human whole brain rna  (TaKaRa)

Journal: bioRxiv

Article Title: Aberrant recursive splicing in a human disease locus

doi: 10.1101/2025.08.14.666599

Figure Lengend Snippet: a . A case-control analysis of noncoding genomic rare copy-number variants (rCNVs), using array-based data, identified intronic deletions of CADM2 recursive splice site 1 (RS1) in genome-wide significant association with hyperactivity (ADHD). ‘ADHD-associated interval’ is the 95% credible interval containing the ADHD rare deletion association from meta-analysis across sub-cohorts. ‘UKBB common del’ is the most frequent deletion allele among the largest rCNV data source, the UK Biobank. ‘Smallest case del’ is the smallest deletion in an individual with ADHD that encompasses CADM2 RS1. See Table S3 for exact coordinates. GTEx transcripts and the local position on chromosome 3 (maroon box) are shown. b . In the present work, we identified rare overlapping intronic deletions in gnomAD, and All of Us (not shown), via WGS data. c. We generated an allelic series deleting or inverting RS1. See Table S6 for exact coordinates. Het, heterozygous. Hom, homozygous. Del, deletion. Inv, inversion. d. This allelic series was generated via CRISPR engineering in human induced pluripotent stem cells (iPSCs), which were then differentiated into glutamatergic neurons (iNs) for downstream functional genomic analyses. e. CADM2 RS sites, which are marked in fetal cortex total RNA-seq read depth (from ) by an uptick of abundance in the absence of an annotated exon. This “sawtooth” pattern of read depth is characteristic of recursive splicing. f. Recursive splicing in humans, modeled after (Cook-Andersen Nature 2015). RE, recursive exon. * Premature termination codon. ss, splice site. The RS site is the 3′ss at the left-hand side of the RE. NMD, nonsense-mediated decay. CADM2 RS1, the first of two RS sites in CADM2 , follows this characteristic architecture of human RS described in long, brain-expressed genes: It is present in the extremely long (767 kb) first intron of the gene, the site is followed by a recursive exon (RE) necessary for recognition of the RS site by splicing machinery but whose own ‘native’ 5′ splice site (5′ss) is weaker than the reconstituted 5′ss that is formed upon splicing of the exon 1 5′ss to the RS site, , and the RE contains a premature termination codon (PTC) predicted to trigger nonsense-mediated decay if included in a mature transcript ( ; Fig. S5). This architecture contrasts with the second RS site in CADM2 (RS2) – not the focus of this study – which is overlapped by an alternative exon that can be included in mature transcripts when CADM2 is driven by a minor promoter ( , ; Fig. S5). g. RE1 (between | marks) and flanking nucleotides.

Article Snippet: To obtain an initial census of CADM2 transcripts and recursive splicing in wt humans and rats, human whole brain RNA and rat brain total RNA were purchased from Clontech (TaKaRa).

Techniques: Control, Genome Wide, Generated, CRISPR, Functional Assay, RNA Sequencing

Journal: bioRxiv

Article Title: Aberrant recursive splicing in a human disease locus

doi: 10.1101/2025.08.14.666599

Figure Lengend Snippet: a . qPCR using primers amplifying from exon 1 to RE1 confirms that RS1 deletion or inversion ablates recursive splicing in a zygosity-dependent manner, with a decrease of 44-47% in het and of 99% in hom lines, compared to wt (one-way ANOVA p=1.53e-14; all comparisons to wt p<1e-6 by Tukey’s HSD test). Normalization gene is GUSB . b. Total RNA-seq demonstrates that RS1 site deletion or inversion decreases the overall expression of CADM2 , in a zygosity-dependent manner, by 23% in hom 500 bp del iNs (FDR=7.4e-3, Wald’s test) as compared to wt. c . qPCR using primers amplifying between constitutive exons 6 and 8 confirms that RS1 site deletion decreases expression of CADM2 in a zygosity-dependent manner, by 8-18% of wt in het lines and by 21% of wt in hom 500 bp del (one-way ANOVA p=0.0021 and individual comparisons to wt significant only for hom 500 bp del iNs with p=0.018 by Tukey’s HSD test; with hets combined, one-way ANOVA p=0.0082 and comparisons to wt significant for hom 500 bp del (p=0.0066) and combined hets (p=0.046) by Tukey’s HSD test). Normalization gene is GUSB . d. Normalized total RNA-seq coverage (log 2 reads/bp) mapped across CADM2 intron 1 shows the sawtooth pattern of read depth corresponding to known waypoints of recursive splicing in this gene (RS1, RS2). RS1 deletion or inversion alters the sawtooth pattern. Blue represents smoothened data. See also Fig. S16. e. Segmentation of the sawtooth pattern over CADM2 ’s first intron for one wt sample. The first segment is from exon 1 to RS1, the middle segment is from RS1 to RS2, and the end segment is from RS2 to exon 2. The red line represents the read count slope within a segment. Chr3 coordinates of the segment boundaries are shown. f. Read count slopes (normalized read depth/cumulative relative position), by genotype, within the segments defined in (e). Ablation of RS1 flattens the slope in segment 1, has no effect in segment 2 except for in het 500 bp del (see text), and has no effect on the slope in segment 3 (First segment: hom 500 bp del: p=3.4e-2, het 500 bp del p=1.2e-2; Middle segment: het 500 bp del p=9.7E–3; all by Student’s t-test). Het 100 kb del was removed from the analysis on account of removing a large part of the intron. Het, heterozygous. Hom, homozygous. Del, deletion. Inv, inversion. Significance markings are * p<0.05, ** p<0.005, *** p<0.0005. (a) and (d) are standard boxplots. See Fig. S9 for similar data from iPSCs.

Article Snippet: To obtain an initial census of CADM2 transcripts and recursive splicing in wt humans and rats, human whole brain RNA and rat brain total RNA were purchased from Clontech (TaKaRa).

Techniques: RNA Sequencing, Expressing

Journal: bioRxiv

Article Title: Aberrant recursive splicing in a human disease locus

doi: 10.1101/2025.08.14.666599

Figure Lengend Snippet: a. CADM2 transcripts identified by de novo transcript reconstruction combining total RNA-Seq and Cap-Seq data in iNs. Similarity to Ensembl transcripts is indicated by names (e.g. CADM2-201), with differences denoted as:’, differs from Ensembl exon definition; *, truncated at this exon; +, additional exon; –, missing exon; U, starts upstream of stated exon. Complete lack of overlap with Ensembl transcripts is denoted “Novel.” b. Normalized counts of reconstructed transcripts in wt iNs. See Table S13 for data. c. Standard boxplots showing fold change of selected transcripts’ TPM in iNs (see Fig. S19 for boxplots of all transcripts). This indicates that there are isolated effects on transcript choice from ablation of RS in CADM2 . Statistics are two-sided t-tests between the wt and edited cell line fold changes. * p<0.05, ** p<0.005, *** p<0.0005. d. Genotype-dependent alterations in the abundance of junctions involving REs (RE1-Ex2, RE2-Ex2), without substantial changes in splicing between either early (Ex1-Ex2, Ex2-Ex3) or later (Ex6-Ex7) coding exons. Data are iN Cap-Seq junction reads, normalized to total reads (which are highly enriched for the CADM2 locus). e. CADM2 exon abundance via DEXSeq , using total RNA-Seq data. Data are read counts from iNs, normalized to library size. Exon 1 follows the definition of “exon1_long_full” in Table S4. The low abundance of RE1 and RE2, alt exon 1, exon 2, and exon 9 are visible. RS1 deletion results in mild decreases of exon abundance of exons downstream of RS1 that are of similar magnitude to, and potentially explain, the decrease in overall expression of CADM2 . See text and Table S12 for statistics; displayed here is FDR. Figure S8 displays a similar pattern in iPSCs. * FDR<0.1.

Article Snippet: To obtain an initial census of CADM2 transcripts and recursive splicing in wt humans and rats, human whole brain RNA and rat brain total RNA were purchased from Clontech (TaKaRa).

Techniques: RNA Sequencing, Isolation, Expressing