Methods for selectively targeting neuronal pathways

Abstract

Disclosed herein, are methods for selectively labeling a neuron, a single projection neuron, or neuronal pathways.

Description

[0001] METHODS FOR SELECTIVELY TARGETING NEURONAL PATHWAYS

[0002] CROSS REFERENCE TO RELATED APPLICATIONS

[0003] This application claims the benefit of the filing date of U.S. Provisional Application No. 63 / 678,394, filed on August 1, 2024, and, U.S. Provisional Application No. 63 / 827,594, filed on June 20, 2025. The content of this earlier filed applications is hereby incorporated by reference in its entirety.

[0004] STATEMENT REGARDING FEDERALLY FUNDED RESEARCH

[0005] This invention was made with government support under grant numbers DA056668, NS132161. and NS132027 awarded by the National Institutes of Health. The government has certain rights in the invention.

[0006] INCORPORATION OF THE SEQUENCE LISTING

[0007] The present application contains a sequence listing that is submitted concurrent with the filing of this application, containing the file name '‘36406_0042Pl_SL’’ which is 61,440 bytes in size, created on July 31, 2025, and is herein incorporated by reference in its entirety pursuant to 37 C.F.R. § 1.52(e)(5).

[0008] BACKGROUND

[0009] While bulk-level neuronal projection data is readily available in the mouse, singleneuron projection data is still challenging to obtain, as it requires high-resolution brain- wide imagining and reconstructions. MAPseq (Multiplexed Analysis of Proj ections by Sequencing) was developed instead of imaging, to rapidly map the projections of thousands of single neurons. In MAPseq, Sindbis viruses (SINV) encoding unique RNA sequences (barcodes) deliver each barcode to each neuron and the barcodes are transported to axon terminals. Then, by reading the barcodes via sequencing in target areas across the brain, projection patterns of individual neurons are identified which were hindered in the bulk-level projection. MAPseq, however, is limited because it cannot discriminate molecular cell types among infected neurons and the MAPseq virus (SINV-MAPseq) is limited to infecting specific brain areas or animal species. Thus, new methods of obtaining single-neuron projection data is needed.

[0010] SUMMARY

[0011] Disclosed herein are methods for selectively labeling a neuron, the methods comprising: a) infecting a neuron or population of neurons with a virus encoding an exogenous gene, wherein the exogenous gene encodes a receptor of an alphavirus; and b) infecting the same neuron or population of neurons in a) with a chimeric virus in an amount to express a target protein, wherein the chimeric virus comprises: i) a RNA nucleic acid encoding a nucleic acid barcode, and ii) a structure protein capable of specifically binding to the alphavirus receptor encoded by the exogenous gene of a); thereby selectively labeling a neuron.

[0012] Disclosed herein are methods for selectively labeling a neuron, the methods comprising: infecting a neuron or population of neurons in a subject with a chimeric virus in an amount to express a target protein, wherein the chimeric virus comprises: i) a RNA nucleic acid encoding a nucleic acid barcode, and ii) a structure protein capable of specifically binding to an alphavirus receptor; wherein the neuron or population of neurons in the subject expresses a receptor of an alphavirus that is capable of specifically binding to the structure protein of ii): thereby selectively labeling a neuron.

[0013] Disclosed herein are methods for labeling a neuronal pathway in an area or species which has not previously been infected by Sindbis virus, the method comprising: infecting a population of neurons with chimeric viruses, wherein the chimeric alphaviruses comprise: i) an self-amplifying but propagation-incompetent RNA nucleic acid encoding a barcode nucleic acid, and ii) structure proteins of alphaviruses with varying infectivity for the neuronal pathway, thereby enhancing the labeling of a neuronal pathway across areas and species.

[0014] BRIEF DESCRIPTION OF THE DRAWINGS

[0015] FIGS. 1A-B show an overview of MAPseq (adapted from Kebschull et al., 2016, Neuron). FIG. 1A shows the principle of MAPseq. FIG. IB shows the single-neuron projections using MAPseq.

[0016] FIGS. 2A-D show the development of chimeric MAPseq viruses. FIG. 2A shows the structure of the Alphavirus. FIG. 2B shows the procedure of making SINV-MAPseq. FIG. 2C shows the members of the Alphavirus family with their associated receptors. FIG. 2D shows the procedure of making chimeric MAPseq viruses.

[0017] FIGS. 3A-D show cell-type specific infection and acquired infectivity by chimeric MAPseq viruses. FIG. 3A shows non-specifically infected neurons by SINV-MAPseq in the mouse cortex. FIG. 3B shows cell-ty pe specifically infected neurons by CHIKV -MAPseq after specific expression of Mxra8 by AAV in the mouse cortex. FIG. 3C shows neurons in the mouse cerebellar nuclei that were not infected by SINV-MAPseq injection. FIG. 3D show infected neurons in the mouse cerebellar nuclei by EEEV -MAPseq.

[0018] FIGS. 4A-L show the validation of cell-type specific infection of CHIKV -MAPseq in the primary motor cortex (Mop). FIG. 4A shows the layer structure of the Mop and molecular markers of L2 / 3 and L5. FIG. 4B shows a schematic for the SINV-MAPseq experiment. FIG. 4C shows neurons non-specifically infected by SINV-MAPseq. FIG. 4D show identified projection patterns from a SINV-MAPseq infected mouse covering IT, ET, and CT. The scale bar indicates normalized projection strength. FIG. 4E show L2 / 3 specific, Cux2-Cre and L5 specific, and Rbp4-Cre lines. FIG. 4F show a schematic for the CHIKV-MAPseq experiments. FIGS. 4G, K show neurons specifically infected by CHIKV-MAPseq in L2 / 3 or L5, respectively in Cux2-Cre and Rbp4-Cre. FIGS. 4H, I show identified projection patterns from a CHIKV-MAPseq infected Cux2-Cre mouse or Rbp4-Cre mouse, covering solely IT or IT and ET, respectively.

[0019] FIGS. 5A-B show a method of labeling a neuron or a population of neurons and creating a sequencing library. FIG. 5 A shows an overview of MAPseq after virus infection. Labeling of each neuron with a unique RNA barcode (green cells were the infected cells and barcodes were represented as numbers); dissection of each target brain areas (represented as A, B. C, and D); extracted viral RNA from each target area: MAPseq sequencing which was made from the extracted viral RNA; and analysis of the sequencing data resulting in singleneuron projection matrix and cell-type clustering by the projection patterns (each dot represent each cell and different color indicates different clusters). FIG. 5B shows the detailed process of making a MAPseq sequencing library from the extracted viral RNA. Black line; viral RNA. gray line; non-viral extra RNA, blue arrow; RT primers, blue line; RT product, navy arrow; PCR1 primer, navy line; PCR1 product, Purple arrow; PCR2 primer, Purple line; and PCR2 product.

[0020] FIG. 6 shows the DH-BB (5 ’SIN; SP-CHIKV) plasmid.

[0021] FIG. 7 shows the DH-BB (5’SIN; SP-CHIKV) sequence (SEQ ID NO: 6).

[0022] FIG. 8 shows the DH-BB (5’SIN; SP-EEEV) plasmid. FIG. 9 shows the DH-BB (5 'SIN; SP-EEEV) sequence (SEQ ID NO: 7).

[0023] FIG. 10 shows the DH-BB (5 ’SIN; SP-SFV) plasmid.

[0024] FIG. 11 shows the DH-BB (5’SIN; SP-SFV) sequence (SEQ ID NO: 8).

[0025] FIG. 12 shows the DH-BB (5’SIN; SP-TR339) plasmid.

[0026] FIG. 13 shows the DH-BB (5’SIN; SP-TR339) sequence (SEQ ID NO: 9).

[0027] FIG. 14 shows the DH-BB (5’SIN; SP-VEEV) plasmid.

[0028] FIG. 15 shows the DH-BB (5’SIN; SP-VEEV) sequence (SEQ ID NO: 10).

[0029] FIG. 16 shows the DH-BB (5’SIN; SP-WEEV) plasmid.

[0030] FIG. 17 shows the DH-BB (5’SIN; SP-WEEV) sequence (SEQ ID NO: 11).

[0031] FIG. 18 shows the pAAV-CAG-tdT-P2A-Mxra8 plasmid.

[0032] FIG. 19 shows the pAAV-CAG-tdT-P2A-Mxra8 sequence (SEQ ID NO: 12).

[0033] FIGS. 20A-P show the development of an example of a method described herein (also referred to as “POINTseq”). FIG. 20A shows diverse projection patterns of individual neurons across target regions (left). Although retrograde tracing can be used to label specific neurons of interest, it cannot distinguish between different connectomic cell types when their projections overlap at the injection site in terms of different target sites (left) and different projection strength (right). FIG. 20B shows that MAPseq efficiently maps single-cell projections but lacks specificity for neurons of interest. FIG. 20C shows the principle of POINTseq relies on the exogenous expression of a specific alphavirus receptor that is not endogenously expressed in the mouse brain. Infection is restricted to receptor-expressing cells through receptor-virus interactions, enabling cell type-specific delivery' of barcodes. FIG. 20D shows the expression of Mxra8 in the mouse brain compared to other know n alphavirus receptors (left) and according to cell types (right). Receptors of Sindbis virus (Slclla2), Venezuelan equine encephalitis virus (LdlradS), Semliki Forest virus (Vldlr and Lrp8) were included. Counts per million mapped reads (CPM) values were adapted from Allen Brain Cell Atlas scRNAseq datasets for the entire mouse brain. The plots display a central line, box, and whiskers representing the median, interquartile range (IQR), and data points within ±1.5x IQR. The level of Mxra8 in the cells is close to zero and significantly lower than that of other well-known neuronal genes (one-way ANOVA with Bonferrom post- hoc test **** / ? < 0.0001). FIG. 20E shows the generation of POINTseq virus by pseudotyping genomic RNA of Sindbis virus used in MAPseq (MAPseq gRNA) with structural proteins of CHIKV. The helper RNA encoding structural proteins of CHIKV w as designed to be propagation-incompetent. FIG. 20F shows three different viral injection conditions in MOp (without AAV, with AAV expressing tdT, with AAV expressing Mxra8 with tdT) to confirm the Mxra8-dependent infection of POINTseq virus. FIGS. 20G-20I show that POINTseq virus infections (GFP-positive) were rare without Mxra8 regardless of AAV (Mxra8-). but highly increased with Mxra8 (Mxra8+). FIG. 20J shows that the number of infections was significantly higher with Mxra8 indicating that approximately 98.7% of POINTseq virus infections are dependent on exogenous Mxra8 (Mxra8-: 6 mice, Mxra8+: 4 mice, Student’s t test, **p < 0.01). FIG. 20K shows high-resolution images from FIG. 20G demonstrated that POINTseq virus infected cells shows typical morphology of cortical neurons. FIGS. 20L and 20M show schematics of experimental designs for Cux2-Cre, Rbp4-Cre, and AAVretro-Cre (left), with descriptions of the targeted neurons (right). FIGS. 20N-20P show specific infections of POINTseq virus in Mxra8-positive cells targeted by the three Cre experiments.

[0034] FIGS. 21A-B show validation of specific mapping of single-neuron projections by POINTseq. FIG. 21 A shows the projection matrix of combined datasets of MAPseq and POINTseq in MOp with hierarchical clustering that identified 6 connectomic types (SINV: 11,153 neurons, Cux2-Cre: 2,544 neurons, Rbp4-Cre: 5,027 neurons, AAVretro-Cre: 2,829 neurons, 2 mice in each). Abbreviations for the targets are provided in Table 2. The color bar represents normalized barcode abundance, which was first normalized by the total barcode abundance across the targets within each neuron, log-transformed, and then normalized by the maximum value of the matrix bringing the scale as from 0 to 1. FIG. 2 IB shows separated projection matrices of MAPseq and POINTseq datasets from FIG. 22A (left) and proportions of expected projection types in each dataset (right).

[0035] FIGS. 22A-H show POINTseq in DAT-Cre mice provides a comprehensive classification of connectomic types of VTA and SNc DA neurons. FIG. 22A shows a schematic of the experimental design for applying POINTseq to map DA neurons in VTA and SNc. involving DAT-Cre mice, Nissl staining, and laser capture microdissection. FIG. 22B shows confirmation of the specific infection of POINTseq virus in DAT+ cells in the VTA and SNc. FIG. 22C shows representative images of target regions from Nissl stained coronal sections for laser capture microdissection along the rostral-caudal axis. The regions were determined based on Allen Brain reference atlas and previous literature (Beier, K.T., et al. (2015) Cell 762. 622-634; Poulin, J.-F., et al. (2018) Nat Neurosci 27, 1260-1271; Hintiryan, H., et al. (2016) Nat. Neurosci. 19, 1100-1 114; and Hunnicutt, B.J., et al. (2016) Elife 5). FIG. 22D shows projection matrix reconstructed from POINTseq data (3,813 neurons from 3 mice; 3,563 VTA neurons and 250 SNc neurons; 519, 362, and 2,932 neurons per mouse). Hierarchical clustering identified 34 connectomic types, each characterized by distinct primary projection targets. Target region abbreviations are listed in Table 2. The color bar represents normalized barcode abundance, which was first normalized by the total barcode abundance across all targets within each neuron, log-transformed, and then normalized by the maximum barcode abundance value of the matrix. FIG. 22E shows the proportion of the number of targets per neuron. FIGS. 22F and 22G show the characterization of each connectomic type by primary projection targets (FIG. 22F) and a dot plot showing mean projection patterns for each type (FIG. 22G). FIG. 22H shows the percentage of VTA and SNc neurons in each connectomic type.

[0036] FIGS. 23A-H show that VTA DA neurons exhibit actively structured co-innervation patterns and projection motifs across target regions with known functions. FIG. 23 A show7reduced projection matrix of 3,452 VTA DA neurons across 9 functionally characterized target regions, each receiving input from more than 10 neurons. Barcode abundance was renormalized by the total barcode count across the reduced set of targets within each neuron, log-transformed, and then normalized to the maximum barcode value in the matrix. FIG. 23B shows the conditional probability7of projecting to region B given projections to region A. FIGS. 23C-23F show co-innervation analysis. FIG. 23C show schematic examples of Dice and rarity- scores and how they influence over- and under-representation relative to a random binomial model. Observed Dice scores (FIG. 23D), rarity scores (FIG. 23E), and deviation of observed Dice scores from those estimated by the random binomial model (binomial test with Bonferroni correction, *p < 0.05, **p < 0.01, **p < 0.001, ****p < 0.0001) (FIG. 23F). FIGS. 23G and23H show motif analysis. Volcano plot of projection motifs with more than 5 observed or estimated neurons (FIG. 23G), and counts of significantly over- or under- represented motifs (binomial test with Bonferroni correction, *p < 0.05, **p < 0.01, **p < 0.001, *«* / ? < 0.0001) (FIG. 23H).

[0037] FIGS. 24A-J show that SNc neurons have broader, more independent projections than VTA neurons in GP and CP. FIG. 24A show GP / CP focused projection matrix of 735 VTA neurons and 239 SNc neurons that have more than 10% of their projections in GP and CP. Barcode abundance was re-normalized by the total barcode count across the reduced set of targets within each neuron, log-transformed, and then normalized to the maximum barcode value in the matrix. FIGS. 24B-24D show the number of targets per neuron. Total proportion of neurons according to number of targets (FIG. 24B), distribution of the number of targets from the three mice (tw o-w ay ANOVA, source x target number, Bonferroni post-hoc test, **p < 0.01. *** / ? < 0.001) (FIG. 24C). distribution of the number of targets according to the main target region (Student’s t test with Bonferroni adjustment, **p < 0.01, **** ? < 0.0001) (FIG. 24D). FIGS. 24E-24J show co-innervation analysis of VTA and SNc neurons across the 10 subregions of GP and CP. Observed Dice scores of VTA and SNc neurons (FIGS. 24E, H), rarity scores of VTA and SNc neurons (FIGS. 24F and 241), and deviation of observed Dice scores from those estimated by the random binomial model in VTA and SNc (binomial test with Bonferroni correction, *p < 0.05, **p < 0.01, **p < 0.001, ****p < 0.0001) (FIGS. 24G and24J).

[0038] FIGS. 25A-N show Mxra8 is not endogenously expressed throughout the mouse brain, and POINTseq virus infects neurons in the Mxra8-dependent manner across multiple brain regions. FIG. 25 A shows expression of Mxra8 across regions of the mouse brain. Counts per million mapped reads (CPM) values were adapted from Allen Brain Cell Atlas scRNAseq datasets for the entire mouse brain. The plots display a central line, box, and whiskers representing the median, interquartile range (IQR), and data points within ±1.5x IQR. CTXsp: cortical subplate, HPF: hippocampal formation, STR: striatum, TH: Thalamus, HY: hypothalamus, PAL: pallidum, MB: midbrain, OLF: olfactory areas, CB: cerebellum, P: pons. MY: medulla. FIG. 25B shows Mxra8-dependent POINTseq virus infection was assessed across brain regions including striatum (STR), dentate gyrus (DG), thalamus (TH), and amygdala (AMY). FIGS. 25C-25F show MAPseq virus efficiently infects these regions. FIGS. 25G-25J show that without exogenous Mxra8 expression, POINTseq virus rarely infects cells in these regions. FIGS. 25K-25N show that with exogenous Mxra8 expression, POINTseq virus infects many cells in these regions. POINTseq virus was injected 18 days after the AAV injection.

[0039] FIGS. 26A-B show confirmation of the propagation-incompetency of the POINTseq virus. FIGS. 26A and 26B show the POINTseq virus (FIG. 26A) does not produce plaques, similarly to the MAPseq virus (FIG. 26B), confirming its propagation incompetency.

[0040] FIGS. 27A-B show that POINTseq virus specifically infects dopamine neurons in the VTA and SNc of DAT-Cre mice. FIG. 27A shows that the POINTseq virus specifically infects tdT+ (DAT ) neurons spanning the VTA and SNc along the rostral-caudal axis. FIG. 27B shows that POINTseq virus infects DA neurons which are TH+ and tdT+ (DAT+). Green arrowheads indicate example cells of the infected cells which are GFP+, TH+, and tdT+.

[0041] FIG. 28 shows the Projection matrices for VTA and SNc DA neurons, separately plotted from the combined matrix shown in FIG. 22D. FIG. 29 shows the comparison of POINTseq-defmed connectomic cell types with previously reported VTA and SNc DA neuron populations, and investigation of corresponding types to the populations.

[0042] FIGS. 30A-C show the co-innervation analysis of FIG. 23 performed separately for individual mice. FIGS. 30A-30C show conditional probability7, Dice score, and over- or under-representation analyses for individual DAT-Cre mice: mouse 1 (FIG. 30A), mouse 2 (FIG. 30B), and mouse 3 (FIG. 30C). The three show patterns consistent with the tendencies observed in FIG. 23.

[0043] FIG. 31 show the quantification of projection strengths of individual neurons that constitute the significantly over- or under-represented motifs shown in FIG. 23H.

[0044] FIGS. 32A-C show that SNc neurons display broader and evener projection strength than VTA neurons. FIG. 32A shows projection densities (projection strength normalized by region size) were calculated for each neuron. FIG. 32B shows that SNc neurons exhibited lower standard deviations of projection densities than VTA neurons, indicating more even projections across GP and CP subregions. FIG. 32C shows projection densities around each neuron's main target gradually declined in SNc neurons, particularly in the caudal direction, while VTA neurons showed sharp decreases, especially when targeting CPr subregions (two- way ANOVA, source x projection density, Bonferroni post-hoc test, *p < 0.05, **p < 0.01, *** ? < 0.001, **** ? < 0.0001).

[0045] FIG. 33 shows neurons projecting to at least two targets show similar projection strength patterns as in FIG. 32. Differences in projection density distribution between SNc and VTA neurons remained consistent when restricting the analysis to neurons projecting to at least two regions (two-way ANOVA, source x projection density, Bonferroni post-hoc test, *p < 0.05, ** ? < 0.01, ***p < 0.001. **** / ? < 0.0001).

[0046] FIGS. 34A-I show that among neurons exclusively projecting to the GP and CP, SNc neurons still exhibit broader projection patterns, while rarity scores become comparable between SNc and VTA neurons. FIG. 34A show- projection matrices of neurons exclusively projecting to the GP and CP. filtered from FIG. 24A (276 VTA neurons, 185 SNc neurons). FIGS. 34B-34C, 34F-34G show7VTA neurons exhibited fewer projection targets and lower co-innervation abundance than SNc neurons, although the differences were reduced compared to the full dataset. FIGS. 34D, and 34H show rarity scores of VTA neuron coinnervations were generally reduced under this filter, except for CPt-involving combinations. FIGS. 34E and 341 show that most co-innervations of VTA and SNc neurons became similarly explained by the random binomial model (binomial test with Bonferroni correction, *p < 0.05, ** ? < 0.01, ** ? < 0.001, **** ? < 0.0001).

[0047] FIGS. 35A-H show that the number of targets and co-innervation analysis of FIG. 24, performed with subsampled neurons of VTA matching the number of SNc neurons. FIGS. 35A-35H show the results of number of targets (FIGS. 35A and 35E), Dice score (FIGS. 35B and 35F), rarity score (FIGS. 35C and 35 G), and over- or under-representation (FIGS. 35D and 35H) remain similar to those of FIG. 24.

[0048] FIGS. 36A-C show that the number of targets and results of the co-innervation analysis performed separately for individual mice using the same methods as in FIG. 24.

[0049] FIGS. 37A-B show co-innervation analysis using Dice scores at maximal regional resolution. FIG. 37A show dice scores of VTA DA neurons across 41 ipsilateral and 13 contralateral target regions. FIG. 37B shows dice scores of SNc DA neurons across 32 ipsilateral target regions.

[0050] FIGS. 38A-B show over- or under-represented co-innervations at maximal regional resolution. FIG. 38A shows over- or under-represented co-innervations of VTA neurons across 41 ipsilateral and 13 contralateral regions. FIG. 38B shows over- or under-represented co-innervations of SNc neurons across 32 ipsilateral regions.

[0051] FIGS. 39A-D show the determination of source parameters. FIG. 39A shows the proportion of number of targets per neurons according to varying source UMI thresholds. MAPseq-1 and MAPseq-2 mice were adapted from a previous published data. The number of targets stabilize beyond a specific range. FIGS. 39B-D show the parameters related to source determination between VTA and SNc. FIG. 39B show the distribution of source ratios (lower UMI / higher UMI between VTA and SNc). FIG. 39C shows the accuracy of source prediction across source ratio bins, estimated by simulations from a model consisting of true source and false source. FIG. 39D shows the accuracy of source prediction in the simulations filtered by varying source thresholds. Bin width is 0.05 and the red line indicates the chosen threshold (0.1) for source ratio.

[0052] FIGS. 40A-B show the Diversity of POINTseq virus. FIG. 40A shows the rank plot of the POINTseq viral barcode library by UMI counts, consisting of 2.2 x 106unique barcodes. FIG. 40B show the estimated fraction of uniquely labeled neurons based on the filtered barcode distribution, excluding barcodes with fewer than 3 UMI counts. DETAILED DESCRIPTION

[0053] The present disclosure can be understood more readily by reference to the following detailed description of the invention, the figures and the examples included herein.

[0054] Before the present methods and compositions are disclosed and described, it is to be understood that they are not limited to specific synthetic methods unless otherwise specified, or to particular reagents unless otherwise specified, as such may, of course, vary. It is also to be understood that the terminology used herein is for the purpose of describing particular aspects only and is not intended to be limiting. Although any methods and materials similar or equivalent to those described herein can be used in the practice or testing of the present invention, example methods and materials are now described.

[0055] Moreover, it is to be understood that unless otherwise expressly stated, it is in no way intended that any method set forth herein be construed as requiring that its steps be performed in a specific order. Accordingly, where a method claim does not actually recite an order to be followed by its steps or it is not otherwise specifically stated in the claims or descriptions that the steps are to be limited to a specific order, it is in no w ay intended that an order be inferred, in any respect. This holds for any possible non-express basis for interpretation, including matters of logic with respect to arrangement of steps or operational flow, plain meaning derived from grammatical organization or punctuation, and the number or type of aspects described in the specification.

[0056] All publications mentioned herein are incorporated herein by reference to disclose and describe the methods and / or materials in connection with which the publications are cited. The publications discussed herein are provided solely for their disclosure pnor to the filing date of the present application. Nothing herein is to be construed as an admission that the present invention is not entitled to antedate such publication by virtue of prior invention. Further, the dates of publication provided herein can be different from the actual publication dates, which can require independent confirmation.

[0057] DEFINITIONS

[0058] As used in the specification and the appended claims, the singular forms “a,” “an” and ‘"the” include plural referents unless the context clearly dictates otherwise.

[0059] The word “or” as used herein means any one member of a particular list and also includes any combination of members of that list.

[0060] The phrase “and / or,” as used herein in the specification and in the claims, should be understood to mean "either or both" of the elements so conjoined, i.e., elements that are conjunctively present in some cases and disjunctively present in other cases. Other elements may optionally be present other than the elements specifically identified by the "and / or" clause, whether related or unrelated to those elements specifically identified unless clearly indicated to the contrary. Thus, as a non-limiting example, a reference to “A and / or B,” when used in conjunction with open-ended language such as "comprising" can refer, in one embodiment, to A without B (optionally including elements other than B); in another embodiment, to B without A (optionally including elements other than A); in yet another embodiment, to both A and B (optionally including other elements); etc.

[0061] Ranges can be expressed herein as from "about" or “approximately” one particular value, and / or to “about” or “approximately” another particular value. When such a range is expressed, a further aspect includes from the one particular value and / or to the other particular value. Similarly, when values are expressed as approximations, by use of the antecedent “about,” or “approximately,” it will be understood that the particular value forms a further aspect. It will be further understood that the endpoints of each of the ranges are significant both in relation to the other endpoint and independently of the other endpoint. It is also understood that there are a number of values disclosed herein and that each value is also herein disclosed as “about” that particular value in addition to the value itself. For example, if the value “10” is disclosed, then “about 10” is also disclosed. It is also understood that each unit between two particular units is also disclosed. For example, if 10 and 15 are disclosed, then 11, 12, 13, and 14 are also disclosed.

[0062] The term “nucleic acid” is used herein to refer to a polymer of deoxyribonucleotides and / or ribonucleotides. “Nucleic acid” shall mean any nucleic acid, including, without limitation, DNA, RNA and hybrids thereof. The nucleic acid bases that form nucleic acid molecules can be the bases A, C, G, T and U, as well as derivatives thereof. “Genomic nucleic acid” refers to DNA derived from a genome, which can be extracted from, for example, a cell, a tissue, a tumor or blood.

[0063] As used herein, the term “amplifying” refers to the process of synthesizing nucleic acid molecules that are complementary to one or both strands of a template nucleic acid. Amplifying a nucleic acid molecule typically includes denaturing the template nucleic acid, annealing primers to the template nucleic acid at a temperature that is below the melting temperatures of the primers, and enzymatically elongating from the primers to generate an amplification product. The denaturing, annealing and elongating steps each can be performed once. Generally, however, the denaturing, annealing and elongating steps are performed multiple times (e.g., polymerase chain reaction (PCR)) such that the amount of amplification product is increasing, often times exponentially, although exponential amplification is not required by the present methods. Amplification typically requires the presence of deoxyribonucleoside triphosphates, a DNA polymerase enzyme and an appropriate buffer and / or co-factors for optimal activity of the polymerase enzyme. The term “amplified nucleic acid molecule’' refers to the nucleic acid molecules, which are produced from the amplifying process.

[0064] As used herein, the term “read” or “sequence read” refers to the nucleotide or base sequence information of a nucleic acid that has been generated by any sequencing method. A read therefore corresponds to the sequence information obtained from one strand of a nucleic acid fragment. For example, an RNA or DNA fragment where sequence has been generated from one strand in a single reaction will result in a single read. However, multiple reads for the same RNA or DNA strand can be generated where multiple copies of that RNA or DNA fragment exist in a sequencing project or where the strand has been sequenced multiple times. A read therefore corresponds to the purine or pyrimidine base calls or sequence determinations of a particular sequencing reaction.

[0065] As used herein, the terms “sequencing”, “obtaining a sequence” or “obtaining sequences” refer to nucleotide sequence information that is sufficient to identify or characterize the nucleic acid molecule, and could be the full length or a partial sequence information for the nucleic acid molecule (e.g., RNA or DNA).

[0066] As used herein the term “sequencing library” refers to a mixture of RNA or DNA fragments comprising the total RNA or DNA from a single organism, single cell or single sample for use in sequencing. Next-generation sequencing libraries are generally size- selected and ligated to sequencing adaptors prior to sequencing.

[0067] As used herein, the term “sequencing adaptor” refers to oligos bound to the 5' and 3' end of each RNA or DNA fragment in a sequencing library. Adaptors may contain platformdependent sequences that allow amplification of the fragment as well as sequences for priming the sequencing reaction.

[0068] As used herein, the term “barcode” can generally refer to any nucleic acid sequence used for identification purposes. A barcode can be a sequence used to identify a source of origin, for example a cell, genome, sample or another nucleic acid to which it is attached. In the context of the methods disclosed herein, a barcode can refer to a stretch of expressed, randomized nucleic acids which is used to uniquely label an individual neuron. Thus, a barcoded nucleic acid contains a barcode portion as well as other portions, e.g. proteinencoding portions. In addition, a sequencing barcode, also known as a sequencing index, refers to a unique nucleic acid sequence, for instance within a sequencing adaptor, which is used to identify the genomic origin of each amplicon in a sequencing library. Similarly, other types of barcodes used in this application include a slice specific identifier (SSI) and a unique molecular identifier (UMI).

[0069] As used herein, the term “multiplex” refers to pooling or otherwise mixing amplicons generated from multiple sources, sequencing the entire collection of amplicons in a single sequencing run and subsequently sorting and identifying the source of each read by a barcode sequence.

[0070] As used herein, the term “uniquely labeled neuron” in the context of the disclosed methods refers to a neuron which contains a barcoded nucleic acid that is not found in any other labeled neuron in the plurality of labeled neurons. In a practical application of the disclosed methods, at least 50%, more preferably at least 75%, more preferably at least 80%, more preferably at least 90%, more preferably at least 95%, 96%, 97%, 98% or 99% of the neurons labeled with a barcoded nucleic acid are uniquely labeled neurons.

[0071] As used herein, the term “expression construct” or “expression vector” refers to any- engineered nucleic acid which is introduced into a cell and is capable of expressing an RNA or protein. Non-limiting examples of expression constructs include recombinant plasmids and recombinant viral nucleic acids. Introduction of an engineered nucleic acid into a cell can be accomplished by a variety of methods including, but not limited to, transformation and transfection techniques, viral transduction, electroporation, chemically-induced uptake of exogenous nucleic acids, hydrodynamic delivery, lipofection, sonoporation and other methods know n to a person of ordinary skill in the field of molecular biology-. In some aspects of the disclosed methods, any of the chimeric viruses described herein can be used to label neurons with unique barcodes. In some aspects, a variety of methods for transgene delivery and expression as described herein can also be used for this purpose.

[0072] As used herein, the term “chimeric protein,” also known as a “fusion protein,” refers to any protein sequence which contains sequences from different sources. In the disclosed methods, a chimeric protein is used to bind a barcoded nucleic acid in a neuron and transport the barcoded nucleic acid to the synapse or axon of the neuron. The chimeric protein may be a modified synaptic or non-synaptic protein. Notably, transport by the chimeric protein should not be necessary when expression of the barcoded nucleic acid is high enough to reach axon terminals.

[0073] As used herein, the term “nucleic acid binding domain” refers to a protein domain or motif which is capable of recognizing and binding to a specific region of a nucleic acid. Examples of nucleic acid binding domains include nZ. which binds a boxB RNA motif, and MS-2 bacteriophage coat protein, which binds an MS-2 RNA stem-loop motif. Other RNA- binding protein domains which can be utilized in a chimeric carrier protein for the disclosed methods will be known to any person of ordinary skill in the art, including, but not limited to, customizable PUF class RNA-binding domains and PP7 bacteriophage coat protein binding site cassettes.

[0074] As used herein, the term "transgene” refers to a gene or genetic material that has been transferred or artificially introduced into the genome by a genetic engineering technique from one organism to another, i.e., the host organism.

[0075] As used herein, the term “transgene expression’" relates to the control of the amount and timing of appearance of the functional product of a transgene in a host organism.

[0076] The term "‘endogenous” as used herein refers to substances and processes originating from within an organism, tissue or cell.

[0077] The term “exogenous” as used herein refers to substances and processes originating from outside an organism, tissue or cell.

[0078] As used herein, the terms “optional” or “optionally” mean that the subsequently described event or circumstance may or may not occur and that the description includes instances where said event or circumstance occurs and instances where it does not.

[0079] As used herein, the term “sample” is meant a tissue or organ from a subject; a cell (either within a subject, taken directly from a subject, or a cell maintained in culture or from a cultured cell line); a cell lysate (or lysate fraction) or cell extract; or a solution containing one or more molecules derived from a cell or cellular material (e.g. a polypeptide or nucleic acid), which is assayed as described herein. For example, a sample can be a neuron or neuronal tissue sample that contains cells or cell components.

[0080] As used herein, the term "‘subject” refers to a mammal, a fish, a bird, a reptile, or an amphibian. The term “subject” also includes domesticated animals (e.g., cats, dogs, etc.), livestock (e.g., cattle, horses, pigs, sheep, goats, etc.), and laboratory7animals (e.g., mouse, rabbit, rat, guinea pig, fruit fly, non-human primate, etc ). In one aspect, a subject can be an animal. In another aspect, a subject can be a transgenic animal. The term does not denote a particular age or sex.

[0081] As used herein, the term “comprising” can include the aspects “consisting of’ and “consisting essentially of.”

[0082] The term “vector” or “construct” refers to a nucleic acid sequence capable of transporting into a cell another nucleic acid to which the vector sequence has been linked. The term “expression vector” includes any vector, (e.g., a plasmid, cosmid or phage chromosome) containing a gene construct in a form suitable for expression by a cell (e.g., linked to a transcriptional control element). “Plasmid” and “vector” are used interchangeably, as a plasmid is a commonly used form of vector. Moreover, the invention is intended to include other vectors which serve equivalent functions.

[0083] The term “expression vector” is herein to refer to vectors that are capable of directing the expression of genes to which they are operatively-linked. Common expression vectors of utility in recombinant DNA techniques are often in the form of plasmids. Recombinant expression vectors can comprise a nucleic acid as disclosed herein in a form suitable for expression of the acid in a host cell. In other words, the recombinant expression vectors can include one or more regulatory elements or promoters, which can be selected based on the host cells used for expression that is operatively linked to the nucleic acid sequence to be expressed.

[0084] “Modulate”, “modulating” and “modulation” as used herein mean a change in activity7or function or number. The change may be an increase or a decrease, an enhancement or an inhibition of the activity, function or number.

[0085] As used herein, the term “cell line” refers to a population of cells capable of continuous or prolonged growth and division in vitro. Often, cell lines are clonal populations derived from a single progenitor cell. It is further known in the art that spontaneous or induced changes can occur in karyotype during storage or transfer of such clonal populations. Therefore, cells derived from the cell line referred to may not be precisely identical to the ancestral cells or cultures, and the cell line referred to includes such variants.

[0086] Although the foregoing invention has been described in some detail by way of illustration and example for purposes of clarity of understanding, certain changes and modifications may be practiced within the scope of the appended claims.

[0087] All publications and patent applications mentioned in the specification are indicative of the level of those skilled in the art to which this invention pertains. All publications and patent applications are herein incorporated by reference to the same extent as if each individual publication or patent application was specifically and individually indicated to be incorporated by reference.

[0088] To understand axonal projections of individual neurons in the mouse brain, a high- throughput technique, MAPseq (Multiplexed Analysis of Projections by Sequencing), was developed using an engineered Sindbis vims (SINV) for MAPseq (SINV-MAPseq). In MAPseq, Sindbis viruses (SINV) encoding unique RNA sequences (barcodes) deliver each barcode to each neuron and the barcodes are transported to axon terminals. Then, by reading the barcodes via sequencing in target areas across the brain, projection patterns of individual neurons are identified which were hindered in the bulk-level projection (FIG. 1).

[0089] SINV is a member of the Alphavirus family which consists of structural and non- structural proteins encoded by its genomic RNA (FIG. 2A). SINV-MAPseq has structural proteins of SINV and genomic RNA encoding nonstructural proteins of SINV together with MAPseq elements. The infectivity of SINV-MAPseq is determined by the structural proteins. SINV infects some types of neurons well but not the other types, which makes it hard to do MAPseq in the later neurons. Also, among the earlier types of neurons, it is hard to discriminate each of them by their molecular types when they are infected together.

[0090] Therefore, to achieve the enforced or cell-type specific infection, a chimeric virus was created. While the receptor that is necessary and sufficient for the SINV infection was not identified, the receptor for Chikungunya virus (CHIKV), Mxra8, has been identified. As disclosed herein, a chimeric virus, CHIKV -MAPseq, was made by replacing the structural proteins of SINV-MAPseq with the structural proteins of CHIKV. Then, the infectivity of CHIKV -MAPseq can be dependent on the expression of the Mxra8 receptor. As also disclosed herein, the enforced or cell-type specific infection of CHIKV -MAPseq was achieved by expressing Mxra8 in the neurons that are difficult or hard to infect or solely in specific types of neurons by helper AAV.

[0091] The compositions and methods disclosed herein can solve at least two problems associated with MAPseq. The compositions and methods disclosed herein can map projections of thousands of individual neurons efficiently and it can infect some or specific neurons as well as discriminate molecular cell ty pes among infected neurons.

[0092] The compositions and methods disclosed herein can be used to show how individual neurons (and the type of neuron) construct a network across the brain.

[0093] Currently, no other techniques are available for high-throughput projection mapping of specific neurons at the single-neuron level. While barcoded anatomy resolved by sequencing (BARseq) which uses in-situ sequencing can be used to identity’ cell-types, in-situ sequencing is commonly used and is further limited to the use of the same viruses as the original MAPseq which cannot discriminate different neurons and many neurons are resistant to infection by that virus. METHODS

[0094] Disclosed herein are methods of selectively labeling a neuron or a population of neurons. In some aspects, two or more neurons can be labeled, wherein the two or more neurons are located in the same anatomical region. Also disclosed herein are methods of selectively labeling a neuronal pathway in area or the same anatomical region.

[0095] In some aspects, the method comprises: a) infecting a neuron or population of neurons with a vims encoding an exogenous gene, wherein the exogenous gene encodes a receptor of an alphavirus; and b) infecting the same neuron or population of neurons in a) with a chimeric virus in an amount to express a target protein, wherein the chimeric vims comprises: i) a RNA nucleic acid encoding a nucleic acid barcode, and ii) a structure protein capable of specifically binding to the alphavirus receptor encoded by the exogenous gene of a); thereby selectively labeling a neuron.

[0096] In some aspects, the method comprises: a) infecting a neuron or population of neurons in a subject with a chimeric virus in an amount to express a target protein, wherein the chimeric virus comprises: i) a RNA nucleic acid encoding a nucleic acid barcode, and ii) a structure protein capable of specifically binding to an alphavirus receptor; wherein the neuron or population of neurons in the subject expresses a receptor of an alphavirus that is capable of specifically binding to the structure protein of a) ii); thereby selectively labeling a neuron.

[0097] In some aspects, the methods described herein can further comprise: isolating the barcode nucleic acid from each labeled neuron; amplifying the barcode nucleic acid from each labeled neuron generating barcode amplicons; sequencing the amplicons; determining barcode abundance in one or more brain regions; converting the barcode abundance to a matrix of single neuron projection patterns.

[0098] Further disclosed herein are methods for labeling a neuronal pathway in an area or species which is hardly (e.g., rarely) infected by Sindbis virus. In some aspects, the methods can comprise: a) infecting a population of neurons with chimeric viruses, wherein the chimeric alphaviruses comprise: i) a self-amplifying but propagation-incompetent RNA nucleic acid encoding a barcode nucleic acid, and ii) structure proteins of alphaviruses with vary ing infectivity for the neuronal pathway, thereby enhancing the labeling of a neuronal pathway across areas and species.

[0099] Also disclosed herein are methods of screening a compound for its effects on neurons or a neuronal pathway. In some aspects, the methods can comprise: obtaining a map of single neuron projections before and after exposure of the neurons or the neuronal pathway to a compound according the method comprising: a) infecting a neuron or population of neurons with a vims encoding an exogenous gene, wherein the exogenous gene encodes a receptor of an alphavirus: and b) infecting the same neuron or population of neurons in a) with a chimeric virus in an amount to express a target protein, wherein the chimeric vims comprises: i) a RNA nucleic acid encoding a nucleic acid barcode, and ii) a structure protein capable of specifically binding to the alphavirus receptor encoded by the exogenous gene of a); or the method comprising a) infecting a neuron or population of neurons in a subject with a chimeric virus in an amount to express a target protein, wherein the chimeric vims comprises: i) a RNA nucleic acid encoding a nucleic acid barcode, and ii) a structure protein capable of specifically binding to an alphavirus receptor; wherein the neuron or population of neurons in the subject expresses a receptor of an alphavirus that is capable of specifically binding to the structure protein of a) ii). In some aspects, wherein the method comprises determining whether the compound is toxic to the neurons or the neuronal pathway or identifying changes to one or more projectional cell types. In some aspects, wherein the method comprises determining the effects of the compound on one or more (or different) projectional cell types. In some aspects, wherein the method comprises determining the effects of the compound on the neurons or the neuronal pathway. In some aspects, the effect of the compound can be a change in the projection pattern of the neurons or the neuronal pathway. In some aspects, the effect of the compound can be a change in projection strength of the neurons or the neuronal pathway. In some aspects, the neurons or the neuronal pathway can be from a model of a human disease. In some aspects, the human disease can be Parkinson’s disease or Alzheimer’s disease. In some aspects, the neurons or the neuronal pathway can be from a model of a human disorder. In some aspects, the human disorder can be a neurodevelopmental disorder, a neuropsychiatric disorder, or an addiction. In some aspects, the methods can comprise obtaining a map of single neuron projections before and after exposure of the neurons to a drug or a compound.

[0100] In some aspects, the virus in a) can be an adeno-associated virus (AAV). In some aspects, the AAV DNA can be from any AAV serotype including but not limited to AAV1, AAV2.1 serotype, AAV 2.5 serotype, AAV 2.8 serotype. AAV 2.9 serotype. AAV DJ serotype, and AAV PhP.eB serotype. In some aspects, the AAV can be an AAV1 serotype.

[0101] In some aspects, AAV encoding the exogenous gene can be administered in sufficient amounts to transfect the cells of a desired region or area and to provide sufficient levels of gene transfer and expression without undue adverse effects. Adeno-associated virus (AAV) is a replication-deficient parvovirus, the singlestranded DNA genome of which is about 4.7 kb in length including 145 nucleotide inverted terminal repeat (ITRs). There are multiple serotypes of AAV. The nucleotide sequences of the genomes of the AAV serotypes are known. For example, the nucleotide sequence of the AAV serotype 2 (AAV2) genome is presented in Srivastava et al., J Virol, 45: 555-564 (1983) as corrected by Ruffing et al., J Gen Virol, 75: 3385-3392 (1994), and the complete genome of AAV-1 is provided in GenBank Accession No. NC_002077.

[0102] As used herein, the term '‘AAV” is a standard abbreviation for adeno-associated virus. Adeno-associated virus is a single-stranded DNA parvovirus that grows only in cells in which certain functions are provided by a co-infecting helper virus. There are currently thirteen serotypes of AAV that have been characterized. General information and reviews of AAV can be found in, for example. Carter, 1989, Handbook of Parvoviruses, Vol. 1, pp. 169-228, and Bems, 1990, Virology, pp. 1743-1764, Raven Press, (New York). However, it is fully expected that these same principles will be applicable to additional AAV serotypes since it is well known that the various serotypes are quite closely related, both structurally and functionally, even at the genetic level. (See, for example, Blacklowe, 1988, pp. 165-174 of Parvoviruses and Human Disease, J. R. Pattison, ed.; and Rose, Comprehensive Virology 3: 1- 61 (1974)). For example, all AAV serotypes apparently exhibit very similar replication properties mediated by homologous rep genes; and all bear three related capsid proteins such as those expressed in AAV2. The degree of relatedness is further suggested by heteroduplex analysis which reveals extensive cross-hybridization between serotypes along the length of the genome; and the presence of analogous self-annealing segments at the termini that correspond to “inverted terminal repeat sequences” (ITRs). The similar infectivity patterns also suggest that the replication functions in each serotype are under similar regulatory control.

[0103] An “AAV vector” as used herein refers to a vector comprising one or more polynucleotides of interest (or transgenes) that are flanked by AAV terminal repeat sequences (ITRs). Such AAV vectors can be replicated and packaged into infectious viral particles when present in a host cell that has been transfected with a vector encoding and expressing rep and cap gene products.

[0104] In some aspects, the neuron or population of neurons can be a peripheral neuron or peripheral neurons. In some aspects, the neuron or population of neurons can be a central nervous system neuron or central nervous system neurons. In some aspects, the neuron or population of neurons can be in any Cre mouse line. In some aspects, the neuron or population of neurons can be in a Cux2-Cre mouse line or a Rbp4-Cre mouse line. In some aspects, when an AAV expressing Mxra8 is used in the Cre dependent manner, any Cre mouse line can be used. For example, considering important role of dopaminergic pathway in Parkinson’s disease, DAT-Cre can be used.

[0105] In some aspects, wherein the neuron or population of neurons are infected with the chimeric virus at a multiplicity of infection of about 1.

[0106] In some aspects, the same neuron or population of neurons can be infected 2 to 4 weeks after the neuron or population of neurons were infected with a virus encoding an exogenous gene, wherein the exogenous gene encodes a receptor of an alphavirus. In some aspects, the same neuron or population of neurons can be infected about 3 weeks after the neuron or population of neurons were infected with a virus encoding an exogenous gene, wherein the exogenous gene encodes a receptor of an alphavirus. In some aspects, the same neuron or population of neurons can be infected after a sufficient amount of time such that the neuron or population of neurons that were infected with a virus encoding an exogenous gene, wherein the exogenous gene encodes a receptor of an alphavirus, is expressed. In some aspects, step b) of the method can occurs 2-4 weeks after step a) of the method.

[0107] In some aspects, the exogenous gene can be regulatable (e.g., CRE). In some aspects, wherein the exogenous gene can beAlxra#. For example, Mxra8 (MXRA8 matrix remodeling associated 8) gene encodes Mxra8 protein.

[0108] In some aspects, the exogenous gene in the virus of step (a) can be a receptor of an alphavirus. In some aspects, the exogenous gene in the virus can be a receptor of an alphavirus. In some aspects, the exogenous gene in the virus can be a receptor of an alphavirus, wherein the alphavirus is rarely expressed endogenously in the area where the virus is injected. In some aspects, the receptor of the alphavirus can be regulatable.

[0109] In some aspects, the target protein can be Mxra8 or the Mxra9 receptor. Mxra8 is a receptor of CHIKV.

[0110] In some aspects, the barcoded nucleic acid can be RNA. In some aspects, the RNA nucleic acid encoding a nucleic acid barcode can be a self-amplifying propagationincompetent RNA nucleic acid. In some aspects, the barcode in each of the barcoded nucleic acids has a length of (k) nucleotides, wherein 4k is greater than the number of neurons to be labeled. In some aspects, the barcoded nucleic acid contains a barcode region that is about 30 nucleotides in length. In some aspects, wherein the barcoded nucleic acid can encode a fluorescent marker. In some aspects, the subject can be an animal. In some aspects, the animal can be a transgenic animal. In some aspects, the subject can be a transgenic animal. In some aspects, the subject can be a mouse. In some aspects, the subject can be a rat. In some aspects, the subject can be a non-human primate. In some aspects, the subject can be a zebra fish. In some aspects, the subject can be a xenopus.

[0111] In some aspects, the subject can be engineered to specifically express the receptor of the alphavirus. In some aspects, the subject can be engineered to specifically express the receptor of the alphavirus in a specific neuronal cell population. In some aspects, the receptor of the alphavirus can be regulatable.

[0112] In some aspects, the alphavirus can be CHIKV, VEEV, EEEV, WEEV, SFV, or SINV (TR339). Tonate, or Everglades virus.

[0113] In some aspects, the chimeric virus can further comprise a detection label. As used herein, a detection label is any molecule that can be associated with amplified nucleic acid, directly or indirectly, and which results in a measurable, detectable signal, either directly or indirectly. In some aspects, the detection label can be any fluorescent protein. In some aspects, the detection label can be GFP.

[0114] The methods described herein can be used to map neurons and their projections to close and distinct brain regions by way of long-range axonal projections. The methods disclosed herein can differentiate functionally distinct population of neurons that may be intermingled within a small region.

[0115] In some aspects, the methods described herein can be used to generate single neuron projection maps (e.g., see, FIG. 5). In some aspects, the disclosed workflow can comprise the steps of: (a) a flash frozen brain is cryosectioned and areas of interest are dissected out. Total RNA from every area can then extracted individually; (b) a known amount of spike-in RNA and RT primers containing unique SSI and UMIs can be added to the total RNA from every area. Double stranded cDNA is produced and leftover RT primers can be digested using Exonuclease I to avoid UMI containing primers to participate in subsequent PCR reactions. Two rounds of nested PCR can be performed, bringing in the PE2 sequencing primer binding site and P7 sequence as 5’ overhangs of the reverse primer. After gel extraction, the amplicons are ready for sequencing.

[0116] Also disclosed herein are methods for obtaining a map of single neuron projections in a region containing projections of a plurality of labeled neurons. In some aspects, the methods can comprise a) infecting a neuron or population of neurons with a virus encoding an exogenous gene, wherein the exogenous gene encodes a receptor of an alphavirus; and b) infecting the same neuron or population of neurons in a) with a chimeric virus in an amount to express a target protein, wherein the chimeric virus comprises: i) a RNA nucleic acid encoding a nucleic acid barcode, and ii) a structure protein capable of specifically binding to the alphavirus receptor encoded by the exogenous gene of a); thereby selectively labeling a neuron and further comprising i) dissecting the region containing projections of the plurality of barcoded neurons into sections; ii) isolating the barcoded nucleic acids from each dissected section; iii) amplifying the isolated barcoded nucleic acids; iv) sequencing the amplified barcoded nucleic acids; and v) determining associations between identical barcode sequences thereby obtaining a map of single neuron projections in the region. In some aspects, the region being mapped can have projections from neurons belonging to the central nervous system or the peripheral nervous system. Thus, a neuronal map of any region e.g., any organ or tissue, for example muscle or gut tissue, which neurons project through can be obtained using the methods disclosed herein.

[0117] In some aspects, the method comprises: a) infecting a neuron or population of neurons in a subject with a chimeric virus in an amount to express a target protein, wherein the chimeric virus comprises: i) a RNA nucleic acid encoding a nucleic acid barcode, and ii) a structure protein capable of specifically binding to an alphavirus receptor; wherein the neuron or population of neurons in the subj ect expresses a receptor of an alphavirus that is capable of specifically binding to the structure protein of a) ii); thereby selectively labeling a neuron and further comprising i) dissecting the region containing projections of the plurality of barcoded neurons into sections; ii) isolating the barcoded nucleic acids from each dissected section; iii) amplify ing the isolated barcoded nucleic acids; iv) sequencing the amplified barcoded nucleic acids; and v) determining associations between identical barcode sequences thereby obtaining a map of single neuron projections in the region. In some aspects, the region being mapped can have projections from neurons belonging to the central nervous system or the peripheral nervous system. Thus, a neuronal map of any region e.g., any organ or tissue, for example muscle or gut tissue, which neurons project through can be obtained using the methods disclosed herein.

[0118] In some aspects of the disclosed methods, wherein in step (i) the region can be dissected by a gross dissection method. In some aspects, wherein in step (i) the region can be dissected by a laser-capture microdissection method.

[0119] In some aspects of the disclosed methods, wherein in step (ii) the barcoded nucleic acids can be isolated by TIVA tagging. In some aspects, wherein in step (ii) a known amount of spike-in nucleic acid molecules can be added to every sample of isolated barcoded nucleic acids in order to determine the efficiency of barcode sequence recovery. In some aspects, wherein step (ii) further comprises reverse transcription of the isolated barcoded nucleic acids. In some aspects, wherein step (ii) further comprises adding a slice specific identifier (SSI) to the barcoded nucleic acids from a dissected area. In some aspects, wherein step (ii) further comprises adding a unique molecular identifier (UMI) to each barcoded nucleic acid from each dissected area.

[0120] In some aspects of the disclosed methods, the sequences of the barcoded nucleic acids can be obtained by a FISSEQ method.

[0121] In some aspects of the disclosed methods, in step (v) a threshold of true, noncontaminating barcode expression can be determined by the number of barcode sequences recovered from cells that lack a barcoded construct. In some aspects of the disclosed methods, in step (v) the barcode sequences in the injection site (reference barcodes) can be matched with the barcode sequences in the target sites to create a barcode matrix of size [number of reference barcodes] x [number of target sites+number of injection sites]. In some aspects of the disclosed methods, each target area of the barcode matrix can be normalized by the number of unique spike-in molecules detected in each. In some aspects of the disclosed methods, each target area of the barcode matrix can be normalized by the amount of I3-actin per pl of total RNA. In some aspects of the disclosed methods, all barcodes can be normalized to sum to 1 across all target areas. In some aspects of the disclosed methods, in step (v) peaks of barcode molecule counts are defined by i) being at least half as high as the maximal barcode count across all target sites; ii) being separated by at least three slices; and iii) rising at least their half maximal height from their surroundings ('prominence'), thereby defining peaks for use in determining associations between identical barcode sequences.

[0122] In some aspects of the disclosed methods, the labeled neurons belong to the central nervous system. In some aspects of the disclosed methods, the labeled neurons belong to the peripheral nervous system.

[0123] In some aspects of the disclosed methods, a map of single neuron projections can be generated.

[0124] Described herein are methods for obtaining a plurality of labeled neurons, comprising infecting neurons with a modified, barcoded virus library. In some aspects of the disclosed methods, the virus can be a Sindbus or CHIKV virus. In some aspects of the disclosed methods, neurons in the brain can be labeled by injecting the barcoded virus library into only a specific portion or structure of the brain or brain-stem. For example, the LC is an example of one such specific site; however, any other specific site or location of the brain may be injected. In some aspects of the disclosed methods, the barcoded virus library can be injected into more than one portion or structure of the brain.

[0125] In some aspects of the disclosed methods, the library can be sufficiently diverse to uniquely label at least 50%, more preferably at least 75%, more preferably at least 80%, more preferably at least 90%, more preferably at least 99% of the total number of neurons that are infected.

[0126] In some aspects of the disclosed methods, the neurons can be infected with a barcoded virus library at a multiplicity’ of infection (MOI) of about 1. In some aspects, the virus can be Sindbus or CHIKV.

[0127] In some aspects, the methods disclosed herein can further comprise infecting the neurons with a nucleic acid library which is capable of altering gene expression. In some aspects, the functional library can be a CRISPR library. In some aspects of the disclosed methods, the functional library’ can be a shRNA library. In some aspects of the disclosed methods, a drug or compound can applied to the plurality of labeled neurons to determine if the drug is capable of counteracting a wiring defect in a neuron caused by the altered gene expression from the functional library.

[0128] In some aspects of the disclosed methods for selectively’ labeling a neuron or selectively labeling a population of neurons, the methods can comprise injecting AAV, wherein the AAV encodes receptor gene in a targeted area. After viral expression in the neuron or the population of neurons (e.g., about 3 weeks), the chimeric virus in injection into the same area. In some aspects, the methods can comprise injection one or more chimeric viruses into the targeted area or species. After expression of the one or more chimeric viruses, the desired target areas of the neuronal pathway can be dissected and extract the total RNA from each dissected area.

[0129] In some aspects of the disclosed methods, to create a sequencing library' using the extracted RNA, each area can be synthesized using RT primers specific to the barcode RNA. Since separate RT primers can be used with area or sample specific indexes (SSI), each target area can be identified after sequencing. RT primers also have UMI to quantify the abundance of barcode RNA. After RT, we digest RT primers using Exo nuclease I to exclude nonspecific fragments. Then, RT products can be amplified by PCR adding adaptors, indexes, and sequencing primer binding sites for the next generation sequencing.

[0130] In some aspects of the disclosed methods a projection matrix (cell x projection strength) can be constructs where each cell is identified by the unique barcode and the projection strength is estimated by the number of UMI of the barcode. Based on the matrix, cells can be clustered according to their projection patterns.

[0131] Kits

[0132] Disclosed herein are kits comprising any of the agents described herein. In some aspects, any of the agents (e.g., chimeric viruses) disclosed herein can be assembled into research kits to facilitate their use in diagnostic or research applications. A kit can include one or more containers housing the components of the disclosure and instructions for use. Specifically, such kits may include one or more agents described herein, along with instructions describing the intended application and the proper use of these agents. In some aspects, the agents in a kit can be in a formulation and dosage suitable for a particular application or for a particular method. Kits for research purposes can contain the components in appropriate concentrations or quantities for running various experiments.

[0133] Also disclosed herein are kits for producing an AAV and chimeric viruses. In some aspects, the kit can comprise a container housing an isolated nucleic acid encoding one or more nucleic acid barcodes. In some aspects, the kits can further comprise instructions for producing the AAV and the chimeric viruses. In some aspects, the kit further comprises at least one container housing an AAV vector, wherein the AAV vector comprises a transgene.

[0134] In some aspects, the kits can be designed to facilitate use of the methods described herein by researchers and can take many forms. Each of the compositions of the kit, where applicable, may be provided in liquid form (e.g., in solution), or in solid form, (e.g., a dry powder). In some aspects, some of the compositions can be constitutable or otherwise processable (e.g., to an active form), for example, by the addition of a suitable solvent or other species (for example, water or a cell culture medium), which may or may not be provided with the kit. As used herein, “instructions” can define a component of instruction and / or promotion, and typically involve written instructions on or associated with packaging of the disclosure. Instructions also can include any oral or electronic instructions provided in any manner such that a user will clearly recognize that the instructions can be associated with the kit, for example, audiovisual (e.g., videotape, DVD, etc.), Internet, and / or web-based communications, etc. The written instructions can be in a form prescribed by a governmental agency regulating the manufacture, use or sale of biological products, which instructions can also reflect approval by the agency of manufacture, use or sale for animal administration.

[0135] The kits disclosed herein can also contain any one or more of the components described herein in one or more containers. In some aspects, the kits can include instructions for mixing one or more components of the kit and / or isolating and mixing a sample and applying to a subject. The kits can include a container housing agents described herein. The agents can be in the form of a liquid, gel or solid (powder). The agents can be prepared sterilely, packaged in syringe, and shipped refrigerated. Alternatively, it can be housed in a vial or other container for storage. A second container can have other agents prepared sterilely. Alternatively, the kits can include the active agents premixed and shipped in a syringe, vial, tube, or other container. The kits can have one or more or all of the components required to administer the agents to an animal, such as a syringe, topical application devices, or iv needle tubing and bag, particularly in the case of the kits for producing specific somatic animal models.

[0136] EXAMPLES

[0137] Example 1: Mapping single-neuron projections with improved sensitivity and cell-type specificity

[0138] Understanding the statistics of single neuron projection patterns of different cell types is important for deciphering information flow between brain regions. While bulk-level projection data is readily available in the mouse, single-neuron projection data is still more challenging to obtain, as it requires high-resolution whole-brain imagining and neuron reconstructions. MAPseq, which uses cellular barcoding and DNA sequencing instead of imaging to rapidly map the projections of thousands of single neurons in individual animals can sidestep these difficulties; however, as MAPseq is based on the RNA virus Sindbis, it cannot be used with popular DNA recombinases such as Cre or Flp to enable cell-type specific projection mapping. To address this shortcoming, an improved version of MAPseq was developed that allows single-cell projection mapping of specific cell types as defined by recombinase expression. In addition, the MAPseq sequencing library preparation w as further improved to capture fine projections in more neurons, achieving close to 4 times improved sensitivity with 6 times decreased cost per sample simultaneously. The method disclosed herein was validated in the mouse motor cortex and then applied to uncover the brain-wide projections of midbrain dopaminergic neurons at single-cell resolution.

[0139] SINV is a family of Alphaviruses, which consists of non-structural (nsP) and structural proteins (sP) encoded by its genomic RNA (FIG. 2A). To produce SINV -MAPseq, modified SINV genomic RNA encoding barcodes and GFP (MAPseq RNA) were cotransfected with RNA encoding SINV sP in BHK cells (FIG. 2B), w here the infectivity of the virus is governed by SINV sP. Given that the infectivity of Alphaviruses varies (Holmes et al., 2020. PLoS Pathog; Zimmerman et al., 2023, Cell) (FIG. 2C), it was reasoned that the infectivity of the MAPseq virus could be regulated by sP of different Alphaviruses. Thus, chimeric MAPseq viruses were generated using the sP of different Alphaviruses (FIG. 2D).

[0140] The viruses and plasmid constructs are summarized in Table 1.

[0141] Table 1. Viruses and plasmids that we generated.

[0142] Among them, CHIKV-MAPseq permits cell-type specific MAPseq in the mouse since the receptor of CHIKV, Mxra8, is rarely expressed in the mouse brain. For example, while SINV -MAPseq infects the mouse cortex without specificity (FIG. 3 A), Mxra8 can be expressed cell-type specifically by Adeno-associated virus (AAV) making those neurons infective by CHIKV-MAPseq (FIG. 3B). In the case of acquired infectivity, for example, while SINV-MAPseq rarely infects neurons in the mouse cerebellar nuclei (FIG. 3C], EEEV- MAPseq displays better infectivity (FIG. 3D) in those neurons. Moreover, while SINV- MAPseq rarely infects neurons of some species other than the mouse, chimeric MAPseq viruses show better infectivity to those neurons.

[0143] Next, the cell -type specific infection of CHIKV-MAPseq was validated in the mouse primary motor cortex (MOp) which consists of layer 2 / 3 (L2 / 3), layer 5 (L5). and layer 6 (L6). Each layer consists of intratelencephalic (IT) neurons. IT and extratelencephalic (ET) neurons, and corticothalamic (CT) neurons based on projection patterns (Munoz-Castaneda et al., 2021, Nature), respectively (FIG. 4A). Since Cux2 is a marker of L2 / 3, Cux2 positive neurons are IT, while Rbp4, a marker of L5, positive neurons are IT and ET. Based this information, Cux2 or Rbp4 neurons were targeted to examine whether CHIKV-MAPseq solely infects the targeted neurons with expected projection patterns. At first. SINV-MAPseq was used in WT mice as a control and it resulted in non-specific infection across layers and all ty pes of IT, ET, and CT neurons as expected (FIGS. 4B-D). Next, to specifically express Mxra8 in Cux2 or Rbp4 neurons, AAV expressing Mxra8 was injected together with tdTomato (tdT) in the Cre recombinase-dependent manner in Cux2-Cre or Rbp4-Cre mouse lines, followed by CHIKV-MAPseq injection (FIGS. 4E, F). As expected, CHIKV-MAPseq was limited to infecting the targeted neurons resulting in IT neurons solely in the Cux2-Cre line and IT and ET neurons in the Rbp4-Cre line (FIGS. 4G-I).

[0144] Example 2: Mapping the connectomic architecture of the mouse midbrain dopamine system using cell type-specific barcoding

[0145] Brain-wide neural circuits are formed by the complex axonal branching patterns of individual neurons. Described herein is a method referred to herein as POINTseq, which is a high-throughput and user friendly barcoded connectomics method that allows cell typespecific single-cell projection mapping for thousands of neurons per animal. POINTseq builds on MAPseq by leveraging viral pseudotyping and a specific alphavirus-cellular receptor pair to allow cell type specific barcoding and thus directly integrates with the viral- genetic circuit analysis toolbox. POINTseq mapping of genetically and projection-defined cell populations was validated in the mouse motor cortex. POINTseq was then applied to midbrain dopaminergic neurons and the brain-wide single-cell projections of 3,813 dopaminergic neurons were reconstructed in ventral tegmental area (VTA) and substantia nigra pars compacta (SNc). 34 connectomic cell t pes, vastly exceeding the known diversity of dopaminergic projection types were defined, and structured projection motifs of VTA and SNc neurons were identified. These data provide insight into the anatomical hardware to support diverse dopamine functions.

[0146] The methods described herein (e.g., POINTseq) allow for massively multiplexed single-cell projection mapping of cell types of interest showing that projections organize into non-random motifs that can mediate parallel dopamine signalling.

[0147] Distant brain regions communicate and collaborate through long range axonal projections to perform computations (Chklovskii, D.B., and Koulakov, A. A. (2004) Annu Rev Neurosci 27, 369-392; Swanson, L.W.. and Bota, M. (2010) Proc Natl Acad Sci U S A 107, 20610-20617; Luo. L. (2021). Science 373, eabg7285; and Whitesell. J.D., et al. (2021) Neuron 109, 545-559). Axons of individual neurons often branch to innervate several and sometimes dozens of dow nstream brain regions (Kebschull, J.M., et al. (2016) Neuron 91, 975-987; and Munoz-Castaneda, R., et al. (2021) Nature 598, 159-166). As the regions innervated by one neuron, to a first approximation, receive the same signals, groups of neurons that have similar projection patterns define the anatomical output pathways of a brain region and can be considered connectomic cell types. Defining these output pathways, understanding how they intersect with other cellular modalities, and how they function in behavior has been a central goal in systems neuroscience (Luo, L. (2021) Science 373, eabg7285; Peng, H., et al. (2021) Nature 598, 174-181; and Christensen, A.J., et al. (2022) Curr. Opin. Neurobiol. 77, 102630). In particular, the output pathways of neuromodulatory systems with their broad projections and many functions have been difficult to understand. For example, in the midbrain dopaminergic (DA) system, approximately 10.000 DA neurons per hemisphere (Brichta, L., and Greengard, P. (2014) Front. Neuroanat. 8, 152) in the mouse are intermingled with other cell types in the ventral tegmental area (VTA) and substantia nigra pars compacta (SNc). These neurons project to many regions including the limbic system, cerebral cortex, striatum, and olfactory system to mediate important functions including reward processing, reinforcement, aggression, and locomotion (Lammel, S., et al. (2014) Neuropharmacology 76 Pt B, 351-359; Garritsen, O., et al. (2023) Nat Rev Neurosci 24, 134-152; Heymann, G., et al. (2020) Neuron 105, 909-920.e5; Beier, K.T., et al. (2015) Cell 162, 622-634; Menegas, W„ et al. (2018) Nat Neurosci 21, 1421-1430; Mahadevia, D., et al. (2021) Nat. Commun. 12. 6796; Howe, M.W., and Dombeck, D.A. (2016) Nature 535, 505-510; and Dai, B., et al. (2025) Nature 639, 430-437). How different types of DA cells map onto different functions and circuits, however, remains an open question.

[0148] The simplest operational definition of an output pathw ay is as neurons of a certain genetic type that project to a single target region as identified by a retrograde tracer (FIG. 20A). This approach has provided a fruitful anatomical basis for functional studies, revealing, for example, that DA neurons projecting to the nucleus accumbens (NAc) or the caudate putamen (CP) preferentially respond to reward and locomotion, respectively (Howe, M.W., and Dombeck, D.A. (2016) Nature 535, 505-510; Schultz, W., et al. (1997) Science 275, 1593-1599; and Hefti, F., et al. (1985) Neuropharmacology 24, 19-23). Similarly, DA neurons projecting to specific subregions of NAc and CP exhibiting differential responses to rewarding or aversive stimuli (Lammel, S., et al. (2014). Neuropharmacology 76 Pt B, 351— 359; Heymann, G.. et al. (2020). Neuron 105, 909-920; Menegas. W., et al. (2018) Nat Neurosci 21. 1421-1430; Lammel, S.. et al. (2011) Neuron 70, 855-862; and Lemer. T.N.. et al. (2015) Cell 162, 635-647). Defining output pathways by a single retrograde tracer, however, does not provide information about the other brain regions that the labeled neurons innervate. Including this information can bring the definition of output pathways closer to the anatomical ground truth and often reveals additional functional heterogeneity (Economo, M.N., et al. (2018) Nature 563, 79-84; and Han, Y , et al. (2018) Nature 556, 51-56). Within the framework of retrograde tracing, some co-innervation information can be incorporated into the analysis of output pathways by using additional retrograde tracers (Beier, K.T., et al. (2015) Cell 162, 622-634; Mahadevia, D„ et al. (2021). Nat. Commun. 12, 6796; Lammel, S„ et al. (2008). Neuron 57, 760-773; and Fallon, J.H. (1981) J. Neurosci. 7, 1361-1368), or by mapping the brain- wide collateralization patterns of neurons that project to a particular target region by combining retrograde and anterograde mapping (Beier, K.T., et al. (2015) Cell 162. 622-634; Beier, K.T., et al. (2019) Cell Rep. 26, 159-167; Schwarz, L.A., et al. (2015) Nature 524, 88-92; Ren, J., et al. (2018) Cell 175, 472-487).

[0149] The definition of output pathways by retrograde tracing, however, is fundamentally limited, as it does not capture the brain-wide projection patterns of individual neurons, or any information about projection strength of individual neurons (FIG. 20A). Instead, it implicitly assumes that the target region injected with the retrograde tracer is exclusively innervated by a specific connectomic cell type. When different connectomic cell types overlap in their projections to this target region, even if they do so at different strengths, these cell types will be labeled by the retrograde tracer, leading to averaging over connectomic types and muddying the analysis of circuit function (KebschulL J.M., et al. (2016) Neuron 91, 975-987; and Schwarz, L.A., et al. (2015) Nature 524, 88-92). While the appropriate selection of target regions for retrograde labeling can mitigate this effect, a clean answer is not guaranteed: neither the ground truth to confirm that a homogenous cell type is labeled nor is it necessarily true that each connectomic type exclusively innervates at least one target region.

[0150] Anterograde single neuron projection mapping does not average over individual neurons and thus is guaranteed to reveal the output pathways of a brain region, independent of projectional overlap of connectomic cell types. Such tracing, however, is traditionally very time intensive, typically limiting the number of reconstructed cells to less than 100 (Economo, M.N., et al. (2018) Nature 563, 79-84; and Aransay, A., et al. (2015) Front Neuroanat 9, 59), preventing its widespread application and limiting the statistical power in defining connectomic cell types (Han, Y., et al. (2018). Nature 556, 51-56). As a result, and despite the widespread interest in the DA system, the single-cell projections of midbrain DA neurons remain poorly understood. A previous study reconstructed a total of 30 DA neurons in the mouse VTA (Aransay, A., et al. (2015) Front Neuroanat 9, 59). Recent technological developments in volumetric imaging such as fMOST (Zheng, T., et al. (2013) Opt. Express 21, 9839-9850; and Zheng, T., et al. (2019) Biomed. Opt. Express 10, 4075-4096), 2-photon serial tomography (Kim. Y., et al. (2017) Cell 171, 456-469; and Winnubst, J., et al. (2019) Cell 179, 268-281), and light-sheet microscopy (Daetwyler, S., and Fiolka, R.P. (2023) Commun. Biol. 6, 502; Tavakoli, M.R., et al. (2025) Nature; and Glaser, A., et al. (2024) doi.org / 10.7554 / elife.91979.2) have enabled more efficient brain-wide tracing, allowing the collection of hundreds to thousands of single neuron reconstructions. However, this increase in throughput comes at the cost of highly specialized equipment and computational problems in handling vast amounts of imaging data and tracing individual cells. Moreover, the number of neurons traced is still limited to 50-100 cells per animal (Winnubst, J., et al. (2019) Cell 179, 268-281; and Gao, L„ et al. (2022) Nat. Neurosci. 25, 515-529).

[0151] To allow rapid single neuron tracing at scale democratically in any laboratory without the need for specialized equipment, a barcoded connectomics method MAPseq was developed, which offers a high-throughput, imaging-free approach for mapping brain-wide single-neuron projections (Kebschull, J.M., et al. (2016) Neuron 91, 975-987). In MAPseq, thousands of neurons per animal are labelled with RNA barcodes by infection with a barcoded library of Sindbis virus, a positive sense RNA virus from the alphavirus family (Kebschull, J.M., et al. (2016) Front Neuroanat 10, 56). Rapid replication of Sindbis virus genomic RNA inside each cell ensures robust expression of a single barcode per cell, a feature important to barcoded connectomics. The barcodes are then trafficked to axon terminals, where they can be detected and quantified by Illumina sequencing as proxies for axonal projection strength. By analyzing the sequencing data, thousands of single-neuron projections are reconstructed from a single animal within a week (Kebschull, J.M., et al. (2016) Neuron 91, 975-987; Munoz-Castaneda. R.. et al. (2021) Nature 598. 159-166; Han. Y , et al. (2018) Nature 556, 51-56; Chen, X , et al. (2019) Cell 179, 772-786; Chen, Y , et al. (2022) Cell 185, 4117-4134; Gergues, M.M., et al. (2020) Nat Neurosci 23, 1444-1452; Huang, L„ et al. (2020) Cell 183, 2040; Sun, Y.-C., et al. (2021) Nat Neurosci 24, 873-885; and Zeisler, Z.R., et al. (2023) Neuron 111, 3307-3320).

[0152] Despite these advances in throughput, MAPseq has a major limitation for its use in the viral genetic dissection of neuronal circuits (FIG. 20B). As it is based on an RNA virus, genetic cell type information cannot be directly incorporated into the tracing experiments using conventional tools such as e g., Cre-driver lines (Whitesell, J.D., et al. (2021) Neuron 109, 545-559; and Klingler, E., et al. (2021) Nature 599, 453-457). MAPseq has been combined with single cell RNAseq (Klingler, E., et al. (2021) Nature 599, 453-457) or in situ sequencing (Chen, X., et al. (2019). Cell 179, 772-786; and Sun, Y.-C., et al. (2021). Nat Neurosci 24, 873-885) of the barcoded neurons to directly bridge barcode identity and hence projections and endogenous gene expression in single neurons. However, specialized hardware requirements, expensive reagents, and work time per animal so far limit the widespread application of these methods across laboratories. A simpler integration of barcoded connectomics and genetics is therefore needed to enable the dissection of neuronal circuits at single neuron resolution in any laboratory.

[0153] Described herein is POINTseq (projections of interest by sequencing), a cell typespecific barcoded connectomics tool, that uses cell type specific infection by a pseudotyped barcoded Sindbis virus and DNA sequencing for rapid single-cell projection mapping of cell types of interest. Importantly, POINTseq readily integrates with the standard genetic neuroscience toolkit and requires no specialized equipment, as an example of the utility of the method, it was applied to mapping the brain-wide projection of VTA and SNc DA neurons. A large number of connectomic cell types that were inaccessible to retrograde tracing based circuit dissection methods were identified, and a number of motifs and co-innervation patterns that are overrepresented over a random target choice model were defined (Han, Y., et al. (2018) Nature 556, 51-56). By using the disclosed method, current knowledge about DA connectomic cell types was refined and the anatomical basis of information flow in the mouse dopaminergic system is provided herein.

[0154] POINTseq enables selective barcoding of cell types of interest via cell type-specific infection. POINTseq can utilize the alphavirus host cell receptors that interact with viral structural proteins and are sufficient for viral infection (Zhang, R., et al. (2018) Nature 557, 570-574; and Zimmerman, O., et al. (2023) J. Clin. Invest. 133). If one of these receptors is not normally expressed in the mouse brain, it could be overexpressed specifically in neurons of interest using, e.g., AAV and standard recombinase-based expression gating systems, rendering those cells selectively susceptible to infection by the alphavirus that uses the receptor (FIG. 20C). To find such a receptor, the expression levels of known alphavirus receptors were assessed in the mouse brain using the Allen Brain Cell Atlas (Y ao, Z., et al. (2023) Nature 624. 317-332). While many receptors are expressed broadly in neurons the receptor for Chikungunya virus (CHIKV) (Zhang, R., et al. (2018) Nature 557, 570-574), Mxra8, is not expressed in neurons or non-neuronal cells in any brain region (FIG. 20D and FIG. 26A). CHIKV-Mxra8 was used as virus-receptor pair for the cell type specific infection strategy.

[0155] To avoid the need to revalidate barcoded projection tracing with CHIKV, Sindbis MAPseq virus was pseudotyped with the structural proteins of CHIKV (FIG. 20D). This approach is an improvement over MAPseq as it can change the infective properties of the viral particles to those of CHIKV. Sindbis MAPseq virus is generated by co-expressing the barcoded genomic RNA with a helper RNA encoding Sindbis structural proteins in packaging cells (Kebschull, J.M., et al. (2016). Front Neuroanat 10, 56). Building on this protocol, A helper RNA was created in which the Sindbis virus structural protein open reading frame was swapped for the CHIKV structural protein open reading frame. By expressing this new helper RNA with barcoded genomic RNA from Sindbis in packaging cells, high titers of propagation incompetent CHIKV -pseudoty ped, barcoded Sindbis virus was generated (FIGS. 20E and 25). This virus is referred to as “POINTseq virus” herein.

[0156] To validate Mxra8-dependent infection of POINTseq virus in the mouse brain, POINTseq virus was injected by itself or follow ing injection of a dummy AAV expressing tdTomato (AAV-CAG-tdT) in the primary' motor cortex (MOp) of adult C57BL6 / J mice (FIG. 20F). Given the absence of endogenous Mxra8 expression in the mouse brain, almost no cells were infected by the GFP expressing POINTseq virus. However, when the POINTseq injection was preceded with the injection of an AAV helper virus expressing Mxra8 and tdTomato (AAV-CAG-tdT-P2A-Mxra8), tdTomato positive neurons were efficiently infected by POINTseq virus (FIGS. 20G-20H). Infection is highly specific to target cells with 98.7% of infections mediated by exogenous Mxra8 (FIGS. 20J and 20K). This strategy w as further tested across a range of brain regions, including the striatum, dentate gyrus, thalamus, and amygdala, and in each case, it was found that POINTseq virus infection was dependent on exogenous Mxra8 overexpression (FIG. 26B).

[0157] Finally, it was tested whether cell type-specific Mxra8 expression would support cell type-specific POINTseq infection. Mouse MOp contains different classes of long-range projection neurons arranged according to cortical layers: intratel encephalic (IT) cells in layer 2 / 3 and 5, extratelencephalic (ET) cells in layer 5, and corticothalamic (CT) cells in layer 6 (Munoz-Castaneda, R., et al. (2021) Nature 598, 159-166). Mouse line Cux2-Cre specifically labels layer 2 / 3 IT neurons whereas line Rbp4-Cre labels layer 5 IT and ET neurons (Munoz- Castaneda, R., et al. (2021) Nature 598, 159-166). Cre-dependent AAV helper virus AAV- CAG-FLEx-tdT-P2A-Mxra8 was injected into the MOp of Cux2-Cre or Rbp4-Cre mice, followed by POINTseq virus injection (FIG. 20L). The results show' that POINTseq virus specifically infected Mxra8-positive cells in layer 2 / 3 and layer 5 of Cux2-Cre and Rbp4-Cre mice, respectively (FIGS. 20N and 200). To test POINTseq in projection defined cells, retrograde AAV encoding Cre (AAVretro-Cre; AAVrg-EFla-Cre) was injected into contralateral MOp in wild ty pe mice, and cre-dependent Mxra8 helper virus in ipsilateral MOp (FIG. 20M). After injecting POINTseq virus into ipsilateral MOp infection was observed in Mxra8-positive cells in layers 2 / 3 and 5, where contralaterally projecting IT (ITc) neurons are located (FIG. 20P). POINTseq accurately maps projections of targeted cortical cell types. Next, it was evaluated whether POINTseq produced the expected single-cell projection patterns of the targeted cell types in Cux2-Cre, Rbp4-Cre, and AAVretro-Cre experiments (FIG. 21). After helper AAV injection into the respective mouse lines (N=2 animals each), barcoded POINTseq virus was injected and 26 regions were dissected including regions in cortex, striatum, thalamus, midbrain, and hindbrain covering the main brain- wide projection targets of MOp neurons (Table 2). Barcode sequencing in these regions with a optimized protocol yielded the projections of 2,544 neurons from Cux2-Cre animals, 5,027 neurons from Rbp4- Cre animals, and 2,829 neurons from AAVretro-Cre animals. These POINTseq data were combined with previously collected MAPseq data from MOp containing 11,153 neurons to produce a projection matrix of 21,553 MOp neurons. Using agglomerative hierarchical clustering, neurons were classified into the three cardinal projection types (IT, CT, ET) and further divided IT cells by whether they project to the striatum (STR+ / -) or only to ipsi or also contralateral cortical regions (ITi / ITc), as before, ((Munoz-Castaneda, R., et al. (2021) Nature 398, 159-166) yielding a total of 6 projection types (ITc STR-, ITi STR-, ITc STR+, ITi STR+, CT, ET; FIG. 21A).

[0158] Table 2. Abbreviations of brain regions.

[0159] Ip: Ipsilateral, Con: Contralateral The results show that neurons from the MAPseq experiment belonged to the six projection types with proportions ranging from 11.8% to 24.6%. In contrast, POINTseq neurons predominantly belonged to specific subsets of projection types according to the targeted neuronal population (FIG. 21B). Specifically, in Cux2-Cre animals, which target layer 2 / 3 IT neurons, 97.0% of the neurons were classified as IT. In Rbp4-Cre animals, which target layer 5 IT and ET neurons, 96.6% of the neurons were classified as IT or ET. Consistent with previous findings that ITc STR+ neurons are more abundant than ITc STR- neurons in layer 5, (Munoz-Castaneda, R., et al. (2021) Nature 598, 159-166) Rbp4-Cre showed a higher proportion of ITc STR+ than ITc STR-. The inverse is true for layer 2 / 3 IT cells in Cux2-Cre animals. In AAVretro-Cre animals, which target neurons with projections to the contralateral MOp, 94.5% of the identified neurons were contralaterally projecting ITc types. Taken together, these results show that POINTseq is a highly specific and straightforward method for mapping single-neuron projections in defined cell populations combined with common Cre-dependent strategies.

[0160] POINTseq reveals the architecture of the midbrain dopaminergic system. DA neurons in the midbrain regions VTA and SNc are the primary source of dopamine in the brain (Chinta, S.J., and Andersen, J.K. (2005) Cell Biol. 37, 942-946; and Bissonette, G.B., and Roesch, M.R. (2016) Genes Brain Behav. 15, 62-73). They play important roles in reward processing (Lammel. S., et al (2014) Neuropharmacology 76 Pt B, 351-359; Heymann, G., et al. (2020) Neuron 105. 909-920; Howe, M.W., and Dombeck, D.A. (2016) Nature 535, 505- 510; Schultz, W„ et al. (1997) Science 275, 1593-1599; and Lerner, T.N., et al. (2015) Cell 162, 635-647), aversion (Lammel, S., et al (2014) Neuropharmacology' 76 Pt B, 351-359; Heymann, G., et al. (2020) Neuron 105, 909-920; Menegas. W., et al. (2018) Nat Neurosci 21, 1421-1430; and Lerner, T.N., et al. (2015) Cell 162, 635-647), aggression (Mahadevia, D., et al. (2021) Nat. Commun. 12, 6796; and Dai, B., et al. (2025) Nature 639, 430-437), and locomotion (Howe, M.W., and Dombeck, D.A. (2016) Nature 535, 505-510; Azcorra, M„ et al. (2023) Nat Neurosci 26, 1762-1774; Mendonca, M.D., et al. (2024) Curr. Biol. 34, 1034-1047; and Dodson. P.D., et al. (2016) Proc. Natl. Acad. Sci. U. S. A. 113, E2180-8) while serving as important circuit nodes in substance use disorders (Juarez. B.. and Han. M - H. (2016) Neuropsychopharmacology 41, 2424-2446; and Volkow, N D., et al. (2007) Arch. Neurol. 64, 1575-1579), psychiatric diseases (Chen, A.P.F., et al. (2021) Biomedicines 9, 641,' and Davis, K.L., et al. (1991) Am. J. Psychiatry' 148, 1474-1486), and neurodegenerative diseases (Zhou. Z.D., et al. (2023) Transl. Neurodegener. 12, 44; Dauer, W., and Przedborski, S. (2003) Neuron 39, 889-909; D’Amelio, M., et al. (2018) Pharmacol. Res. 130, 414-419; and Nobili, A., et al. (2017) Nat. Commun. 8. 14727). Modulating those many functions, midbrain DA neurons project broadly across the brain. Recent single-cell transcriptomic and bulk anatomical studies have revealed large heterogeneity among dopaminergic neurons that are reflected in different dopaminergic functions (Garritsen, O., et al. (2023) Nat Rev Neurosci 24, 134-152; Azcorra, M., et al. (2023) Nat Neurosci 26, 1762- 1774; Poulin, J.-F.. et al. (2018) Nat Neurosci 21, 1260-1271; and Poulin, J.-F., et al. (2020) Trends Neurosci. 43. 155-169). How these distinct functional circuits are formed from different populations of branching DA neurons and what subcircuits target the same target regions, however, remains poorly understood.

[0161] To address this gap, the brain- wide single cell projections of midbrain DA neurons were mapped. DA and non-DA cells are intermingled in the VTA and SNc, (Morales, M., and Margolis, E.B. (2017) Nat. Rev. Neurosci. 18, 73-85) preventing the use of MAPseq to map DA neurons only. Therefore, POINTseq was applied in the DA specific Cre driver mouse line DAT-Cre. Cre-dependent Mxra8 helper AAV was injected into the VTA and SNc of DAT- Cre mice, followed by barcoded POINTseq virus (diversity > ***; FIG. 22A). The results show that POINTseq virus specifically infected tdT- and TH-positive cells in the VTA and SNc (FIGS. 22B and 27).

[0162] In three animals, the two injected source regions VTA and SNc, and a total of 81 target regions ipsi and contralateral to the injection site were dissected. These regions include the major targets of VTA and SNc DA neurons, including medial prefrontal cortex (rnPFC), nucleus accumbens (NAc), lateral septal nucleus (LS), olfactory tubercle (OT), ventral pallidum (VP), globus pallidus (GP), caudate putamen (CP), and lateral and basolateral amygdala (AMY), as well as regions such as olfactory' bulb (OB), entorhinal cortex (EC) and lateral habenula (LHb) (FIGS. 22C, Table 2) (Beier, K.T., et al. (2015) Cell 162, 622-634; Mahadevia, D., et al. (2021) Nat. Commun. 12, 6796; Lammel, S., et al. (2008) Neuron 57, 760-773; Aransay, A., et al. (2015) Front Neuroanat 9, 59; Fields, H.L., et al. (2007) Annu. Rev. Neurosci. 30, 289-316; Menegas, W., et al. (2015) Elife 4; Zhang, Z., et al. (2017) Elife 6; Lee, J.Y., et al. (2021) Nature 598, 321-326; and Stamatakis, A.M., et al. (2013) Neuron 80. 1039-1053).

[0163] Accounting for the distinct bulk projections and functional differences between subregions of NAc, (Heymann, G., et al. (2020) Neuron 105, 909-920.e5; Beier, K.T., et al. (2015) Cell 162, 622-634; Lammel, S„ et al. (2008) Neuron 57, 760-773; Poulin, J.-F., et al. (2018) Nat Neurosci 21, 1260-1271; and Tolve, M„ et al. (2021) Cell Rep. 36. 109697) the NAc was further subdivided into NAc lateral core (NAcLatC), NAc medial core (NAcMedC), NAc lateral shell (NAcLatS), NAc medial shell (NAcMedS), and posterior NAc (NAc-p). Similarly, based on the distinct bulk projections and functional roles of VTA / SNc DA neurons (Beier, K.T., et al. (2015) Cell 162, 622-634; Howe, M.W., and Dombeck, D.A. (2016) Nature 535, 505-510; Lerner, T.N., et al. (2015) Cell 162, 635-647; Poulin, J.-F., et al. (2018) Nat Neurosci 21, 1260-1271; Menegas, W„ et al. (2015) Elife 4 and Tolve, M„ et al. (2021) Cell Rep. 36, 109697) and distinct cortical inputs in the CP (Foster, N.N., et al. (2021) Nature 598. 188-194; Hintiryan. H., et al. (2016) Nat. Neurosci. 19. 1100—1114; and Hunnicutt, B.J., et al. (2016) Elife 5), the CP was divided into four subdomains: lateral CP (LatCP), dorsomedial CP (dMedCP), ventromedial CP (vMedCP), and CP tail (CPt). These domains were further split across the rostral-caudal axis where appropriate [CPr(rostral), Cpi (intermediate), and CPc (caudal)], resulting in the following 9 dissected CP subregions: LatCPr, LatCPi, LatCPc, dMedCPr, dMedCPi, dMedCPc, vMedCPr, vMedCPi, and CPt. Each subregion is further segmented from rostral to caudal with 300 pm resolution as indicated by alphabetical labeling (e g. LatCPc-a). To ensure maximum accuracy in the dissection, tissue handling and staining was optimized (Chen. Y., et al. (2022) Cell 185, 4117-4134; and Huang, L., et al. (2020) Cell 183, 2040), resulting in a protocol that uses laser capture microdissection of Nissl stained brain sections comparable in quality to that achieved with conventional Nissl staining (FIG. 22C). Then, the barcodes were sequenced in the dissected regions, and reconstructed projection matrices (Kebschull, J.M., et al. (2016) Neuron 91. 975-987). Importantly, each barcoded neuron was assigned to reside either in the VTA or SNc based on barcode abundance (Huang, L., et al. (2020) Cell 183, 2040) (STAR methods) and thus map VTA and SNc neurons in parallel in the same animal.

[0164] Midbrain dopaminergic neurons are diverse and form many connectomic cell types. The experiments yielded a total of 3,813 high quality DA neurons, with 3,563 located in the VTA and 250 in the SNc (FIG. 22D). DA neurons branch widely, with 82% of neurons innervating more than 1 and up to 17 regions (FIG. 22E). To understand the overall structure of the DA neuron projections, hierarchical clustering of the dataset was performed. To be conservative, the coarsest clustering resolution that allowed to distinguish known distinct populations (Heymann, G., et al. (2020) Neuron 105, 909-920; Beier. K.T.. et al. (2015) Cell 162, 622-634; Menegas, W., et al. (2018) Nat Neurosci 21, 1421-1430; Mahadevia, D., et al. (2021) Nat. Commun. 12, 6796; Lerner, T.N., et al. (2015) Cell 162, 635-647; Lammel, S„ et al. (2008) Neuron 57, 760-773; Poulin, J.-F., et al. (2018) Nat Neurosci 21, 1260-1271; Menegas, W„ et al. (2015) Elife 4; Tolve, M.. et al. (2021) Cell Rep. 36, 109697; Wu, J., et al. (2019) Cell Rep. 28, 1167-1181; and Kramer, D.J., et al. (2018) eNeuro 5) was chosen, such as preferentially CPt-projecting neurons, and then collapsed overclustered populations based on significance of that node, in which the null model assumes that all values within a node are drawn from a Gaussian distribution and the significance of the observed linkage value of the node was assessed by comparing it to simulated linkage values under the null hypothesis (Kimes, P.K., et al. (2017) Biometrics 73, 811-821) (STAR methods). In total, 34 connectomic cell types were identified (FIGS. 22D and 22F), far exceeding the number of known molecular DA neuron subtypes (Garritsen. O.. et al. (2023) Nat Rev Neurosci 24. 134-152; and Poulin, J.-F., et al. (2020) Trends Neurosci. 43, 155-169).

[0165] 96.7% of DA neurons project predominantly ipsilaterally (Aransay, A., et al. (2015) Front Neuroanat 9, 59; and Jaeger. C.B.. et al. (1983) J. Comp. Neurol. 218, 74-90), and form 32 out of the 34 total connectomic cell types. These types can be roughly grouped by their strongest projection targets (FIGS. 22F and 22G). Several types preferentially target NAc (types 1-8, 30), OT (types 6-12), and CP (types 20-29, 31-34), while other regions are the strongest target for one or two connectomic cell types, such as mPFC (ty pe 11), LS (types 12-13), VP (type 16), BST (type 17), AMY (type 18), GP (type 19). Interestingly, whereas many cell types display some innervation of NAc, connectomic cell types that have their strongest projects in LS, GP, CPt, or dMedCP (types 13, 19, 20, 32-34) rarely project to NAc. CP-projecting neurons show preferences for different CP subregions and positions along the rostral-caudal axis, reflecting the functional organization of this region. Notably, type 11 neurons uniquely project to cortical regions with substantial projections to multiple regions including OT, VP, and AMY, presenting hub-like neurons that project to cortical and subcortical regions in parallel. 3.3% of the traced neurons primarily project to contralateral regions. They form 2 connectomic types (types 14 and 15) out of these type 14 preferentially projects to the contralateral NAcMedS.

[0166] Molecular DA cell types are distributed somewhat continuously but spatially biased across the VTA and SNc (Garritsen, O., et al. (2023) Nat Rev Neurosci 24, 134-152; and Poulin, J.-F., et al. (2020) Trends Neurosci. 43, 155-169). Therefore, the contribution of VTA and SNc to connectomic cell types were investigated and big differences in projection patterns between VTA and SNc neurons were observed (FIGS. 22G, 22H, and 28). For example, 92.4% of SNc neurons belong to types preferentially projecting to the GP / CP (types 19-28, 31-34), whereas only 15.3% of VTA neurons belong to these types. In contrast, 56.3% of VTA neurons belong to types that have substantial projections to NAc (types 1-8, 29, 30), whereas 4.0% of SNc neurons belong to these types. The results show that VTA DA neurons tend to project rostral and ventromedial areas of CP while SNc DA neurons tend to project across all CP areas (Howe, M.W., and Dombeck, D.A. (2016) Nature 535, 505-510; Chen, A.P.F., et al. (2021) Biomedicines 9. 647; Poulin, J.-F., et al. (2018) Nat Neurosci 27, 1260-1271; and Farassat, N., et al. (2019) Elife S). SNc neurons are abundant in all CP- projecting types (types 20-28, 31-34), except for type 28 which preferentially projecting to LatCPr-a. In contrast, VTA neurons are abundant in type 28 and types preferentially projecting to vMedCPr or vMedCPi (types 24, 31).

[0167] Taken together, the clustering results align well with reported bulk projection patterns of DA neurons, but add important granularity. In particular, it was found that most previously investigated projection-defined populations likely consist of multiple connectomic cell types (FIG. 29), providing additional functional specificity' of DA neuronal types. For example, while distinct populations that project to the NAcLatS or NAcC and to the CP were reported (Lammel, S., et al. (2008) Neuron 57, 760-773), the methods used herein can further classify these into multiple connectomic types based on their projection patterns: NAcLatS- or NAcC- projecting neurons and CP-projecting neurons are divided into distinct types, types 2-4, 6, 29, and types 18-27, 32-34, respectively, according to their projection patterns involving other regions or across CP subregions. Similarly, a population previously described to project to the NAcC, MedCP, VP, and LS given NAcMed input (Beier, K.T., et al. (2015). Cell 762, 622-634), can be further separated into multiple types (types 1, 2, 8), depending on whether they preferentially project to the NAcMedS or additionally innervate other regions such as the VP and OT.

[0168] Non-random co-innervation of target regions by individual VTA dopaminergic neurons demonstrates functional importance of information sharing. Neurons that coinnervate different regions, to a first approximation, transduce shared signals to these target regions. Understanding the nature of such co-innervations will therefore yield important insights into information routing and shared functions of these regions. Co-innervations, or lack thereof, might be established by a simple binomial model during development where neurons choose their target regions statistically independently by a random draw. Alternatively, co-innervations could be hard-coded developmentally by an active process (Han. Y.. et al. (2018) Nature 556. 51-56). Information is shared between co-innervated regions and segregated between not co-innervated regions under either model. However, deviations from the binomial model show that sharing or not sharing information is functionally significant as energy' was spent to set them up. Although the functions of DA signaling in individual regions have been extensively studied, the importance of shared or segregated DA signaling across regions remains poorly understood. Therefore, the statistics underlying DA neuronal projection patterns was assessed (FIG. 23).

[0169] The analysis included VTA DA neurons, given the restricted projections of SNc DA neurons to GP and CP. The analysis also included a subset of target regions (mPFC, EC, NAc, OT, LS, VP, AMY, CP*: CP excluding CPt, and CPt), where DA function has been studied and which are targeted by at least 10 neurons (FIG. 23 A). CP was subdivided into CP* and CP* considering the specific anatomy and functions of the CPt compared to the other regions (Menegas, W., et al. (2018) Nat Neurosci 21, 1421-1430; Menegas, W., et al. (2015) Elife 4; and Tsutsui-Kimura, I., et al. (2025) Nat. Neurosci. 28, 795-810). While CPt- projecting neurons are primarily arise from the SNc (Menegas, W., et al. (2018) Nat Neurosci 21, 1421-1430), they are also found in the VTA (FIGS. 22D and 22F) (Menegas, W., et al. (2015) Elife ).

[0170] To describe the frequency and relationship of co-innervations between pairs of target regions, three complementary7measures were calculated: First, the conditional projection probability P(B| A) between pairs of regions was calculated to assess the likelihood of projecting to a region B that a neuron projects to another region A (FIG. 23B). Second, the Dice score (Dice, was calculated to assess the overlap between projections to the two regions. The Dice score ranges from 0 to 1, where 0 indicates complete exclusivity7and 1 indicates complete overlap. Finally, the rarity score, (FIGS. 23C and 23E; STAR Methods) was introduced, defined as the inverse of the Dice score estimated under the binomial model. The rarity score thus represents the abundance of projections belonging to the co-innervated regions relative to the total neuron population. Similar and lower rarity7scores across regions indicate a more even distribution of projections, whereas varying and higher rarity scores indicate inconsistent proportions of projections to the regions. By definition, the product of the observed Dice score and the rarity score indicates whether a given co-innervation is over- or under-represented relative to an independent model of target choice (FIGS. 23C and 23F; STAR Methods), thus indicating if an active process was set up this co-innervation or lack thereof during development or not. Results are consistent across individual animals (FIG. 30).

[0171] The analyses revealed a close association between NAc, OT, and VP, regions in which DA signaling is involved in encoding reward and reward prediction error (Zhang, Z., et al. (2017) Elife 6; Richard, J.M., et al. (2018). Elife 7; Ottenheimer, D.J., et al. (2020) Sci. Adv. 6; Ottenheimer, D.J., et al. (2020) Nat. Neurosci. 23, 1267-1276; Murata, K.. et al. (2015) J. Neurosci. 35, 10581-10599; Ikemoto, S. (2007) Brain Res. Rev. 56, 27-78; Heymann, G., el al. (2020) Neuron 105, 909-920; Schultz, W., et al. (1997) Science 275. 1593-1599; Lammel, S„ et al. (2011) Neuron 70, 855-862; and Tobler, P.N., et al. (2005) Science 307, 1642-1645). Specifically, NAc is likely to be innervated by neurons that projection to OT and VP (FIG. 23B). Additionally, NAc has high a high Dice score with OT and VP (FIG. 23C), and these co-innervations are over-represented over what would be expected from the binomial model (FIG. 23F).

[0172] The hub-like cell type 11 (FIG. 23F) projects to a number of regions involved in associative and aversive learning (Menegas, W., et al. (2018) Nat Neurosci 21, 1421-1430; Zhang, Z., et al. (2017) Elife 6; Lee, J.Y., et al. (2021) Nature 598, 321-326; Tsutsui- Kimura, I., et al. (2025) Nat. Neurosci. 28, 795-810; Richard, J.M.. et al. (2018) Elife 7; Ottenheimer, et al. (2020) Sci. Adv. 6; Ottenheimer, D.J., et al. (2020) Nat. Neurosci. 23, 1267-1276; Murata, K„ et al. (2015) J. Neurosci. 35, 10581-10599; Ikemoto, S. (2007) Brain Res. Rev. 56, 27-78; Jun, H„ et al. (2024) Nature 633, 864-871; Vander Weele, C M., et al. (2018) Nature 563. 397-401; Kienast, T., et al. (2008) Nat. Neurosci. 11. 1381-1382; Lutas, A., et al. (2019) Nat. Neurosci. 22, 1820-1833; Tang, W., et al. (2020) J. Neurosci. 40, 3969- 3980; Duvarci, S. (2024) Trends Neurosci. 47, 1014-1027). Analyzing the co-innervation statistics of these regions, it is noted that the high and overrepresented co-innervation of mPFC and EC by VTA DA neurons (FIGS. 23C and 23F). This bifurcation nicely matches to the reported co-dependence of mPFC and EC in associative learning (Jun, H., et al. (2024) Nature 633, 864-871 ). Given that a neurons projects to the mPFC or EC, the OT, VP, and AMY are highly likely to be innervated (FIG. 23B), and co-innervations of the mPFC or EC with the OT, VP, AMY, or CPt are over-represented (FIG. 23F) over the random model of target choice, indicating functionally specific DA subnetworks involving cortical regions.

[0173] In addition to co-innervations that are more abundant than expected, it was also observed co-innervations that happen significantly less frequently than expected under a binomial model, indicating that distinct functions are carried out by neurons innervating these regions (FIG. 23F). The co-innervation between NAc and CPt is under-represented, consistent with the fact that DA neurons projecting to the NAc are primarily located in the medial VTA, whereas those projecting to the CPt are laterally located (Menegas, W , et al. (2015) Elife 4). NAc-LS co-innervation is also under-represented (Mahadevia, D., et al. (2021) Nat. Commun. 12, 6796). Interestingly, many other co-innervations that involve LS are also under-represented, such as LS-VP, LS-CP*, and LS-CPt (FIG. 23F). In contrast. LS exhibits over-represented co-innervations with mPFC and EC, demonstrating a role of shared DA signalling between LS and cortical regions which is segregated from other regions. Finally, the CP*-OT co-innervation is under-represented. As CP* receives inputs related to auditory, visual, and sensory motor functions (Foster, N.N., et al. (2021) Nature 598, 188- 194; Hintiryan, H., et al. (2016) Nat. Neurosci. 19, 1100-1114; and Hunnicutt, B.J., et al. (2016) Elife 5), and OT deals with olfactory information, this underrepresentation shows that DA signalling can modulate these modalities separately.

[0174] VTA dopaminergic neurons form non-random projection motifs. To obtain a comprehensive view of the projections of individual DA neurons, over- and under- represented projection motifs, defined as innervation patterns across the possible combinations of target regions (FIGS. 23G and 23H) were assessed (Han, Y., et al. (2018) Nature 556, 51-56). For this analysis, the mPFC and EC were grouped into '’cortex" (CTX). based on their highly similar co-innervation patterns, and included motifs for which the observed or expected number of contributing neurons exceeded 5. 21 motifs that w ere significantly over- or under-represented relative to the binomial model were identified, providing the substrate for parallel dopaminergic signal transduction pathways across multiple regions.

[0175] NAc-OT, OT-VP, and NAc-OT-VP, motifs (motifs 3, 7, 12) are over-represented, consistent with over-represented pair-wise co-innervations betw een the three regions. Strikingly, motifs containing NAc and CP* with or without OT and VP (motif 5, 14, 18) are also over-represented and abundant, although co-innervations between CP* and NAc, OT. and VP is not over-represented. This indicates that innervations to NAc and CP* are specifically enriched when not involving other regions, except for OT and VP. Furthermore, motifs containing CP* and OT or VP are over-represented when NAc is also included (motif 14, 17, 18, 21). but under-represented without NAc (motifs 8. 9), demonstrating that innervation to NAc can facilitate co-innervations between OT and CP*, and betw een VP and CP*.

[0176] Motifs containing CTX-NAc-OT-VP-AMY (motifs 19-21) are highly over- represented, consistent with the over-represented co-innervations between pairs of CTX, OT, VP, and AMY (FIG. 23F). and the high probability of NAc projection given projections to other regions (FIG. 23B). The role of DA signaling in these regions converges to associative learning of aversive stimuli (Zhang, Z., et al. (2017) Elife 6; Lee, J.Y., et al. (2021) Nature 598, 321-326; Richard, J.M., et al. (2018) Elife 7: Ottenheimer, D.J., et al. (2020) Sci. Adv. 6; Ottenheimer, D.J.. et al. (2020) Nat. Neurosci. 23, 1267-1276; Murata, K., et al. (2015) J. Neurosci. 35, 10581-10599; Ikemoto, S. (2007) Brain Res. Rev. 56, 27-78; Jun, H., et al. (2024) Nature 633. 864-871; Vander Weele, C.M., et al. (2018) Nature 563, 397-401; Kienast, T., et al. (2008) Nat. Neurosci. 11, 1381-1382; Lutas, A., et al. (2019) Nat. Neurosci. 22, 1820-1833; Tang, W., et al. (2020) J. Neurosci. 40, 3969-3980; and Duvarci, S. (2024) Trends Neurosci. 47, 1014-1027), in which these motifs may play roles. Furthermore, motif 21 contains CPt, which is also implicated in the learning of aversive or threatening stimuli (Menegas, W., et al. (2018) Nat Neurosci 21, 1421-1430; Tsutsui- Kimura, I., et al. (2025) Nat. Neurosci. 28, 795-810; and Duvarci, S. (2024) Trends Neurosci. 47, 1014-1027). Motifs 10 and 15 containing CPt and AMY without CTX, NAc, OT, and VP, demonstrating more specific roles related to those aversive or threatening stimuli.

[0177] Additionally. CTX-LS motif (motif 2) which does not involve other regions and is over- represented, consistent with the over-represented co-innervations of the LS with cortical regions and the under-represented co-innervations with regions such as the NAc, VP, and CP*.

[0178] It was also observed under-represented motifs (motifs 1, 4, 6, 8, 9, and 13). Notably, many of these motifs occur between regions of the NAc, OT, and VP and regions of the AMY, CP*, and CPt (motifs 4, 6, 8, 9, and 13), whereas over-represented motifs (motifs 12 and 15) are observed within each group separately. This shows that overall signals may be largely separately shared within the NAc, OT, and VP group and the AMY, CP*, and CPt group. Interestingly, when the cells also project to CTX, over-represented motifs between these two groups (motifs 17 and 19-21) were observed, demonstrating that DA projections to the CTX can facilitate co-innervations between the NAc, OT, and VP group and the AMY, CP*, and CPt group.

[0179] To complement the analysis of binarized projection motifs, the projection strengths of individual neurons in each motif was investigated (FIG. 31). Neurons from under-represented motifs tended to strongly project to a single region, whereas neurons from over-represented motifs projected more evenly across multiple regions. These data show that shared signal transduction within over-represented motifs is relatively balanced, while signal transduction within under-represented motifs is more biased towards single regions.

[0180] SNc dopaminergic neurons have broader projections to GP and CP than VTA neurons. CP is one of the major projection targets of midbrain DA neurons. CP projecting neurons are primarily found in the SNc but are also present in the VTA (Beier, K.T., et al. (2015) Cell 162, 622-634; Howe, M.W., and Dombeck, D.A. (2016) Nature 535, 505-510; Menegas, W., et al. (2015) Elife 4; and Farassat, N.. et al. (2019) Elife 8). CP targeting DA neurons have been shown to innervate different CP subregions and to be selectively processing rewarding stimuli, aversive stimuli, or locomotion (Beier, K.T., et al. (2015) Cell 162, 622-634; Howe, M.W.. and Dombeck, D A. (2016) Nature 535. 505-510: Lerner, T.N., et al. (2015) Cell 162, 635-647; and Poulin, J.-F., et al. (2018) Nat Neurosci 21, 1260-1271). The large diversity of DA neuron connectomic types revealed by POINTseq (FIG. 22F), however, shows additional structure in CP innervating DA neurons. Specifically, it is unknown how many subregions individual DA neurons innervate and whether there is structure in the co-innervation patterns of different subregions. Moreover, it is unclear how these statistics differ between VTA and SNc neurons. Although SNc DA neurons are known for their wide and dense axonal arborizations compared to VTA neurons in general, their projection patterns within CP have not been compared (Pacelli, C., et al. (2015) Curr. Biol. 25, 2349-2360; Matsuda, W., et al. (2009) J. Neurosci. 29, 444-453; and Giguere, N., et al. (2019) PLoS Genet. 15, e!008352).

[0181] To answer these questions, projection patterns of VTA and SNc DA neurons were compared across 9 functionally distinct CP subregions (LatCPr, LatCPi, LatCPc, dMedCPr, dMedCPi, dMedCPc, vMedCPr. vMedCPi, CPt) (Foster, N.N., et al. (2021) Nature 598, 188- 194; Hintiryan, H., et al. (2016) Nat. Neurosci. 19, 1100-1114; and Hunnicutt, B.J., et al. (2016) Elife 5). Given the close coordination of CP and GP in regulating locomotion, and that GP and CP projecting DA neurons share a similar spatial distribution in VTA and SNc (Menegas, W., et al. (2015) Elife 4), GP was also included in the analysis. Neurons were selected whose projections to CP and GP account for at least 10% of their total projections (FIG. 24A). The results from the analysis of connectomic types in VTA and SNc (FIGS. 22F and 28), show that VTA and SNc neurons differently contributed to connectomic types with distinct projection patterns. For example, VTA neurons rarely belong to type 21, which primarily projects to LatCPc and dMedCPc. whereas SNc neurons were rarely found in types 28, 29, and 30, which mainly project to LatCPr or vMedCPr. Additionally, many types of SNc neurons project to the GP, whereas VTA neurons show GP projections only in a few specific types. Overall, many VTA DA neurons preferentially project to specific subregions, particularly to LatCPr and vMedCPr, while SNc DA neurons have broader and more even projections across CP subregions.

[0182] Consistently, the number of target regions per VTA and SNc neuron is strikingly different (FIG. 24B and 6C). 22% of VTA neurons vs 4% of SNc neurons project exclusively to 1 target, while 62% of SNc neurons vs 32% of VTA neurons project to more than 3 targets. The distribution of those proportions per mouse was significantly different between VTA and SNc neurons (FIG. 24C). To examine if these differences are specific to certain target regions, neurons were grouped based on their maximum projection target. It was found that SNc neurons project to more targets than VTA neurons when their main target project is LatCP, dMedCP, vMedCP, CPr, CPi, or CPc. The difference was biggest in neurons with maximum projection to LatCP, vMedCP, and CPr. In contrast, no significant difference was observed in target numbers for neurons that have their strongest projection to GP or CPt (FIG. 24D).

[0183] As the analysis of target regions is based on binarized projection patterns , it was next assessed how their projection strengths varied across target regions (Figure 32). To do this, projection densities were calculated (projection strength divided by region size; FIG. 32A). The standard deviation of projection densities for each neuron were analyzed. SNc neurons had smaller standard deviations than VTA neurons (FIG. 32B), indicating that individual SNc neurons project more evenly to GP and CP subregions. When the projection densities were plotted according to the main target regions, SNc neurons gradually decreased around the main target, in particular in the caudal direction, whereas of VTA neurons dropped sharply, especially when the main target was in the CPr subregions (FIG. 32C). Importantly, the results were not driven by a larger number of dedicated projection neurons in the VTA, as they are preserved when the neurons were required to innervate at least two regions (FIG. 33).

[0184] SNc dopaminergic neurons have more random projections to GP and CP than VTA neurons. To understand the projection statistics of VTA and SNc neurons, the frequency of pairwise co-innervations were assessed using the Dice and rarity scores and assessed over- or under-representation relative to the independent target choice model, combined with rarity score as in FIGS. 23D-23F (FIGS. 24E-24J). SNc neurons exhibit higher co-innervations than VTA neurons especially between GP and CP, within LatCP, and between LatCP and MedCP (FIGS. 24E and 24H). For example, while VTA neurons show high co-innervation of GP only with spatially close regions such as LatCPc and vMedCPi, SNc neurons exhibit high coinnervation also with more distal regions including LatCPi, dMedCPi, and dMedCPc. Similarly, SNc neurons are more likely to co-innervate LatCP and dMedCP as well as LatCP and vMedCP than VTA neurons. Within LatCP. the lower co-innervations of VTA neurons can be attributed to their preferential proj ections to LatCPr compared to LatCPi and LatCPc (FIG. 23 A). In contrast, within MedCP the co-innervation rates are comparable between SNc and VTA.

[0185] The rarity scores for SNc neurons are more similar and lower than the scores for VTA neurons, indicating that SNc neurons project more evenly to GP and CP subregions than VTA neurons (FIGS. 24F and 241). In particular, CPt and other caudal areas including LatCPi, LatCPc, and dMedCPc are rarely targeted by VTA neurons, consistent with their preferential projections to the rostral regions (FIG. 24 A).

[0186] Intriguingly, most co-innervations by SNc neurons agree with the random binomial model of projections, whereas co-innervations by VTA neurons are often significantly different from what would be expected (FIGS. 24G and 24J), driven by both the Dice and rarity scores. Specifically, VTA neurons exhibit many over-represented co-innervations within MedCP or involving caudal regions, while co-innervations between rostral regions and caudal regions tend to be more under-represented. Notably, DA signaling in MedCP (Seiler, J.L., et al. (2022) Curr. Biol. 32, 1175-1188) and VTA DA neuron activity (Pascoli, V., et al. (2015) Neuron 88, 1054-1066) promote compulsive reward seeking, demonstrating that the strongly over-represented co-innervations of VTA DA neurons within MedCP may play important role in this behavior.

[0187] To test whether the different projection patterns of VTA and SNc neurons in GP and CP are driven by VTA neurons projecting other third regions, the analysis was repeated focusing on neurons whose projections are confined exclusively to the GP and CP (FIG. 34). Most of the neurons eliminated by this filter are VTA neurons that project to the CPr within GP / CP and NAc outside of it (FIG. 22F). In the CP / GP exclusive set, the number of targets per neuron and abundance of co-innervations remained lower for VTA than SNc neurons, albeit to a lesser degree than in the previous analysis. This consistency indicates fundamental differences of projection patters of VTA and SNc neurons in GP and CP regardless of whether they innervate other regions or not (FIGS. 34B, 34C, 34F, and 34G). However, the rarity score of VTA neuron co-innervations decreased for those tested co-innervations involving CPt (FIGS. 34D and 34H), such that most co-innervations of VTA and SNc neurons are now explained by the binomial model of target choice (FIGS. 34E and 341). This demonstrates two distinct phases in the development of VTA neuron projections, where during the initial phase the projections are structured by non-random, selective targeting, but when the projections are restricted in GP and CP at the initial phase, the development follows more random manner dunng the second phase.

[0188] Taken together, individual SNc neurons innervate more regions in GP and CP more evenly than VTA neurons, and their projections are well described by the binomial model of target choice. In contrast, VTA neurons are biased to innervate rostral regions of the CP, and co-innervations deviate from the binomial model with over-represented co-innervations within MedCP and between regions involving caudal regions, and under-represented co- innervations between rostral and caudal regions. These results remain robust when subsampling VTA neurons to match SNc neuron numbers (FIG. 35) and similar trends are observed across individual animals (FIG. 36).

[0189] Co-innervation analyses at maximum region resolution reveal the detailed architecture of shared and segregated dopaminergic signal transduction. Finally, the brainwide co-innervation patterns of VTA and SNc DA neurons for the dissected regions at maximum resolution were analyzed (FIGS. 37 and 38), including the sub-divisions of NAc and CP, and minor target regions such lateral habenula (LHb) and contralateral regions. For each co-innervation, the Dice score was measured and it was tested whether the frequency of these co-innervations is explained by an independent model of target choice. To ensure robustness, regions targeted by at least 10 neurons were included, resulting in a dataset that included the 41 ipsilateral regions and 13 contralateral regions for VTA projections and 32 ipsilateral regions for SNc projections.

[0190] Ipsi- and contralateral regions are rarely co-innervated by VTA neurons, as expected from the exclusive contralaterally projecting types 14 and 15 (FIG. 22F). Co-innervations among contralateral regions are high and more abundant than expected from the binomial model. Among ipsilateral regions, there is a tendency that co-innervations of OB, cortical regions, or AMY over-represented over the binomial model; NAc and limbic regions are highly co-innervated; and GP and CP subregions are highly co-innervated and over- represented. Also, consistent with FIG. 22F. LatCPr and vMedCPr exhibit high coinnervations with NAc. Within NAc, co-innervations of NAcLatS-NAcLatC and NAcMedS- NAcMedC were abundant and over-represented demonstrating shared signal transductions according to lateral and medial divisions. Additionally, the NAcLatS-NAcMedS Dice score is high, but not over-represented, while NAcLatC-NAcMedC is relatively low but over- represented, showing important shared signal transduction over NAc subregions. NAc-p generally showed similar patterns to NAcLatS. Within GP and CP, SNc neuron coinnervations tended to be higher across subregions, and were less over-represented compared to VTA neurons, consistent with FIG. 24.

[0191] Described herein are methods (e.g., POINTseq). which allow massively multiplexed single neuron projection mapping of genetically defined cell populations, and applied it to understand the brain-wide projectional architecture of VTA and SNc dopaminergic neurons. POINTseq is a user-friendly technology that exploits MAPseq and a specific alphavirus- cellular receptor pairing to limit Sindbis virus based barcode delivery to cells of interest. It thus introduces barcoded connectomics into the established neural circuit dissection toolkit that defines cell types of interest (Luo. L., et al. (2018) Neuron 98, 865; Luo, L., et al. (2008) Neuron 57. 634-660). POINTseq was validated in motor cortex using both cre-driver mouse lines and projection-specific Cre expression. Then mapped the brain-wide projections 3,813 individual DA neurons in the VTA and SNc. We identified a large diversity of connectomic cell ty pes that align with expected bulk projection patterns, but also highlight wi despread heterogeneity within previously defined DA subpopulations. We show that co-innervation patterns are often not explained by independent target choice, suggesting functional specificity and importance of these projection motifs. Furthermore we find that SNc neurons broadcast their information more broadly and indiscriminately across the GP and CP than VTA neurons.

[0192] Simple cell type specificity for barcoded connectomics . Barcoded connectomics methods, such as MAPseq, convert neuroanatomy into a format amenable to high-throughput sequencing yielding high multiplexing and single-cell resolution. However, cell type specific mapping was not possible so far. limiting the uptake of these methods. Barcoded connectomics can directly bridge connectomics and transcriptomics for individual cells (Chen, X., et al. (2019) Cell 179, 772-786; and Klingler, E„ et al. (2021) Nature 599, 453- 457) thus circumventing the need for cell type specific tracing. While those combined approaches are powerful, they rely on single cell sequencing or specialized in situ sequencing equipment. As a result, these tools have a high barrier of entry and are costly on a per neuron and animal level and are relatively slow. Their use is currently restricted to expert labs, with limited application to behavioral or disease studies that require a large number of animals to overcome animal variability.

[0193] A method described herein (e.g., POINTseq) was developed to permit mapping circuits with cell type specificity while maintaining the single cell resolution and mapping efficiency and requiring no specialized equipment. POINTseq specificity is fully determined by Mxra8 expression, which can be controlled using any established tools for the genetic dissection of neuronal circuits (Luo, L., et al. (2018) Neuron 98, 865; and Luo, L., et al. (2008) Neuron 57, 634-660), including but not limited to Cre-driver lines (Whitesell, J.D.. et al. (2021) Neuron 109. 545-559; Munoz-Castaneda. R.. et al. (2021) Nature 598, 159-166; Harris, J. A., et al. (2014) Front Neural Circuits 8, 76; and Zhang, F., et al. (2010). Nat Protoc 5, 439-456). (FIGS. 20L, 20N, and 200), viral delivery of Cre in the retrograde (Lavoie, A., and Liu, B.-H. (2020) Front. Mol. Neurosci. 13, 9; Peng, S., et al. (2024) Cell Rep 43, 114277; and Gradinaru, V., et al. (2010) Cell 141. 154-165) (FIGS. 20M and 20P) or anterograde direction (Lo, L., and Anderson, D.J. (2011) Neuron 72, 938-950; and Zingg, B., et al. (2017) Neuron 93, 33-47), cell type specific promoters (Walther, W., and Stein, U. (1996) J. Mol. Med. 74, 379-392; and Radhiyanti, P.T.. et al. (2021) Neurosci. Lett. 756, 135956), intersectional strategies (Hughes, A C., et al. (2024) Nat. Neurosci. 27, 1400-1410; Fenno, L.E., et al. (2014) Nat. Methods 77, 763-772; and Jeong, M., et al. (2024) Neuron 772, 56-72), and activity dependent labeling (Guenthner, C.J., et al. (2013) Neuron 78, 773- 784; Allen. W.E., et al. (2017) Science 357, 1149-1155; DeNardo, L„ and Luo, L. (2017) Curr Opin Neurobiol 45. 121-129; Lee, D., et al. (2017) Nat Biotechnol 35, 858-863; Wang, W ., et al. (2017) Nat Biotechnol 35, 864-871; and Barykina, N.V., et al. (2022) Prog Neurobiol 216, 102290). Furthermore, as cells of interest are defined prior to barcoding in POINTseq. virally induced changes to the transcriptome do not affect specificity or decrease cellular yield (Uyaniker. S., et al. (2019) Front. Cell. Neurosci. 13, 362).

[0194] Alphavirus pseudotyping to control Sindbis virus tropism. Pseudotyping of viruses is a well-established method to change viral tropism. In neuroscience, however, its application has so far primarily focused on Lenti (Gutierrez-Guerrero, A., et al. (2020) Viruses 72, 1016; and Joglekar. A.V., and Sandoval, S. (2017) Hum. Gene Ther. Methods 28, 291-301) and Rabies virus (Wickersham, I.R., et al. (2007) Neuron 53, 639-647; and Jin, L., et al. (2023) Cell Rep. Methods 3, 100644). The work described herein now expands this toolkit to propagation-incompetent Sindbis virus (Kebschull, J.M., et al. (2016) Front Neuroanat 10, 56). While POINTseq exploits the lack of CHIKV receptor expression in the mouse brain to drive cell type specific barcoding, similar strategies using CHIKV or other alphaviruses will also allow addressing difficulties in infecting cells of interest more broadly in mice and other species.

[0195] Single-cell projection mapping defines diverse and overlapping dopaminergic pathways. Previous studies demonstrate functionally diverse subpopulations of midbrain DA neurons (Lammel, S., et al. (2014) Neuropharmacology 76 Pt B, 351-359; Heymann, G., et al. (2020) Neuron 105, 909-920; Mahadevia, D., et al. (2021) Nat. Commun. 72, 6796; Howe, M.W., and Dombeck, D.A. (2016) Nature 535, 505-510; Lerner, T.N., et al. (2015) Cell 162, 635-647; and Azcorra. M., Gaertner, Z., Davidson, C , He, Q., Kim, H., Nagappan, S., Hayes, C.K.. Ramakrishnan. C.. Fenno, L., Kim, Y.S., et al. (2023). Unique functional responses differentially map onto genetic subtypes of dopamine neurons. Nat Neurosci 26, 1762-1774; and Azcorra, M., et al. (2023) Nat Neurosci 26, 1762-1774). These populations are generally defined by a single projection target through retrograde tracing, by bulk anterograde collateralization mapping conditioned on projections to a particular target region (Beier, K.T., et al. (2015) Cell 762, 622-634; Lerner, T.N., et al. (2015) Cell 762, 635-64; and Beier, K.T., et al. (2019) Cell Rep. 26, 159-167), multiplexed retrograde tracing (Beier, K.T., et al. (2015) Cell 762, 622-634; Mahadevia, D„ et al. (2021) Nat. Commun. 72, 6796; and Lammel, S., et al. (2008) Neuron 57, 760-773), or by genetically defined adult subtypes of DA neurons (Garritsen, O., et al. (2023) Nat Rev Neurosci 24, 134-152; Azcorra, M., et al. (2023) Nat Neurosci 26, 1762-1774; Poulin, J.-F., et al. (2018) Nat Neurosci 27, 1260-1271; Poulin, J.-F., et al. (2020) Trends Neurosci. 43, 155-169; Poulin, J.-F., et al. (2014) Cell Rep. 9, 930-943; La Manno, G., et al. (2016) Cell 767. 566-580; and Pereira Luppi. M.. et al. (2021) Cell Rep. 37, 109975). While these methods provide powerful tools to dissect the dopaminergic system, they are not guaranteed to yield pure proj ectional populations based on first principles: Retrograde tracing and associated bulk collateralization mapping treat all cells projecting to the region injected with retrograde label as equal. They therefore struggle especially with resolving (partially) overlapping populations (Kebschull, J.M., et al. (2016) Neuron 91, 975-987; and Schwarz, L.A., et al. (2015) Nature 524, 88-92). Moreover, multiplexed retrograde tracing is limited in the number of target regions that can be assessed, due to the exponential decrease of co-labeling efficiency caused by the independent nature of each tracer injection, overall resulting in an underestimation of collateralization and a coarse classification of connectomic types. Additionally, in retrograde tracing, target regions are determined by the (unknown) diffusion of injected tracers, resulting in difficulty resolving adjacent structures and the blurring of projection patterns. Finally, retrograde tracing is limited to providing binary information of projections (project or not) without quantifying projection strength, fundamentally limiting its ability to define connectomic cell types.

[0196] While genetic types are defined at single cell resolution in transcriptomic space, these molecularly homogeneous cell populations are often composed of cells with heterogeneous projection patterns (Peng, H., et al. (2021) Nature 598, 174-181; and Liu, L.. et al. (2025) Nat. Methods 22, 861-873). As a result, functional measurements will reflect the aggregate activity' of distinct neuronal populations, obscuring true functional segregation. Single neuron reconstructions are important to define proj ectomi cally homogeneous DA output pathways. In a first step, a pioneering study reconstructed the single-cell projections of 30 VTA DA neurons, revealing remarkable heterogeneity’ and as a result defining 5 cell classes (Aransay. A., et al. (2015) Front Neuroanat 9, 59). The POINTseq data now- increase cell numbers over lOOx and also include the SNc, thus allow ing a comprehensive overview of the diverse single-cell projection patterns of midbrain DA neurons.

[0197] Based on the data described herein, 34 connectomic cell types were defined, far exceeding the number of reported subdivisions of DA neurons despite a conservative clustering approach (STAR Methods). Importantly, the classification recapitulates previously reported projection patterns but provides further heterogeneity within them (FIGS. 22D, 22F, and 29). For example, neurons were identified corresponding to a known population that project to CP but not NAc or mPFC (Lammel, S., et al. (2008) Neuron 57, 760-773). However, in the data described herein this population is further split into a total of 13 connectomic types (types 18-27. 32-34). which innervate CP with different preferential target subregions across LatCP, dMedCP, and vMedCP along with rostral-caudal axis. Similarly, a previously identified population which project to NAcMed while showing bulk projections to LS and MedCP (Beier, K.T., et al. (2015) Cell 162, 622-634), can be divided into NAcMed and LS projecting types (types 11-13). and NAcMed and vMedCP projecting types (types 28- 31), respectively. The data also shows heterogeneity of molecular subtypes. For example, VTA Sox6+neurons that project to the NAcLatS or NAcC, but not to the NAcMedS (Poulin, J.-F., et al. (2018) Nat Neurosci 21, 1260-127), may comprise four distinct connectomic types (types 5, 6, 28, and 30), while SNc AldhlaH- neurons that project in aggregate to CPr, LatCPi, dMedCPi, and LatCPc (Poulin, J.-F., et al. (2018) Nat Neurosci 21, 1260-127), may consist of 5 different types (types 23, 24, 26, 27, 32).

[0198] Branching patterns deviate from a random model, demonstrating function importance. This large diversity7of cell types seems to be encoded at least in part by an active process, as the frequencies of many co-innervations and projection motifs are poorly explained by a binomial model of random and independent target choice per neuron (FIGS. 23 and 24). It thus stands to reason that some, if not all, projection motifs carry functional implications. For example, the abundant over-represented NAc-CP* motif might be important for parallel processing of reward and related movements. Similarly, over-represented coinnervations and motifs between NAc. OT, and VP were identified. Given the known functions of NAc (Heymann, G., et al. (2020) Neuron 105, 909-920; Schultz, W., et al. (1997) Science 275, 1593-1599; Lammel, S„ et al. (2011) Neuron 70, 855-862; and Tobler, P.N., et al. (2005) Science 307, 1642-1645), OT (Zhang, Z , et al. (2017) Elife 6; Murata, K„ et al. (2015) J. Neurosci. 35, 10581-10599; and Ikemoto, S. (2007) Brain Res. Rev. 56, 27- 78), and VP (Richard, J.M., et al. (2018) Elife 7; Ottenheimer, D.J.. et al. (2020) Sci. Adv. 6; and Ottenheimer, D.J., et al. (2020) Nat. Neurosci. 23, 1267-1276) in the associative learning involving rewarding or aversive stimuli, and related approach or avoidance behavioural outcomes, it is thus tempting to speculate that these regions may collaborate on the processing of stimuli and learning through shared DA signalling. It was also found that over-represented co-innen ations and motifs of mPFC, EC, AMY, and CPt. DA signalling in mPFC, AMY, and CPt are implicated in negative valence, avoidance and aversive learning (Lammel, S., et al. (2014) Neuropharmacology 76 Pt B, 351-359; Menegas, W„ et al. (2018) Nat Neurosci 27, 1421-1430; Lammel, S„ et al. (2011) Neuron 70, 855-862; Tsutsui -Kimura, I., et al. (2025) Nat. Neurosci. 28, 795-810; Vander Weele, C.M., et al. (2018). Nature 563, 397-401; Zafiri, D., and Duvarci, S. (2022) Front. Behav. Neurosci. 16. 1041929; Weele, C.M.V., et al. (2019) Brain Res. 1713, 16—31; and Rosenkranz, J. A., and Grace, A. A. (2002) Nature 417, 282-287) and EC itself encodes negative valence (Liu, P., et al. (2023) Cell Rep. 42, 113204). DA signalling in EC is also involved in associative learning (Lee, J.Y., et al. (2021) Nature 598, 321-326), and mPFC and EC are co-dependent on each other for this process (Jun, H., et al. (2024) Nature 633, 864-871). Shared DA signalling among these regions may therefore be deeply involved in responses to aversive stimuli and aversive learning. Partially overlapping with but distinct from with this network is the over-represented CTX-LS motif and under-represented coinnervations of LS with regions including NAc, VP, CP*, and CPt, demonstrating that shared DA signaling between CTX and LS can serve distinct functions. Interestingly, DA signaling in LS promotes aggressive behaviors (Mahadevia, D., et al. (2021) Nat. Commun. 72, 6796; and Dai, B., et al. (2025) Nature 639, 430-437) and DA signaling in mPFC is implicated in aggression (van Erp, A.M., and Miczek, K.A. (2000) J. Neurosci. 20, 9320-9325; Shokry, I.M., et al. (2019) Exp. Neurol. 313. 26-36; Li, X., et al. (2025) Behav. Brain Res. 476, 1 15285; and Bai, F., et al. (2023). iScience 26, 107718). LS further exhibits over-represented co-innervations with other aggressive behavior-related regions such as BST (Bayless, D.W., et al. (2019) Cell 176, 1190-1205; and Padilla, S.L., et al. (2016) Nat. Neurosci. 19, 734-741) and LHb (Takahashi, A., et al. (2022); Nat. Commun. 13, 4039; and Gouveia, F.V., and Ibrahim, G.M. (2022) Front. Psychiatry 13, 817302).

[0199] CP projections of VTA and SNc DA neurons are fundamentally differently organized. SNc neurons projecting to GP and CP innervate subregions more broadly and more randomly than VTA neurons (FIG. 24). This shows that SNc neurons are functionally relatively homogeneous, broadcasting their signals across the CP. In contrast. VTA neurons participate in overrepresented projection motifs and co-innervations providing distinct functional properties of these projectionally defined VTA subpopulations.

[0200] POINTseq mapping resolution and accuracy is determined by dissection. Dissection accuracy was maximized using Nissl staining and Laser Capture Microdissection and the experiments were repeated in independent mice. Nevertheless, dissection artifacts cannot be fully excluded. In particular, some dissected brain regions are continuous without clear-cut borders. Variation in dissection in these cases might result in artifactual over-represented coinnervations of the two adjacent regions. While this possibility cannot be excluded, it is noted that many instances of non-over-represented co-innervations even between neighboring regions, such as VP-GP by VTA neurons and among CP subregions by SNc neurons was observed (FIG. 38). This finding demonstrates that the observed over-represented coinnervations or motifs are not likely to result from border artifacts.

[0201] Animals. Mice were maintained on a 12-hour light / dark cycle with ad libitum access to food and water. For POINTseq development and validation, mouse lines were purchased from the Jackson Laboratory or MMRRC centers: C57BL / 6J wildt pe (Jackson Laboratory', #000664), Cux2-Cre (MMRRC at University of Missouri, #032778-MU), Rbp4-Cre (MMRRC at UC Davis, #037128-UCD). These lines were maintained on a C57BL / 6J background, and male offspring were used for experiments. DAT-Cre (Jackson Laboratory', #006660) mice were purchased from the Jackson Laboratory' or obtained Johns Hopkins University. Both male and female mice were used for DA projection mapping. For the experiments, AAV was injected in 8-10 weeks old mice, and MAPseq virus or POINTseq virus when mice were 11-13 weeks old.

[0202] Cell culture. BHK and HEK293T cell lines yvere used to generate MAPseq viruses and AAV, respectively. HEK293T cells were maintained in medium consisting of 10% FBS (Cytiva, SH30070.03). DMEM (Gibco, 11995073), and lx Antibiotic-Antimycotic (Gibco, 15240096), yvhile BHK cells yvere maintained in medium containing 5% FBS (Cytiva, SH30070.03), MEM Alpha (Gibco, 12571063), lx Antibiotic-Antimycotic (Gibco, 15240096), lx MEM Vitamin (Gibco, 11120052), and lx L-Glutamine (Gibco, A2916801). Cells were incubated in a humidified incubator at 37°C with 5% CO2 under sterile conditions.

[0203] Constructs. pAAV-CAG-tdT (Addgene, #59462) was obtained and pAAV-CAG-tdT- P2A-Mxra8 and pAAV-CAG-FLEx-tdT-P2A-Mxra8 were created using a construct encoding mouse Mxra8 using standard cloning methods. Additionally, a helper plasmid, DH- BB[5'SIN;181 / 25ORF], was generated for POINTseq virus by replacing the region encoding the structural proteins of the Sindbis virus in the original MAPseq helper plasmid, DH- BB[5'SIN;TE12ORF] (Addgene #72309), with the CHIKV 181 / 25 structural protein sequence.

[0204] Viruses. POINTseq virus. The POINTseq viral library was generated (Kebschull, J.M., et al. (2016) Neuron 91, 975-987 for MAPseq virus. Briefly, the HKL1 plasmid library’ and the CHIKV helper plasmid BB[5'SIN;181 / 25ORF] was linearized, and genomic and helper RNAs were produced by in vitro transcription using the mMessage mMachine SP6 kit (Invitrogen, AM 1340). Genomic and helper RNAs were co-transfected into BHK cells at 80- 90% confluence in 10 cm dishes at a 1: 1 molar ratio using Lipofectamine 2000 (Invitrogen, 11668027). After 40-44 hours, the supernatant was harvested and the virus was concentrated by ultracentrifugation (160,000 x g, 2 hours). Viral titers were measured by qPCR following reverse transcription.

[0205] Plaque assay. Plaque assays were conducted (Zheng, T., et al. (2019) Biomed. Opt. Express 10, 4075-4096). Briefly, 90% confluent BHK cells were infected with 100-fold diluted MAPseq or POINTseq virus (1 pL in 200 pL medium) for 1 hour in 24-well plates. After infection, cells were overlaid with 0.4% SeaPlaque Agarose (Lonza, 50101) prepared in the same culture medium. 24 hours post-infection, 3 to 4 independent regions were imaged per w ell and the existence of plaques were manually examined based on previously described shapes (Zheng, T., et al. (2019) Biomed. Opt. Express 10, 4075-4096).

[0206] A A Vs. AAV2 / 1 w as generated using pAAV genomic plasmids (pAAV-CAG-tdT, pAAV-CAG-tdT-P2A-Mxra8, and pAAV-CAG-FLEx-tdT-P2A-Mxra8). adapting a previously described protocol. Briefly, one of the genomic plasmids, and helper plasmids p5E18(2 / l) (Addgene, #112862) and pAdDeltaF6 (Addgene, #112867) were transfected into 80% confluent HEK293T cells in 15 cm dishes using polyethylenimine (PEI) at a 1: 1: 1 molar ratio. 6 to 8 hours after transfection, the medium was replaced with fresh medium. 80 hours after transfection, the medium we harvested and it was loaded onto iodixanol gradients (54% 4 ml, 40% 5 ml, 25% 5 ml, and 15% 5 ml) prepared using OptiPrep (Sigma, DI 556). After ultracentrifugation (200,000 x g, 2 hours), the 40% layer was collected. Then, the AAV was concentrated using an Amicon Ultra-15 Centrifugal Filter, 100 kDa (Sigma, UFC910024). Viral titers were measured by qPCR. Additionally, retrograde AAV (AAVrg-EFla-Cre) were purchased from Addgene (Addgene, #55636- AAVrg).

[0207] Virus injections. Mice w ere anesthetized with isoflurane (4% for induction, 1-1.5% for maintenance) and placed in a stereotaxic frame. Lidocaine (5 mg / kg) and meloxicam (2 mg / kg) were administered subcutaneously, and an incision was made on top of the head. After drilling at the desired location, a virus was injected via glass micropipettes using a Nanoject III (Drummond Scientific) at a rate of 1 nl / sec. For MOp, the injection was paused for 10 minutes every 100 nl injection to prevent leaking. After injection, the needle and sutured the incision were removed after waiting for 10 minutes.

[0208] To evaluate Mxra8-dependent infection of POINTseq virus in MOp, striatum, dentate gyrus, thalamus, and amygdala, each region was injected with AAV-CAG-tdT (5 x 1012 vg / ml, 500 nl) or AAV-CAG-tdT-P2A-Mxra8 (5 x 1012vg / ml, 500 nl), followed by POINTseq virus (4 x 1011vg / ml, 200 nl) three weeks later. MAPseq virus (4 x 1011vg / ml, 200 nl) was also injected in each regions as controls in other mice. The following coordinates relative to bregma were used: MOp, AP: +0.5 mm, ML: ±1.5 mm, DV: -0.6 mm; Striatum, AP: ±0.5 mm, ML: ±1.8 mm, DV: -3.5 mm; Dentate gyrus, AP: -1.5 mm, ML: ±1.2 mm, DV: -1.9 mm; Thalamus, AP: -1.5 mm. ML: ±1.2 mm, DV: -3.5 mm; AMY (AP: -1.5 mm, ML: ±3.3 mm, DV: -4.6 mm.

[0209] To assess cell type-specific infections, for Cux2-Cre and Rbp4-Cre, AAV-CAG- FLEX-tdT-P2A-Mxra8 (5 x 1012vg / ml, 300 nl) was injected at 3 depths in the left MOp (AP: ±0.5 mm, ML: ±1.5 mm, DV: -0.3, -0.6. -1.0 mm relative to bregma) of Cux2-Cre and Rbp4- Cre mice. For AAVretro-Cre, AAV-CAG-FLEX-tdT-P2A-Mxra8 was injected in the same manner in left MOp, and AAVrg-EFla-Cre (1 x 1012vg / ml, 500 nl) in the right MOp (AP: ±0.5 mm, ML: -1.5 mm, DV: -0.6 mm relative to bregma). 20 days (for Cux2-Cre or Rbp4- Cre) or 27 days (for AAVretro-Cre) after AAV injections, POINTseq virus (4 x 1011vg / ml, 200 nl) was injected at 3 depths in the same locations in left MOp.

[0210] To map midbrain DA neurons in the VTA and SNc using POINTseq, AAV-CAG- FLEX-tdT-P2A-Mxra8 (5 x 1012vg / ml, 400 nl) was injected at four mediolateral locations in the left hemisphere of DAT-Cre mice (AP: -2.8 mm, ML: ±0.4, ±0.9, +1.4, +1.8 mm, DV: - 4.5 mm relative to bregma). 20 days after the AAV injection, POINTseq virus (4 x 1011vg / ml, 300 nl) was injected at the same locations.

[0211] Tissue processing. 40-44 hours after MAPseq or POINTseq virus injections, the mice were euthanized to collect their brains. For fresh brains, there were immediately froze on a metal plate at -80°C with OCT (Electron Microscopy Sciences). For fixed brains, the mice were perfused with PBS and 4% paraformaldehyde (PF A), followed by a 24-hour fixation in 4% PFA at 4°C. After fixation, the brains were stored in a 30% sucrose and 200mM glycine solution prepared in PBS at 4°C for 24 hours, then froze them at -80°C with OCT. The frozen brains were cryosectioned coronally for MAPseq, POINTseq, or imaging. For imaging, 50 pm PFA-fixed slices were mounted on slides using Fluoromount-G Mounting Medium with DAPI (Invitrogen, 00-4959-52).

[0212] For POINTseq of MOp-injected brains, 300 pm fresh frozen brain slices were mounted on slides and target regions were dissected by hand. TRIzol (Invitrogen, 15596026) was added to the dissected samples and they were homogenized using the TissueLyser III (Qiagen, 9003240). For POINTseq of VTA / SNc-injected brains, Nissl staining of 150 gm fixed brain slices was performed on ice. Briefly, after mounting, the slices were rinsed with RNase-free water (Invitrogen, 10977023) and they were stained with 0.05% toluidine blue for 45 seconds, followed by another rinse in RNase-free water. The slices were then incubated in 100% EtOH for 2 minutes before rinsing them with RNase-free water. The slices were then mounted on 4 pm-lhick PEN (polyethylene naphthalate) steel frames (Leica, 11600289), pretreated with 2% gelatin to prevent detachment, for laser capture microdissection. The frames were placed in a vacuum desiccator overnight at room temperature to dry. Once dried, the desired regions were dissected using a lOx objective on a Leica LMD7 system and they were collected in 96-well plates. No RNA degradation was observed within a week of collection when slides are stored in the dessicator. and completed the dissections within 4 days. After dissection, 100 pl of digestion solution (50 rnM Tris, 10 mM EDTA, 0.5% SDS) and 5 pl of Proteinase K (NEB, P8107S) was added to each well and the samples were incubated for 15 minutes at 50°C with 800 rpm shaking (USA Scientific, #8012-0000 and #8012-0018). The samples were then incubated at 80°C for an additional 15 minutes without shaking, and TRIzol was finally added to the digested samples.

[0213] Imaging. The images of viral infections were captured using a fluorescence microscope (Keyence BZ-X710) with 4x (NA = 0.20) and lOx (NA = 0.45) objectives. For the plaque assay (FIG. 25), a 20x objective (NA = 0.45) was used. For high magnification images in FIG. 20K and FIG. 27, a confocal laser scanning microscope (Zeiss LSM 810) with a 40x oil-immersion objective (NA = 1 .3) was used. Within each experiment, consistent imaging settings, including excitation power and exposure time, was applied.

[0214] Immunohistochemistry. Fixed brain slices were mounted on glass slides and hydrophobic borders were drawn using a PAP pen (Sigma, Z672548). The slices were then permeabilized with 0.2% Triton X-100 for 30 minutes at RT. After permeabilization, the slices were incubated in a blocking solution (5% goat serum in PBS) for 1 hour at RT. Next, a primary antibody, rabbit anti-TH (Sigma, T8700-1VL, 1 :500), was applied in the blocking solution overnight. The slices were then washed three times with PBS for 5 minutes each and incubated with a secondary antibody, goat anti-rabbit Alexa Fluor 647 (Invitrogen. A55055. 1 : 1000), in the blocking solution. After three additional PBS washes, the samples were mounted with Fluoromount-G Mounting Medium containing DAPI (Invitrogen, 00-4959-52).

[0215] MAPseq2 procedures. Amplicon sequencing libraries were prepared using the MAPseq2 protocol. Briefly, total RNA was extracted from dissected samples using TRIzol (Invitrogen, 15596026) according to manufacturer's instructions. Reverse transcription (RT) was performed using MashUp Reverse Transcriptase and sample-specific primers containing unique a molecular identifier (UMI) and a sample-specific identifier (SSI). A know n amount of spike-in RNA (105or IO7molecules to the target or source samples, respectively) resembling the viral RNA but distinguishable by an 8 nt identifier (CGTCAGTC) was also added in the barcode sequence. The samples were then treated with Exo I and RNase If immediately after RT, followed by bead clean-up. The RT products were then amplified by two rounds of PCR (PCR1 and PCR2). A 5x qPCR master mix prepared with 5x SYBR green I (Invitrogen, S7563), 5x AccuPrime reaction mix, 2.5mM MgSCU. 5pM forward primer, 5LIM reverse primer, 10% ROX Reference Dye (Invitrogen, 12223012), and 25% DMSO was used. The amplification was monitored in real time using a LightCycler96 and stopped cycling before saturation. PCR2 primers contain sample-specific identifier pairs, i5 and i7. preventing sample cross-contaminations during sequencing. After PCR2, we gel-extracted the 219bp PCR product with the QIAquick kit (Qiagen, 28704), and verified library size and concentration using a TapeStation (Agilent, G2992AA). Sequencing was performed on the NovaSeq X Plus platform using 2x150 reads. Libraries were demultiplexed by i5 / i7 index pairs, with R1 reads covering barcodes (>32 nt) and R2 covering SSI and UMI (>20 nt). Following sequences w ere used for primers.

[0216] RT primer : 5' GAC GTG TGC TCT TCC GAT CTN NNN NNN NNN NNN NNN NNN NCA CGA CGG CAA TTA GGT AGC 3’ (SEQ ID NO: 1). The 5' 12 Ns are UMI (random) and later 8 Ns are SSI (sample-specific).

[0217] PCR1 forward primer: 5' CGA GAA GCG CGA TCA CAT G 3' (SEQ ID NO: 2).

[0218] PCR1 reverse primer: 5' CTG GAG TTC AGA CGT GTG CTC TTC CGA TCT 3' (SEQ ID NO: 3).

[0219] PCR2 forward primer: 5' AAT GAT ACG GCG ACC ACC GAG ATC TAC ACN NNN NNN NAC ACT CTT TCC CTA CAC GAC GCT 3’ (SEQ ID NO: 4). 8 Ns are 15 (sample-specific).

[0220] PCR2 forward primer: 5' CAA GCA GAA GAC GGC ATA CGA GAT NNN NNN NNG TGA CTG GAG TTC AGA CGT GTG CTC TTC 3' (SEQ ID NO: 5). 8 Ns are 17 (sample-specific).

[0221] Expression ofMxra8 in the mouse brain. The scRNAseq datasets of the adult mouse whole brain were downloaded from the Allen Brain Cell Atlas (alleninstitute.github.io / abc_atlas_access / descriptions / WMB_dataset.html, WMB-10Xv2 and WMB-10Xv3), which includes log-transformed gene expression data for 3,724,993 cells. accompanied by cell metadata such as cell identity (“cell_label”) and cluster annotations C’cluster annotation term laber). While the cells were classified at various scales, the Division scale, consisting of Pallium glutamatergic, Subpallium GABAergic, PAL-sAMY- TH-HY-MB-HB neuronal, CBX-MOB-other neuronal, Neuroglial, Vascular, and Immune, was selected. The first four divisions were grouped as Neuron and Neuroglial was renamed as Glial to avoid confusions, resulting in four major cell classes of Neuronal, Glial, Vascular, and Immune. The mean expression of specific genes across the cells, classes or regions, was then assessed using the following gene IDs: Mxra8 (ENSMUSG00000029070), Slcl la2 (ENSMUSG00000023030), Ldlrad3 (ENSMUSG00000048058), Vldlr (ENSMUSG00000024924), Lrp8 (ENSMUSG00000028613).

[0222] Quantification of Mxra8-dependent infection. To quantify Mxra8 dependent infection of the POINTseq virus, mice were injected with POINTseq virus in MOp (Mxra8-, N=6 mice) or AAV-CAG-tdT-P2A-Mxra8 followed by POINTseq virus (Mxra8+, N=4 mice). The images of the slices across the injection sites were captured to count GFP-positive cells using imageJ (version 1.540. Each image was converted to 8 bit, and a threshold was applied to generate binary images of cell bodies. Then, the cells were isolated and counted using the “Watershed” and “Analyze Particles” function. Particles larger than 400 pm2were counted. In cases where cell segmentation was inaccurate, these areas were manually counted .

[0223] Estimation of POINTseq viral diversity and unique labeling. To assess the diversity of the library. Illumina sequencing was performed, which revealed approximately 2.2 x io6unique barcodes with varying abundance (FIG. 39A). It was then determined whether this distribution was sufficient to uniquely label the number of infected neurons per animal. Briefly, the barcodes with at least 3 UMI counts we included, not to overestimate the uniquely labeled fraction, and calculated the fraction using the following equation, where k is the number of infected neurons, N is the number of unique barcodes (i.e., barcode diversity), and ptis the probability7of each barcode based on the barcode distribution (FIG. 40B). Importantly, even in the case of highest number of infected neurons (2,932 neurons), the uniquely labelled fraction was close to 1 (0.9968).

[0224] Producing raw barcode matrices. To produce barcode matrices (barcodes X regions) from FASTQ sequencing files, the MAPseq2 data processing pipeline was followed. Briefly, the 30 nt barcode and a 2 nt pyrimidine anchor were stripped from Read 1 and the 8 nt SSI and a 12 nt UMI we stripped from i5,i7 demultiplexed FASTQ files, and the reads with the correct SSI were kept. Since specific SSI, i5, and i7 were used for each sample, most of the demultiplexed files should contain the correct SSI, and it was found that 98-99% of the reads had the correct SSI. Next, the SSI was removed, and the remaining 44 nt reads (barcode, anchor, UMI) were used to generate rank plots of each unique read and its abundance using Matlab (MathWorks, R2022a). Rank plots exhibit a plateau and a tail, with the plateau being reliable and the tail being unreliable. The minimal read thresholds for the samples were determined to exclude non-reliable reads and the abundance of each 30 nt barcode + 2 nt pyrimidine anchor were quantified by the number of different UMIs they were observed with. The sequences were then divided into two groups: viral barcodes and spike-in barcodes, based on whether they contained the spike-in specific sequence (CGTCAGTC). To correct for sequencing and PCR errors, any 32 nt sequences within a Hamming distance of 3 or less was collapsed. To do so, a connectivity matrix of the 32 nt sequences was constructed using Matlab (Mathworks), where connections were determined by Bow tie alignments of the barcodes with up to 3 mismatches allowed. In any connected component of this matrix the most abundant sequence represented the group. Finally, barcodes were matched up across samples by perfect identity and a raw barcode matrix (barcode x region) was constructed as a Matlab variable, where each element represented the number of UMIs of the barcode in each region.

[0225] Producing normalized and filtered projection matrices. The following procedures were conducted using Matlab. First, it was confirmed that negative controls from uninjected brains processed in parallel with the sample brain had zero counts for approximately 99.9% of barcodes. Any barcode whose maximum UMI count across target regions was <10 to only include reliable barcodes, as false-positive barcodes with up to 8 UMIs were observed (albeit rarely) in negative control samples was excluded. To ensure barcodes with well infected cell bodies in the VTA / SNc were included, a minimal UMI threshold was set for the source region. To determine an appropriate threshold, the proportion of number of targets per neuron was plotted as a function of UMI threshold and the source threshold was set where the measure stabilizes (FIG. 39A). The source threshold was 30 UMIs For MOp-injected samples and 40 UMIs for VTA / SNc-injected samples. Then, the UMIs of the filtered barcodes in each region were normalized by the abundance of spike-in RNA in that region. The resulting values for each barcode were further normalized by the total of abundance of that barcode across the target regions. Finally, the values were log-transformed and scaled such that the maximal projection value of the matrix was 1. Hierarchical clustering. Agglomerative hierarchical clustering were performed for cell type clustering on the filtered barcode matrix using Matlab. First, projection matrices were combined into rows and shuffled the rows. Then, each row was clustered through agglomerative clustering based on Euclidean distances among elements with Ward’s algorithm, using Matlab's “linkage” and “dendrogram” functions. To identify major clusters, the “maxclusf ' parameter of the dendrogram function was set to a specific number, which cuts the dendrogram at a position that yields that number of clusters. After unbiased clustering, the rows were sorted according to the order of clusters in the dendrogram and plotted the matrix. Separate projection matrices were generated by extracting rows of each experiment from the combined matrix while maintaining the sorted order of rows.

[0226] Validation of the cell type-specific mapping ofPOINTseq. To evaluate the specific mapping of POINTseq, projection matrices of 6 mice were produced using POINTseq from the Cux2-Cre, Rbp4-Cre, and AAVretro-Cre experiments (Cux2-1 : 728 neurons, Cux2-2: 1,816 neurons, Rbp4-1: 2,280 neurons, Rbp4-2: 2,747 neurons, Retro-1: 576 neurons, Retro- 2: 2,253 neurons). Then they were combined with a previous MAPseq dataset of MOp neurons with the same target regions obtained from 2 mice (MAPseq-1 : 7,506 neurons, MAPseq-2: 3,647 neurons), resulting in a projection matrix of neurons and 26 targets, without batch corrections.

[0227] Given that each experimental condition was expected to enrich for distinct connectomic cell types, hierarchical clustering on the combined projection matrix was performed. 5 major projection-defined clusters previously identified in MOp: intratel encephalic ipsilateral (ITi), intratelencephalic contralateral with or without striatal projections (ITc STR- and ITc STR+), corticothalamic (CT), and extratelencephalic (ET) neurons were recovered. Since the ITi group was previously not well resolved in relation to striatal projections, ITi into ITi STR- and ITi STR+ was further subdivided, yielding a total of 6 connectomic types.

[0228] Mapping midbrain dopaminergic neurons using POINTseq. The source region of each neuron between VTA and SNc, leveraging the property of the Sindbis genomic RNA, which exhibits much higher level of accumulation in source regions than the target regions, was determined. To do this, a source ratio was defined as the ratio of UMI counts (lower / higher) between VTA and SNc. Source ratio of 0 indicates that the barcode is detected in one region, allowing unambiguous source assignment, while source ratio of 1 suggests equal UMI counts in both regions, making the source indeterminable. Intermediate values indicates that one region with higher UMI value is likely the source while the other region is likely the target. The distribution of source ratios were then plotted and it was found that over 90% of neurons exhibited a source ratio below 0. 1 (FIG. 39B). demonstrating that most of cases, one region shows much higher UMI counts than the other. To determine whether this threshold could reliably distinguish true source region from strongly projecting target regions, a simulation was conducted using a dataset. For each simulated neuron, the higher UMI count between VTA and SNc (true source), and the highest UMI count among the target regions as the strongly projecting target (false source) were used. The region with the higher UMI count were assigned as the simulated source and the accuracy of source determination across the ratio between the two UMI values (lower / higher between true source and strong target, simulated source ratio) was assessed. In the simulations, source ratio within 0-0.05 and within 0.05-0.1 yielded more than 95% accuracy with overall accuracy more than 96% for source ratios under 0.1 (FIG. 39C and 39D). Based on this result, neurons with a source ratio less than 0. 1 in the analyses was included, ensuring high confidence in the assignment of either VTA or SNc as the source.

[0229] Agglomerative hierarchical clustering was then performed. Previously known distinct projection-defined populations, such as neurons projecting to MedCP but not to LatCP or vice versa, as well as neurons preferentially projecting to CPt, were identified. These biologically meaningful groups were used as a reference for selecting an appropriate clustering resolution and ending up with 40 clusters which was a minimum to avoid overclustering of the data. To ensure confident clusters, it was estimated whether each node of dendrogram is statistically significant using SHC (Significance of Hierarchical Clustering) testing developed to estimate statistical significance of agglomerative hierarchical clustering. In SHC, the null hypothesis assumes that the data points within a given node are drawn from a single Gaussian distribution. A null Gaussian model is then generated based on the mean and standard deviation of the data points of subjected node. For each node, a simulated Ward linkage value is calculated from the null model with matched sample sizes. The observed linkage value is then tested against the distribution of the simulated values. We simulated 500 times and p < 0.05 was considered to be significant. To correct for multiple comparisons across the dendrogram, SHC employs a sequential testing procedure that controls the familywise error rate (FWER). Starting from the root, if a node fails to reject the null hypothesis, its child nodes are considered non-significant. Using this approach, 34 statistically significant clusters we identified. SHC testing was conducted using the R package sigclust2 (version 1.2.4). Estimating the total number of neurons. To compare the observed results against expectations under an independent binomial model for co-innervation and motif analyses, the total number of neurons, Nt, which includes the observed neurons as well as neurons that do not project to any of the dissected target regions, were first estimated. Here, projection is defined as the presence of a barcode with nonzero UMI count in a given region.

[0230] When a projection matrix consists of k regions in which j -th region receives projections from Nj neurons, the probability that a neuron does not project to any regions is k n o 4) 7=1

[0231] Ntcan be then inferred, as the sum of neurons that do not project to any regions and neurons projecting to at least one region, Nobs. is Nt, where eydenotes the j -th elementary symmetric polynomial in the variables

[0232] NlrN2, ... , Nk. This polynomial was solved using the “roots” function in MATLAB and selected the largest positive real root, rounded to the nearest integer, as the estimate Nt.

[0233] Co-innervation analysis. To assess how similarly two regions, A and B, are innervated, the dice score ranging from 0 to 1. defined as N2NAnBwhere NAnB, NA, and NBare A+NB number of neurons of projecting to both regions. A, and B, respectively, was used. A dice score of 0 indicates no co-innervations while 1 indicates complete co-innervations. where the probability of projecting to region A is P(A) and projecting to region B is and we assume independent projections. Therefore, the dice score under the binomial model is equal to Nt2(NNAa+NNbB) .

[0234] The rarity score, defined as the inverse of the Dice score under the binomial assumption, can be calculated by

[0235] It indicates how rare each co-innervation is relative to the total number of neurons. The rarity score ranges from 1 to infinity, where a value of 1 indicates that A and B are coinnervated by all neurons. A rarity score of infinity- indicate at least one of the two regions is not innervated by any neuron.

[0236] Finally, the ratio of observed Dice score and estimated Dice score then can be calculated as the product of observed Dice score and rarity score by definition, which is expressed as, where the value of 1 indicates that co-innervations following the random model, while more than 1 or less than 1 indicates over- or under-represented co-innervations, respectively.

[0237] Motif analysis .

[0238] The motifs, binarized co-innervation patterns were examined across the possible region combinations, among 8 regions, CTX, NAc, OT, LS, VP, AMY, CP*, and CPt. For example, CTX-NAc motif represents neurons that project to CTX and NAc, but do not project to any other region. There are 28-l (excluding the non-projecting “motif') possible motifs with 8 target regions. Under the independent target choice model, the probability of each motif can be calculated by multiplying the probabilities of neurons projecting or not projecting to each region. Then the ratio between the observed and expected neuron counts for each motif was calculated. To minimize the sampling artifacts, motifs with either observed or expected counts higher than 5 were included in the analysis.

[0239] Statistical analyses.

[0240] Student’s t test, one-way or two-way ANOVA with Bonferroni post-hoc tests, linear regression analysis, and Pearson correlation test were performed using R functions (“t.tesf’, “aov” and “pairwise.t.tesf '). The significance of co-innervation and motifs were determined by the binomial test, and p-values were adjusted using Bonferroni correction using Matlab.

[0241] The tests were conducted as two-sided, and their significance levels are noted in the text.

Claims

CLAIMSWHAT IS CLAIMED IS:

1. A method for selectively labeling a neuron, the method comprising: a) infecting a neuron or population of neurons with a virus encoding an exogenous gene, wherein the exogenous gene encodes a receptor of an alphavirus; b) infecting the same neuron or population of neurons in a) with a chimeric virus in an amount to express a target protein, wherein the chimeric virus comprises: i) a RNA nucleic acid encoding a nucleic acid barcode, and ii) a structure protein capable of specifically binding to the alphavirus receptor encoded by the exogenous gene of a); thereby selectively labeling a neuron.

2. A method for selectively labeling a neuron, the method comprising: a) infecting a neuron or population of neurons in a subject with a chimeric virus in an amount to express a target protein, wherein the chimeric virus comprises: i) a RNA nucleic acid encoding a nucleic acid barcode, and ii) a structure protein capable of specifically binding to an alphavirus receptor; wherein the neuron or population of neurons in the subject express a receptor of an alphavirus that is capable of specifically binding to the structure protein of a)ii); thereby selectively labeling a neuron.

3. The method of claims 1 or 2, wherein the subject is a transgenic animal.

4. The method of claims 1 or 2, wherein the subject is an animal.

5. The method of claim 4. wherein the animal is a mouse, a rat, a non-human primate, a zebra fish, or a xenopus.

6. The method of claims 1 or 2, wherein the subject is engineered to specifically express the receptor of the alphavirus.

7. The method of claims 1 or 2, wherein the receptor of the alphavirus is regulatable.

8. The method of claim 1. wherein the exogenous gene in the vims of step (a) is a receptor of an alphavirus.

9. The method of claim 1. further comprising: isolating the barcode nucleic acid from each labeled neuron; amplifying the barcode nucleic acid from each labeled neuron generating barcode amplicons; sequencing the amplicons; determining barcode abundance in one or more brain regions; converting the barcode abundance to a matrix of single neuron projection patterns.

10. The method of claim 1, wherein the virus in a) is an adeno-associated virus (AAV) serotype.

11. The method of claim 10, wherein the AAV serotype is AAV1, AAV2.1, AAV 2.5, AAV 2.8, AAV 2.9, AAV DJ, or AAV PhP.eB.

12. The method of claim 1. wherein the exogenous gene is Mxra8.

13. The method of claim 1, wherein the target protein is a Mxra8 receptor.

14. The method of claims 1 or 2, wherein the chimeric virus further comprises a detection label.

15. The method of claim 14, wherein the detection label is a fluorescent protein.

16. The method of claim 15, wherein the fluorescent protein is GFP.

17. The method of claims 1 or 2, wherein the neuron or population of neurons is / are a peripheral neuron.

18. The method of claims 1 or 2, wherein the neuron or population of neurons is / are is a central nervous system neuron.

19. The method of claims 1 or 2, wherein the neuron or population of neurons are infected with the chimeric virus at a multiplicity of infection of about 1.

20. The method of claims 1 or 2, wherein the alphavirus is CHIKV, VEEV, EEEV, WEEV, SFV. or SINV (TR339), Tonate, or Everglades virus.

21. The method of claim 1, wherein step b) occurs 2-4 weeks after step a).

22. The method of claim 1. wherein the neuron or population of neurons are in a Cre mouse line.

23. The method of claim 22, wherein the Cre mouse line is Cux2-Cre, Rbp4-Cre, or DAT-Cre.

24. A method for screening a compound for its effects of on neurons or a neuronal pathway, the method comprising: obtaining a map of single neuron projections before and after exposure of the neurons or the neuronal pathway to a compound according to claim 1 or 2.

25. The method of claim 24, method comprising determining whether the compound is toxic to the neurons or the neuronal pathway or identifying a change in the projection pattern of the neurons or the neuronal pathway.

26. The method of claim 24, wherein the neurons or the neuronal pathway are from a model of a human disease or human disorder.

27. The method of claim 26, wherein the human disease is Parkinson’s disease or Alzheimer’s disease.

28. The method of claim 26, wherein the human disorder is a neurodevelopmental disorder, a neuropsychiatric disorder, or an addiction.

29. A method for labeling a neuronal pathway in an area or species which is rarely infected by Sindbis virus: a) infecting a population of neurons with chimeric viruses, wherein the chimeric alphaviruses comprise:i) an self-amplifying but propagation-incompetent RNA nucleic acid encoding a barcode nucleic acid, and ii) structure proteins of alphaviruses with varying infectivity for the neuronal pathway, thereby enhancing the labeling of a neuronal pathway across areas and species.

Images