Multiplexed retrograde labeling and monosynaptic tracing using barcoded rabies virus.

(A) In multiplexed retrograde labeling, rabies viruses carrying different barcodes are injected into different brain regions, and retrogradely labeled neurons can be distinguished based on the barcodes they carry. (B) In multiplexed monosynaptic tracing, source cells labeled by helper AAV viruses expressing TVA and rabies G can be infected by barcoded rabies virus, which then passes the barcode to presynaptic neurons. Both rabies barcodes and endogenous mRNAs can be read out to infer cell/type connectivity. However, if multiple source cells share the same barcode, they may obscure single-cell connectivity mapping and must be filtered out.

High-quality cell typing of rabies virus-barcoded neurons using single-cell RNA-seq.

(A) Illustration of the design of barcoded rabies virus libraries. (B) Barcode distribution in the CCS and non-CCS libraries. (C) Outline of the experiments. (D) Cluster mapping confidence of rabies-labeled cells from this study and AAV-labeled cells from (Graybuck et al., 2021). (E) Select marker gene expression in uninfected cells from (Tasic et al., 2018) and rabies-infected cells of matching cell types. (F) UMAP plot of gene expression of cells infected with rabies virus (black) overlaid on non-infected cells from (Tasic et al., 2018). The uninfected cells are color-coded by cluster identities and cluster names are indicated. (G) The expression of rabies-encoded genes in the sequenced cells. The count for each rabies-encoded gene was scaled by the sum of all reads that mapped to viral constructs multiplied by 10,000. Natural logarithms of resulting scaled counts, ln(scaled counts+1), were used to represent expression levels of virally encoded genes. The bar on top indicates donor animals. (H) Example barcode sequences in three sequenced cells. Letter height indicates probability at each position. Gray boxes indicate the barcode region. (I) Distribution of cell types of retrogradely labeled cells. Colors indicate cell types and match those in (F), and disc sizes indicate number of cells.

Number of cells in the scRNA-seq-based transsynaptic tracing and retrograde tracing experiments.

In situ sequencing resolves cell types of neurons infected with barcoded rabies virus.

(A) Illustration of probe designs and amplification approach for in situ sequencing on rabies barcodes. (B) Illustration of the experiments. (C) Left, example image of a coronal section (outlined by a dashed line) during the hybridization cycle. The final sequencing cycles for genes and barcodes in the boxed area are shown on the Right. Genes and nucleotides corresponding to each color are indicated. Scale bars = 50 µm. (D)(E) All sequenced cells shown on a UMAP plot (D) or on a representative coronal section (E). Colors indicate subclass-level cluster labels as shown in the legend. (F) Cluster matching between BARseq subclass-level clusters and subclasses from scRNA-seq (Tasic et al., 2018). (G) The number of cells with the indicated number of barcodes per cell. (H) The count of the primary barcode and the second most abundant barcode in each barcoded cell. Cells above the dotted line have more than one barcode per cell. (I) Counts of the most dominant barcode and the remaining barcodes in each cell. (J) Summary of the number of cells with more than one barcodes (blue) and/or in which the primary barcodes accounted for less than half of all barcode reads (red). Shapes are not drawn to scale. (K) UMAP plot as plotted in (D), color coded by whether the cells had barcodes. (L)(M) The distribution of endogenous mRNA reads per cell and unique gene counts per cell in all cells (L) or the QC-filtered cells (M). In (L), the dotted vertical lines indicate QC thresholds. BC: Barcoded cells.

Multiplexed retrograde labeling recapitulates known cortical projections.

(A) The minimum Hamming distance between each barcode and all other barcodes for barcodes in the VISal library (blue), the SC/RSP library (red), and random barcodes (dashed). (B) The distribution of mismatch between each sequenced barcode and the closest barcode in the two libraries. Colors indicate whether and which library each barcode was mapped to. (C) The frequencies of barcodes in the libraries (x-axis) are plotted against the number of sequenced cells carrying those barcodes (y-axis). Jitter is added on the y-axis to resolve overlapping dots. (D) Two representative slices, one from each brain showing all sequenced cells on each slice. Barcoded cells are color-coded by projections and non-barcoded cells are color-coded by brain regions. (E)(F) The number of neurons from each cortical area and each subclass that project to either VISal (E) or SC/RSP (F). Dot sizes and colors indicate number of cells. (G) The distribution of the number of projecting cells for each subclass. (H)-(J) Total number of cells (H) or the fraction of cells with projections to SC/RSP (I) or VISal (J) of the indicated L4/5 IT type in each cortical area. (K)(L) Fractions of variance in the probability of projections explained by combinations of the compositional profiles of cell types (S for subclass and T for type) and cortical areas (Area) for L4/5 IT neurons (K) or for all excitatory neurons (L). In (L), the first column indicates variance explained by compositional profiles of cell types at the subclass level. *p =6×10-70, **p =6×10-58, ***p =7×10-78, ****p =6×10- 37 comparing the means to 100 iterations of shuffled controls using two-tailed t-tests.

Multiplexed transsynaptic labeling by sequencing rabies barcodes in situ.

(A) Five possible types of barcode-sharing networks in a barcoded transsynaptic tracing experiment using rabies virus. Whether each network is compatible with monosynaptic tracing and/or mapping synaptic convergence is indicated below. *see text for considerations regarding connected-source networks. (B) Outline of transsynaptic labeling experiment using barcoded rabies virus and in situ sequencing. (C) Example image of a representative coronal section during sequencing. Images of the first gene sequencing cycle, the hybridization cycle, and the first barcode sequencing cycle of the boxed area are shown on the right. Scale bars = 50 µm. (D) First nine barcodes sequencing images of example bisected neuron from two adjacent sections. Scale bars = 10 µm. (E) The fraction of manually curated barcoded cells with the indicated barcode (“BC”) complexity and barcode read counts per cell. Dashed lines indicate quality control thresholds for barcodes. (F) The distribution of endogenous mRNA reads per cell and unique gene counts per cell in non-barcoded cells (gray), barcoded cells (blue), and source cells (red). Dashed lines indicate quality control thresholds for gene expression. (G)(H) UMAP plots of all barcoded cells color-coded by the cluster label at the subclass level (G) or whether the cell has barcodes or is a source cell (H). (I) Locations and cell types of two source cells (red cross) and presynaptic cells (dots) that shared the same barcodes. Colors of dots indicate transcriptomic types of presynaptic neurons. Transcriptomic types of source cells are indicated below each plot. All other cells from the coronal sections that the source cells were on were plotted in gray. (J) Estimated numbers of barcode/source cell combinations that belonged to each of the three networks with source cells. (K) The posterior probability of the number of independent infection events to generate the same number of barcodes found in the experiment. (L) The distribution of the number of cells that shared each barcode that was not found in a source cell. (M) The distribution of the maximum barcode frequency in the virus library that ensures single infection for 95% of barcodes across 10,000 simulations. (N) The ratios between the observed number of converging outputs and the expected number from random connectivity between cortical subclasses of neurons. Colors correspond to log10 of ratios, and the ratios are indicated in the plot. Only values with false positive rate (FPR) < 0.05 are shown (see Methods). As seen from the blue squares associated with L6 IT cells, these neurons were less likely to synapse on to the same post-synaptic neurons with other neuronal types (in particular L6 CT neurons, with only 50% of converging connections compared to those expected from random connectivity).

Quality control of scRNA-seq in rabies virus-infected cells.

(A) Quality control plots of scRNA-seq for all animals used in both multiplexed retrograde labeling experiments and barcoded transsynaptic labeling experiments. (B) Cluster mapping confidence for rabies barcoded neurons. Boxes indicate quartiles and medians, and whiskers indicate range of data excluding outliers. Outliers are shown as individual dots.

The expression of immune response-related genes in rabies virus-infected cells

(A) Volcano plots showing differential expression of genes in clusters with more than 10 rabies infected cells across all six animals. (B) The expression of select immune response related genes in uninfected cells from (Tasic et al., 2018), AAV-infected cells from (Graybuck et al., 2021), and rabies infected cells of matching cell types from this study.

The expression of activity-related genes in rabies virus-infected cells and uninfected cells from (Tasic et al., 2018)

scRNA-seq is insufficient to resolve connectivity among transsynaptically labeled neurons using barcoded rabies virus.

(A) Outline of the experiment. (B) The expression of rabies encoded genes (top) and AAV-encoded genes (bottom) in sequenced cells. The count for each gene was scaled by the sum of all viral reads multiplied by 10,000. Natural logarithms of resulting scaled counts, ln(scaled counts+1), were used to represent expression levels of virally encoded genes. The bar on the top indicates whether a cells is considered to have AAV or not. The second bar on top indicates animals. (C) The number of source cells (left) and presynaptic cells (right) of each cell type (rows) from each animal (columns). Colors indicate cluster identity and dot size indicates number of cells. (D) The minimum Hamming distance between each barcode and all other barcodes in the pool of sequenced barcodes (solid line) or random barcodes (dashed line). (E) The number of barcodes that were found in the indicated number of animals. (F) Box plots showing the library frequency of barcodes found in one or four animals. All barcodes are indicated by dots. The boxes show median and quartiles and the whiskers indicate range. BC: Barcode.

Barcoded transsynaptic labeling resolved by in situ sequencing

(A) Locations of source cells (red dots) relative to cortical layers, which are color coded as indicated. (B) Fraction of presynaptic cell barcodes that were also found in source cells (y axis) when thresholding barcodes in source cells at the indicated reads per cell (x axis).

Presynaptic cells and source cells in all single-source networks

In each plot, a source cell (red cross) and presynaptic cells (dots) that shared the same barcodes were plotted. Colors of dots indicate transcriptomic types of presynaptic neurons. Transcriptomic types of source cells are indicated below each plot. All other cells from the coronal sections that the source cells were on were plotted in gray.

List of filters and lasers used for in situ sequencing