefficients. For each cell type, only the top edges are shown (standardized quantile scores above 0.99). Placozoan nodes are color-coded by cell type. Other metazoan nodes are custom color-coded based on similarity to placozoan cell types. (F) Heatmaps representing the transcriptomic similarity between pairs of cell types of the four placozoan species (rows) compared to seven species from other lineages (columns; including three cnidarians, two bilaterians, one sponge, and one ctenophore). Heatmap color reflects the Pearson correlation score between the expression of genes in each cell type (weighting each gene pair with their expression conservation score in that pair of species, using the ICC procedure).

Article Snippet: Two-dimensional projections of the metacells were produced using a force-directed layout based on the metacell co-clustering graph ( mcell_mc2d_force_knn function in Metacell ), using K = 20 nearest neighbors and a maximum confusion degree = 5.

Techniques: Derivative Assay, Expressing

Journal: Cell

Article Title: Stepwise emergence of the neuronal gene expression program in early animal evolution

doi: 10.1016/j.cell.2023.08.027

Figure Lengend Snippet: Stepwise evolutionary emergence of the neuronal gene expression program (A) Network summarizing pairwise similarities (weighted Pearson correlation) between neurons from cnidarians and bilaterians (middle) with placozoan cell types (top and bottom). Only similarities above 0.2 are shown. All pairwise cell type similarities across phyla are shown in Figure S3 E. (B) Left, ancestral state reconstruction of neuronal gene expression programs across Metazoa. Pie charts indicate presence, gains and losses at each extant or ancestral node. Ancestral nodes are inferred using Dollo parsimony. Neuronal genes in each species are selected from single-cell atlases as having a FC ≥ 2 in at least 25% of the metacells annotated as neurons/neuron-like cells. Right, number of GPCRs and ion channels expressed in neuronal/neuronal-like metacells (threshold FC ≥ 2) versus non-neuronal metacells. (C) Gene ontology enrichments of gene gains in ancestral gene expression programs, based on annotations of the mouse orthologs. (D) Schematic representation of the major functional gains in the neuronal gene expression programs in early animal evolution. See also Figure S3 .

Article Snippet: Two-dimensional projections of the metacells were produced using a force-directed layout based on the metacell co-clustering graph ( mcell_mc2d_force_knn function in Metacell ), using K = 20 nearest neighbors and a maximum confusion degree = 5.

Techniques: Gene Expression, Functional Assay

Journal: Cell

Article Title: Stepwise emergence of the neuronal gene expression program in early animal evolution

doi: 10.1016/j.cell.2023.08.027

Figure Lengend Snippet: Characterization of intermediate metacells and gene modules, related to Figures 2 and (A) Barplots representing the number of cells classified in each intermediate category (gray) compared with the number of cells doublets in each category (green) that would be expected given the relative frequency of the terminal cell types in each case. We used two-tailed exact binomial tests to determine whether the observed number of intermediate cells significantly differed from the expectation (p values next to each set of bars). (B) Top, barplots representing the number of genes shared in intermediate metacells between the placozoan species where each cell type is found (gray) compared with the number of genes shared by the respective terminal cell types (green). We used one-tailed exact binomial tests to determine whether the number of genes shared across species was higher for terminal than for intermediate cell types (p values shown for each cell type). Notice that in most cases the difference is small and non-significant, indicating that the genes expressed in intermediate cells are conserved across species and not a stochastic sampling of genes expressed in the respective terminal cell types. Bottom, Venn diagrams detailing the number of shared genes across species for lipophil-1/gland intermediate cells (gray) compared with the shared genes by lipophil-1 and gland cell types (green). (C) Intermediate cells exhibit intermediate transcriptional signatures between their terminal cell types. For each pair of cell type in each species, we show the sum of the fraction of UMIs (per 1,000 UMIs) of the top markers (FC ≥ 2). Panels are arranged to indicate the detection of specific intermediate cell types (rows) in each of the species (columns). (D) Flow cytometry scatterplots of Trichoplax sp. H2 cells labeled by HCR-ISH against markers specific for lipophil (fatty acid-binding protein 4, Alexa Fluor-647), gland (chymotrypsin, Alexa Fluor-546) and fiber (angiotensin I-converting enzyme, Alexa Fluor-488) cells. Selected areas in each panel denote the percentage of cells with single or double label, which would correspond to intermediate cells. (E) Heatmaps representing the eigengenes across metacells of gene modules calculated using WGCNA in each placozoan. x axis colors indicate the broad cell type classification of metacells. Module colors (y axis) are arbitrary. (F) Left, gene-gene expression correlation matrix, grouping genes into the same modules as in (E). Right, normalized expression across metacells of genes grouped into modules. Transcription factors are highlighted with a dot to the right of the heatmap. Notice the presence of “lateral” gene modules expressed in individual metacells across cell types. These include, for example, the cell cycle and ciliary apparatus modules. (G) Top 10 gene ontology terms enriched in each multi-species gene module. x axis colors indicate the cell type where each module is most active (manual curation). (H) Fold-change expression of selected genes with immune-related functions across cell types of all four placozoans.

Article Snippet: Two-dimensional projections of the metacells were produced using a force-directed layout based on the metacell co-clustering graph ( mcell_mc2d_force_knn function in Metacell ), using K = 20 nearest neighbors and a maximum confusion degree = 5.

Techniques: Two Tailed Test, One-tailed Test, Sampling, Flow Cytometry, Labeling, Binding Assay, Gene Expression, Expressing

Journal: Cell

Article Title: Stepwise emergence of the neuronal gene expression program in early animal evolution

doi: 10.1016/j.cell.2023.08.027

Figure Lengend Snippet: Placozoan gene expression programs (A) Multi-species clustering of gene modules across placozoans. Each node represents a gene module (group of genes co-expressed across metacells; see Figures S4 E–S4G), and each node is color-coded according to the species. Edges link modules sharing orthologs across species, and their width reflects the Jaccard index of ortholog overlap between modules (only edges with Jaccard ≥0.125 are shown). We curated 34 multi-species modules, the majority of which are composed of modules from four species (pie plot). Most modules are specific to individual cell types (bar plot), with the exception of cross-cell type modules that include genes related to pan-peptidergic cells, cell cycle (S-phase and G2-phase), meiosis, and the ciliary apparatus. (B) Gene ontology enrichments in selected gene modules (left), and expression of transcription factor (TF) regulators and associated enriched motifs (right). (C) Left, multi-species clustering of non-peptidergic (top) and peptidergic cell types (bottom). The cell type tree has been obtained as in Figure 1 C. Gray boxes list selected TFs specific to various cell type clades. Right, heatmap depicting the fraction of orthologous genes from each gene module expressed across cell types. Modules have been color-coded according to their cell type specificity, with the cross-cell type modules highlighted with asterisks. (D) Number of TFs, GPCRs, and neuropeptides (NPs) expressed (fold change ≥ 2) in each cell type. See also Figures S4 , , and .

Article Snippet: Two-dimensional projections of the metacells were produced using a force-directed layout based on the metacell co-clustering graph ( mcell_mc2d_force_knn function in Metacell ), using K = 20 nearest neighbors and a maximum confusion degree = 5.

Techniques: Gene Expression, Expressing

Journal: Cell

Article Title: Stepwise emergence of the neuronal gene expression program in early animal evolution

doi: 10.1016/j.cell.2023.08.027

Figure Lengend Snippet: Molecular signatures of neurogenesis in placozoans peptidergic cell progenitors (A) 2D projection of metacells of a Trichoplax sp. H2 single-cell pooled transcriptome of individuals grown under four conditions: treatment with the Notch antagonists DAPT (3,453 cells) and LY411575 (4,666 cells), the Notch signaling agonist Yhhu3792 (5,114 cells), and an untreated control (4,765 cells). Metacells have been color-coded by broad cell type based on comparison to the reference Trichoplax sp. H2 dataset ( Figure 1 C). (B) Normalized expression of Delta, Notch, Hes, and Hey in the 2D projection of Trichoplax sp. H2 metacells. (C) Pie plot with cell type proportions among the control cells (top) and fold-change enrichment in cell type fractions for each drug treatment (bottom). p values from a two-sided Fisher’s exact test of cell type counts relative to the control. (D) Differential expression of the Hes, Hey, and Myc TFs in lower epithelial and gland cells (two cell types with broad Notch expression), measured using the difference in UMIs/10 4 between treatment and control. p values indicate significant differential expression based on a two-sided Fisher’s exact test on UMI counts. (E) Expression of selected marker genes related to peptidergic progenitor specification across all four placozoans, including markers used for HCR-ISH experiments. p values from an FDR-adjusted two-sided Fisher’s exact test of UMI counts in a given cell type, relative to the control. (F) Sox TF maximum likelihood phylogenetic analysis supporting the orthology of placozoan Sox1/2/3 and Sox4/11/12. (G) Left, fluorescent HCR-ISH of C. collaboinventa H23 showing the expression of the peptidergic progenitor-specific marker HoiH23_PlH23_008135 (NN peptide, red) in animals with (Gii) and without (Gi) treatment with 10 μM LY411575 for 24 h. Images are maximum projections of 50 (Gi) and 40 (Gii) optical sections. The dotted lines indicate the sections used for the extended orthogonal views (40 slices). Arrowheads in the orthogonal projections indicate the upper part of the animals. Middle, fluorescent HCR-ISH of C. collaboinventa H23 (Giii and Gvi) showing the expression of the peptidergic progenitor-specific markers HoiH23_PlH23_008135 (NN peptide, red) and Klf13 (yellow). Image (Giii) is a maximum projection of 22 optical sections. Images (Giv) to (Gvi) highlight the detail of three individual cells expressing both markers and correspond to the squared sections of image (Giii). Right, fluorescent HCR-ISH of Trichoplax sp. H2 (Gvii and Gviii) showing the expression of the peptidergic progenitor-specific markers Klf13 (red) and Delta receptor (yellow). Image (Gvii) is a maximum projection of 16 optical sections. Inset (Gviii) highlights the detail of a cell expressing both markers. Dotted line depicts the shape of the cell as delineated by the membrane marker (green). Scale bars correspond to 100 μm in i and ii, 10 μm in (Giii) and (Gviii), and 1 μm for (Giv)–(Gvi) and (Gviii). (H) Expression of selected TFs, RNA-binding proteins and chromatin factors specific to placozoan peptidergic progenitors (E) along the neural developmental trajectories described in scRNA-seq experiments in M. musculus (gastrula to pharyngula stage ), N. vectensis (gastrula to adult ), and Hydra vulgaris (regenerating adult ). Genes with expression FC ≥ 1.25 in any cell type of a given developmental trajectory are indicated as colored squares in each (overexpressed genes with FC > 1 and < 1.25 in stages intermediate between two other stages are indicated with a white asterisk). For each developmental trajectory, we also indicate the number of orthologous TFs and RBPs shared with each placozoan species (barplots to the right). See also Figure S7 .

Article Snippet: Two-dimensional projections of the metacells were produced using a force-directed layout based on the metacell co-clustering graph ( mcell_mc2d_force_knn function in Metacell ), using K = 20 nearest neighbors and a maximum confusion degree = 5.

Techniques: Control, Comparison, Expressing, Quantitative Proteomics, Marker, Membrane, RNA Binding Assay

Journal: Genome Biology

Article Title: Metacell-2: a divide-and-conquer metacell algorithm for scalable scRNA-seq analysis

doi: 10.1186/s13059-022-02667-1

Figure Lengend Snippet: Robustness of the divide-and-conquer metacell algorithm. A Distribution of metacell normalized inner variance for the PBMC dataset, using the Baran et al. algorithm (orange) vs. MC2 two-sided stability score optimization, working on the entire data in a single pile (i.e., no divide and conquer, green). B Distribution of normalized inner variance for the PMBC dataset using the full MC2 algorithms (blue) vs. the single-pile algorithm (green). C Metacell graph derived by MC2 on the PMBC dataset. Annotation as in Baran et al. D Distribution of metacell normalized inner variance for HSC and MPP cells, when using full MC2 on the HCA bone marrow data set (orange) or when restricting analysis to MPP/HSC cells alone (blue). E Metacell graph for the full HCA BM data set and for metacells computed on the zoomed-in HSC/MPP subset

Article Snippet: We first wished to ensure that the MC2 algorithm’s efficient graph partition algorithm is not losing significant quality compared to the original, resampling-based Metacell implementation (MC1) [ ].

Techniques: Derivative Assay

Journal: Genome Biology

Article Title: Metacell-2: a divide-and-conquer metacell algorithm for scalable scRNA-seq analysis

doi: 10.1186/s13059-022-02667-1

Figure Lengend Snippet: Scaling MC2 to millions of cells. A Graphs show scaling of MC2 (multi-pile) compared to a naïve metacell on a single pile or a PCA+2-Phase Louvain clustering implementation in Seurat, using the PMBC 160K cell data (resampled to datasets of increasing sizes— X -axis). B Comparison of MC2 and two-phase clustering performance for the organogenesis datasets (MOCA). C Effects of scaling the pile sizes on the normalized inner variance for MC2 on the organogenesis data. D Distribution of normalized inner variance for MC2 and PCA+Louvain original sub-clusters on the organogenesis data. E Marker heat map and metacell graph projection of the organogenesis data. Clustering of metacells is used for coloring and cross-reference purpose, in support of, but not in place of supervised annotation. F Distribution of metacells linkage with different embryonic time points over the metacell graph. Color coding is based on metacell clustering as in D . To compensate for differences in the number of cells, we randomly sampled 2000 points for each time point and weighted by the fraction of the cells of each age in each metacells

Article Snippet: We first wished to ensure that the MC2 algorithm’s efficient graph partition algorithm is not losing significant quality compared to the original, resampling-based Metacell implementation (MC1) [ ].

Techniques: Comparison, Marker

Journal: Genome Biology

Article Title: Metacell-2: a divide-and-conquer metacell algorithm for scalable scRNA-seq analysis

doi: 10.1186/s13059-022-02667-1

Figure Lengend Snippet: MC2-sensitive detection of rare behavior. A Correlation matrix between log gene expression frequency of 260 genes with rare expression signatures (see text). We highlight several gene clusters at the left. B Each bar graph show specificity (left) and fold change enrichment (right) of top genes separating three exemplified rare transcriptional behaviors. Also shown for each rare behavior are the distribution of total gene expression of rare genes per single cell within the top-enriched metacell(s) (shades of green) and within the top enriched PCA+Louvain subcluster (orange). C Shown are single-cell gene expression for rare behavior marker genes and for genes correlated and anticorrelated with them, plotted for cells within the most strongly enriched PCA+Louvain sub-cluster for the observed behavior

Article Snippet: We first wished to ensure that the MC2 algorithm’s efficient graph partition algorithm is not losing significant quality compared to the original, resampling-based Metacell implementation (MC1) [ ].

Techniques: Gene Expression, Expressing, Marker