Article Figures & Data
Figures
- Figure 1.
Transcriptomic profiling of excitatory neurons in AD, DLB, and PD using short-read snRNA-seq. A, Experimental schematic of snRNA-seq and targeted Iso-Seq. B, UMAP colored by cell type from AD, DLB, PD, and age-matched controls. C, Proportion of cell types in each disease group summed across samples. D, Venn diagram of DEGs in excitatory neurons from each disease group. E, UpSet plot of excitatory neuron DEGs colored by directional overlap or discordance. F, Heat map of overlapping GO pathways affected by each disease in excitatory neurons. Red is indicative of pathways involving upregulated genes and blue is indicative of pathways involving downregulated genes. Ex, excitatory neurons; Inh, inhibitory neurons; Mic, microglia; Oli, oligodendrocytes; OPC, oligodendrocyte precursor cells; Ast, astrocytes; End, endothelial cells; Per, pericytes. See Extended Data Figure 1-1 and Extended Data Tables 1-1, 1-2, and 1-3 for more details.
- Figure 2.
Isoform classification of 50 targeted genes in neurodegenerative diseases and control. A, Heat map of gene expression changes of 50 targeted genes categorized by disease in the six major cell types. Asterisks (*) indicate significant DEGs (differentially expressed genes) in the specific cell type. B, C, Isoform structural classification and proportion of reads (B) in each individual, grouped by disease and (C) summed across individuals in each disease group by cell type. D, UpSet plot of isoforms in different disease groups. Intersection and set size represent the number of isoforms. See Extended Data Table 2-1 for more details.
- Figure 3.
Proportion of FSM (full splice match) isoforms in APP, CLU, BIN1, and MAPT by cell type. Bar plot indicates average of sample read proportions for each isoform. Gene structure of the top expressed FSM isoforms is shown below. A, APP; B, MAPT; C, CLU; D, BIN1. For comparisons of isoform proportions between groups, the proportion of each gene isoform was calculated for a given sample and averaged within each group. Each isoform was compared to the control group, and one-way ANOVA, followed by Sidak's multiple-comparisons correction, was used to determine statistical significance. See Extended Data Figure 3-1 and Extended Data Tables 3-1 and 3-2 for more details.
- Figure 4.
SQANTI3 categories and framework for annotating NNC (novel not in catalog) features. A, Seven features of NNC isoforms that create a novel junction (or two). Multiple features can be observed in a single transcript, but only one is necessary to assign the isoform to the NNC category. Dotted line indicates the new junction. Black (UTR) and gray (CDS) bars represent the reference annotation with the spaces indicating introns. Yellow bars represent examples of NNC isoforms. B, New categorization of spliced transcripts that map to a gene with a single exon. C, Example gene with multiple reference isoforms and a novel isoform with a new splice site that occurs in the CDS of one reference isoform and UTR of the other reference isoform. D, Novel isoform with a new splice site that occurs in the intron of one reference isoform of the gene and the CDS of the other reference isoform.
- Figure 5.
In-depth analysis of novel isoforms identified using targeted snIso-seq. A, B, Novel isoforms detected in each gene ranked by the number of unique transcripts (top). Proportion of novel isoforms [excluding ISM (incomplete splice match)] in each disease group represented in a heat map (bottom). Green represents (A) NIC (novel in catalog) isoforms and yellow represents (B) NNC (novel not in catalog isoforms). C, D, NNC features identified using modified SQANTI3 and quantified by (C) number of transcripts or (D) number of reads. Each black dot represents a sample. Statistical significance was assessed by one-way ANOVA, p < 0.0001. See Extended Data Figures 5-1, 5-2, and 5-3 for more details.
- Figure 6.
An example NNC (novel not in catalog) APP isoforms with a novel exon. A, An APP NNC (novel not in catalog) isoform that is potentially translated into truncated protein isoforms lacking the soluble APP domain. B, Alignment between the potential novel APP protein isoform with canonical full-length APP-770. Brown-shaded regions represent conserved sequences. Dotted vertical lines represent respective secretase cleavage sites. See Extended Data Figure 6-1 for more details.
Extended Data
Figure 1-1
Sample metadata and cell type proportions. (A) Plots showing the distribution of sex (left), age (middle), and RIN (right) of all samples. (B) UMAP plots colored by sex (left), RIN (middle), and age (right). (C) PCA plots colored by sex (left), RIN (middle), and age (right). (D) Cell type proportions in each individual sample. Download Figure 1-1, TIF file.
Table 1-1
Sample information and sequencing QC. Neuropathology and brain bank information are provided for each sample. Basic 10X QC and PacBio metrics for each sample demonstrate how many reads were obtained and various filters that were used. Download Table 1-1, XLS file.
Table 1-2
Differential gene expression. For each cell type and disease, comparisons were made to control samples. The resulting gene lists with adjust p-values are provided. Download Table 1-2, XLS file.
Table 1-3
Gene Ontology Analysis. Gene Ontology (GO) terms reported for each cell type’s differentially expressed genes in each of the diseases. Download Table 1-3, XLS file.
Table 2-1
Target gene panel. Various metrics associated with the fifty genes that were targeted for long-read sequencing and the number of isoforms and reads that were captured. Download Table 2-1, XLS file.
Figure 3-1
Proportion of Full Splice Match isoforms. Stacked bar plots of averaged FSM isoform proportion for each disease group and cell type. Each plot represents a gene from the target enrichment panel and only includes reads from isoforms categorized as FSM. NA represents reads from unassigned cell types. Download Figure 3-1, TIF file.
Table 3-1
isoSeQL output. Tabular output from isoSeQL listing the structural information for each isoform and the corresponding counts separated by cell type and sample. Download Table 3-1, XLS file.
Table 3-2
InterProScan output. Tabular output from InterProScan annotating the functional domains of protein sequences predicted from the isoform sequences detected through PacBio long-read sequencing. Download Table 3-2, XLS file.
Figure 5-1
Structural category proportions of targeted genes. Stacked bar plots showing the proportion of reads that support isoforms in each structural category for each gene from our enrichment panel. The number on top of each bar represents number of reads. Download Figure 5-1, TIF file.
Figure 5-2
Novel isoform correlation with gene characteristics. Relationship between average transcript length and number of NIC (top left) or NNC (top right) reads. Each dot corresponds to a gene in our enrichment panel. Relationship between average number of exons in isoforms of a gene and the number of NIC (bottom left) or NNC (bottom right) reads. Each dot represents a gene. Download Figure 5-2, TIF file.
Figure 5-3
NNC features by disease group and cell type. (A) Number of NNC isoforms with a particular NNC feature per sample grouped by disease. Bar represents the mean across samples, and error bars represent standard deviation. (B) Number of NNC isoforms with a particular NNC feature per sample grouped by cell type. Bar represents the mean and error bars indicate standard deviation. Download Figure 5-3, TIF file.
Figure 6-1
Select examples of NNC transcripts. The NNC transcript is shown in reference to Ensembl reference transcripts. In-frame translations of these NNC transcripts are aligned with the canonical protein isoform. Domains of full-length canonical proteins are highlighted to demonstrate loss of functional domains in the theoretically translated diseased isoform. Download Figure 6-1, TIF file.
In this issue
