CoChat

Identity rather than 3D position informs splicing of rare introns in the human genome

Open access

The DOI matches its Crossref record; 122 of the 122 references it cites were found in a registry or an index; none of the 48 sources its claims cite is checked for support yet.

Abstract

Three-dimensional genome organization influences splicing, yet how introns, particularly rare introns, are arranged within the nucleus and whether this organization contributes to splicing have not been fully explored. We mapped the organization of six intron classes, major, minor, minor-like, hybrid, major-like, and non-canonical, across four human cell lines using Hi-C, TSA-seq, and DamID-seq. Minor intron-containing genes and their introns were enriched in compartment A and speckle-associated domains (SPADs) and depleted from lamina-associated domains. Integrating TSA-seq with RNA-seq indicated that splicing efficiency depends on intron identity rather than position. For example, despite shared SPAD proximity, major-like and minor-like introns were less efficiently spliced than major and minor introns. While minor introns in SPADs were consistently efficiently spliced, minor introns outside SPADs were more efficiently spliced in cancer cells. These findings reveal subclass-specific intron organization, yet identity rather than position governs splicing efficiency, particularly for minor introns in cancer cells.

Overview

CoChat Summary

TL;DR

This study maps six intron classes across four human cell lines and finds that minor introns are enriched in compartment A and speckle-associated domains (SPADs) and depleted from lamina-associated domains (LADs). Splicing efficiency is governed primarily by intron identity rather than 3D position, with minor introns consistently efficiently spliced in SPADs and, in cancer cells, also outside SPADs.

Key findings

  • Rare introns are dispersed across the linear genome; median inter-intron distance increases with rarity: major 141 bp, major-like 27 kb, non-canonical 105 kb, minor 1.8e06 bp, minor-like 4.0e06 bp, and hybrid 6.1e06 bp, with a near-linear relationship between inter-intron distance and intron abundance (R² = 0.98).
  • Minor introns are enriched in compartment A (71% vs 62% of major introns; p = 4.7e-08) and SPADs (25% vs 21% of major introns; p = 0.018), and depleted from LADs (14% vs 19% of major introns; p = 7.7e-05).
  • Within SPADs, major and minor introns are efficiently spliced, whereas major-like and minor-like introns show higher intron retention than their major/minor counterparts; non-canonical introns show high retention and exon skipping. Minor introns in SPADs are consistently efficiently spliced across K562, HFFc6, H1, and HCT116 cells.
  • Minor introns outside SPADs are more efficiently spliced in cancer cells (K562 and HCT116) than in non-cancerous HFFc6 fibroblasts; MIGs with enhanced splicing in both cancer cell lines are enriched for cell cycle regulators.

Methods

  • Public Hi-C, TSA-seq, DamID-seq, and RNA-seq datasets from K562, HFFc6, H1, and HCT116 cells were used; introns were classified into major, major-like, hybrid, minor-like, minor, and non-canonical classes using position weight matrices and mapped to compartments, subcompartments, SPADs, LADs, TADs, loop anchors, and boundaries.
  • Bootstrapping and pollution analyses tested whether spatial biases were tethered to intron identity; splicing was quantified as intron retention and exon skipping using modified bioinformatics pipelines, with statistical testing by Fisher’s exact test, Mann-Whitney U, Kruskal-Wallis with Dunn post hoc, and Gene Ontology enrichment.

Limitations

  • Stated in the paper: small numbers of minor-like and hybrid introns limit statistical power; strict expression and read-coverage thresholds further reduce analyzable introns; the chosen cell lines had vertically integrated datasets, so additional cell lines or new data could not be included; donor sex, gender, ancestry, race, and ethnicity were not reported and were not evaluated.
  • Evident from the text: minor intron LAD comparisons were based on only 8 and 2 minor introns at ≥1 and ≥20 TPM thresholds, respectively; only 6 hybrid introns in SPADs had sufficient data for splicing analysis; the shared minor intron set across all four cell lines was 11 introns.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1016/j.isci.2025.114532

Cites

122 of 122 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Graphical abstract
Graphical abstract CC BY-NC-ND · PubMed Central
Figure 1 Distribution of intron classes across the linear genome (A) Bar graph showing the percentage of the coding and non-coding genome occupied by exons, major introns, and rare introns. (B) Bar graph showing the percentage of the coding genome occupied by rare intron classes, including minor, minor-like, hybrid, major-like, and non-canonical introns, as defined by Olthof, Schwoerer et al.49 Details regarding the unique sequence composition of introns per class can be found in Figure S1. (C) Boxplot showing the distribution of inter-intron distance (in base pairs, bp) for each intron class. Boxes represent the interquartile range (25th to 75th percentile), center lines indicate the median, and whiskers denote the 1.5× interquartile range. Median distance per class is listed to the right of each boxplot. Log transformed inter-intron distance values can be found in Figure S2E. (D) Scatterplot establishing the relationship between inter-intron distance and intron number per class. Line of best fit and R2 value are shown in the graph area. (E) Chromosomal map depicting the genome-wide distribution of non-canonical, minor, minor-like, and hybrid introns. Major and major-like introns were excluded from plotting due to their overabundance relative to the other intron classes (E, insets). Introns are color coded in accordance with the spectrum listed in B. (F) Zoom-in of chromosome 1 from E. Visual summary of depicting quantification of intron density in fixed genomic windows. Approximate genomic location is shown in mega bp (Mbp). (G) Intron density per 250 kb bin along chromosome 1. Each bar represents the total number of introns within a single 250 kb genomic window. Inset highlights a gene cluster located within a single 250 kb bin composed of multiple minor intron-containing genes (MIGs) from the GBP gene family. (H) Intron density per 250 kb bin along the Y chromosome. Each bar represents the total number of introns within a single 250 kb genomic window. Inset highlights a gene cluster located within a single 250 kb bin composed of multiple MIGs from the TSPY gene family. Results of 250 kb binning, including numbers of introns per bin and parent genes, can be found in Data S1. See also Figures S1–S3, Table S1 and S2, and Data S1–S4.
Figure 1 Distribution of intron classes across the linear genome (A) Bar graph showing the percentage of the coding and non-coding genome occupied by exons, major introns, and rare introns. (B) Bar graph showing the percentage of the coding genome occupied by rare intron classes, including minor, minor-like, hybrid, major-like, and non-canonical introns, as defined by Olthof, Schwoerer et al.49 Details regarding the unique sequence composition of introns per class can be found in Figure S1. (C) Boxplot showing the distribution of inter-intron distance (in base pairs, bp) for each intron class. Boxes represent the interquartile range (25th to 75th percentile), center lines indicate the median, and whiskers denote the 1.5× interquartile range. Median distance per class is listed to the right of each boxplot. Log transformed inter-intron distance values can be found in Figure S2E. (D) Scatterplot establishing the relationship between inter-intron distance and intron number per class. Line of best fit and R2 value are shown in the graph area. (E) Chromosomal map depicting the genome-wide distribution of non-canonical, minor, minor-like, and hybrid introns. Major and major-like introns were excluded from plotting due to their overabundance relative to the other intron classes (E, insets). Introns are color coded in accordance with the spectrum listed in B. (F) Zoom-in of chromosome 1 from E. Visual summary of depicting quantification of intron density in fixed genomic windows. Approximate genomic location is shown in mega bp (Mbp). (G) Intron density per 250 kb bin along chromosome 1. Each bar represents the total number of introns within a single 250 kb genomic window. Inset highlights a gene cluster located within a single 250 kb bin composed of multiple minor intron-containing genes (MIGs) from the GBP gene family. (H) Intron density per 250 kb bin along the Y chromosome. Each bar represents the total number of introns within a single 250 kb genomic window. Inset highlights a gene cluster located within a single 250 kb bin composed of multiple MIGs from the TSPY gene family. Results of 250 kb binning, including numbers of introns per bin and parent genes, can be found in Data S1. See also Figures S1–S3, Table S1 and S2, and Data S1–S4.CC BY-NC-ND · PubMed Central
Figure 2 Introns exhibit class-specific biases in 3D genome organization (A) Schematic showing the levels of genome organization interrogated in this study. Annotations were derived from K562 cells and include compartments (A and B), subcompartments (A1, A2, B1, B2, and B3), domains associated with nuclear bodies (SPADs and LADs), and 3D features (TADs, boundaries, and loop anchors). (B) Proportional distribution of each intron class across compartments A and B. Introns that mapped to unassigned regions, meaning they did not intersect with either compartment A or B, were listed as “Neither.” (C) Fisher exact test comparing the compartment distribution of major versus minor introns. (D) Bootstrapping and permutation analysis depicting the proportion of introns per compartment across 1,000 bootstrapped intron sets that vary the percentage of major introns versus minor introns (see also Figure S5). Boxplots depict the distribution of 1,000 bootstrapping replicates where boxes represent the interquartile range (25th to 75th percentile), center lines indicate the median, and whiskers denote the 1.5× interquartile range. (E) Proportional distribution of introns localized to compartments A and B across their respective subcompartments (A1, A2, B1, B2, and B3), shown for each intron class. (F–H) Intersection of genes in the total essentialome with genes containing each intron class (Figure S3B) revealed essential genes that contain each intron class (F). Functional enrichment and intersection of the resulting terms from hybrid, minor-like, minor, and non-canonical intron-containing genes produced 1,314 GO terms (G), including “cell cycle.” Intersection of the 146 genes underlying cell cycle enrichment in each intron class revealed little overlap (H). (I–K) Fisher exact test comparing the compartment distribution of (I) all major introns versus major introns in essential genes, (J) all minor introns versus minor introns in essential genes, and (K) major introns versus minor introns in non-essential genes. Raw intron counts used in proportion calculations and additional statistical tests for B, C, and E can be found in Figure S4. Raw intron counts used in proportion calculations and additional statistical tests for I-J can be found in Figure S10. See also Figures S4–S10, Table S3, and Data S5–S12.
Figure 2 Introns exhibit class-specific biases in 3D genome organization (A) Schematic showing the levels of genome organization interrogated in this study. Annotations were derived from K562 cells and include compartments (A and B), subcompartments (A1, A2, B1, B2, and B3), domains associated with nuclear bodies (SPADs and LADs), and 3D features (TADs, boundaries, and loop anchors). (B) Proportional distribution of each intron class across compartments A and B. Introns that mapped to unassigned regions, meaning they did not intersect with either compartment A or B, were listed as “Neither.” (C) Fisher exact test comparing the compartment distribution of major versus minor introns. (D) Bootstrapping and permutation analysis depicting the proportion of introns per compartment across 1,000 bootstrapped intron sets that vary the percentage of major introns versus minor introns (see also Figure S5). Boxplots depict the distribution of 1,000 bootstrapping replicates where boxes represent the interquartile range (25th to 75th percentile), center lines indicate the median, and whiskers denote the 1.5× interquartile range. (E) Proportional distribution of introns localized to compartments A and B across their respective subcompartments (A1, A2, B1, B2, and B3), shown for each intron class. (F–H) Intersection of genes in the total essentialome with genes containing each intron class (Figure S3B) revealed essential genes that contain each intron class (F). Functional enrichment and intersection of the resulting terms from hybrid, minor-like, minor, and non-canonical intron-containing genes produced 1,314 GO terms (G), including “cell cycle.” Intersection of the 146 genes underlying cell cycle enrichment in each intron class revealed little overlap (H). (I–K) Fisher exact test comparing the compartment distribution of (I) all major introns versus major introns in essential genes, (J) all minor introns versus minor introns in essential genes, and (K) major introns versus minor introns in non-essential genes. Raw intron counts used in proportion calculations and additional statistical tests for B, C, and E can be found in Figure S4. Raw intron counts used in proportion calculations and additional statistical tests for I-J can be found in Figure S10. See also Figures S4–S10, Table S3, and Data S5–S12.CC BY-NC-ND · PubMed Central
Figure 3 Genes containing different intron classes exhibit distinct expression patterns that correspond with 3D genome organization (A) Bar charts of median transcript per million (TPM) values for genes containing each intron class. (B–D) Schematics depicting the organization of compartments and subcompartments from the center to periphery of the nucleus. Expression of genes containing each intron class across compartments (C) and subcompartments (D). Lines represent median TPM values. (E) Proportion of introns per class that are in compartment A and fall within expressed genes. (F) Fisher exact test comparing the distribution of major versus minor introns in compartment A among expressed genes. (G) Proportion of introns per class that are in compartment B and fall within expressed genes. (H) Fisher exact test comparing the distribution of major versus minor introns in compartment B among expressed genes. Raw intron counts used in proportion calculations and additional statistical tests for E-H can be found in Figure S12. See also Figures S11–S14 and Data S13–S15.
Figure 3 Genes containing different intron classes exhibit distinct expression patterns that correspond with 3D genome organization (A) Bar charts of median transcript per million (TPM) values for genes containing each intron class. (B–D) Schematics depicting the organization of compartments and subcompartments from the center to periphery of the nucleus. Expression of genes containing each intron class across compartments (C) and subcompartments (D). Lines represent median TPM values. (E) Proportion of introns per class that are in compartment A and fall within expressed genes. (F) Fisher exact test comparing the distribution of major versus minor introns in compartment A among expressed genes. (G) Proportion of introns per class that are in compartment B and fall within expressed genes. (H) Fisher exact test comparing the distribution of major versus minor introns in compartment B among expressed genes. Raw intron counts used in proportion calculations and additional statistical tests for E-H can be found in Figure S12. See also Figures S11–S14 and Data S13–S15.CC BY-NC-ND · PubMed Central
Figure 4 Minor introns are enriched in SPADs, but this enrichment does not correlate with an increase in splicing efficiency relative to LADs (A) Proportional distribution of each intron class across domains associated with nuclear bodies, including SPADs and LADs. Introns that mapped to unassigned regions, meaning they did not intersect with either SPADs or LADs, were listed as “Neither.” (B) Fisher exact test comparing the nuclear domain distribution of major versus minor introns and major versus non-canonical introns. (C) Proportion of expressed introns per class in SPADs. (D) Fisher exact test comparing the distribution of major versus minor introns and major versus non-canonical introns in SPADs among expressed genes. (E and F) Boxplots of major (E) and minor (F) intron retention for genes filtered at ≥ 1 or 20 TPM. (G and H) Boxplots of major (G) and minor (H) exon skipping for genes filtered at ≥ 1 or 20 TPM. Raw intron counts used in proportion calculations and additional statistical tests for A and B can be found in Figure S15. Raw intron counts used in proportion calculations and additional statistical tests for C and D are listed in Figure S16. See also Figures S15–S17 and Table S4. Boxplots depict the median (center line), interquartile range (box), and the 1.5× interquartile range (whiskers).
Figure 4 Minor introns are enriched in SPADs, but this enrichment does not correlate with an increase in splicing efficiency relative to LADs (A) Proportional distribution of each intron class across domains associated with nuclear bodies, including SPADs and LADs. Introns that mapped to unassigned regions, meaning they did not intersect with either SPADs or LADs, were listed as “Neither.” (B) Fisher exact test comparing the nuclear domain distribution of major versus minor introns and major versus non-canonical introns. (C) Proportion of expressed introns per class in SPADs. (D) Fisher exact test comparing the distribution of major versus minor introns and major versus non-canonical introns in SPADs among expressed genes. (E and F) Boxplots of major (E) and minor (F) intron retention for genes filtered at ≥ 1 or 20 TPM. (G and H) Boxplots of major (G) and minor (H) exon skipping for genes filtered at ≥ 1 or 20 TPM. Raw intron counts used in proportion calculations and additional statistical tests for A and B can be found in Figure S15. Raw intron counts used in proportion calculations and additional statistical tests for C and D are listed in Figure S16. See also Figures S15–S17 and Table S4. Boxplots depict the median (center line), interquartile range (box), and the 1.5× interquartile range (whiskers).CC BY-NC-ND · PubMed Central
Figure 5 Splicing of rare intron classes in SPADs is conserved across cell lines (A and B) Boxplots depicting the distribution of intron retention (A) and exon skipping (B) in SPADs of K562 cells for each intron class. (C–E) Boxplots depicting the distribution of intron retention values for each intron class in SPADs of HFFc6 (C), H1 (D), and HCT116 (E) cells. (F–H) Boxplots depicting the distribution of exon skipping values for each intron class in SPADs of HFFc6 (F), H1 (G), and HCT116 (H) cells. Boxplots depict the median (center line), interquartile range (box), and the 1.5× interquartile range (whiskers). Comparisons involving more than two groups were performed using Kruskal-Wallis tests followed by the Dunn post hoc test with Benjamini-Hochberg adjustment. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. See also Figure S17 and Data S16 and S17.
Figure 5 Splicing of rare intron classes in SPADs is conserved across cell lines (A and B) Boxplots depicting the distribution of intron retention (A) and exon skipping (B) in SPADs of K562 cells for each intron class. (C–E) Boxplots depicting the distribution of intron retention values for each intron class in SPADs of HFFc6 (C), H1 (D), and HCT116 (E) cells. (F–H) Boxplots depicting the distribution of exon skipping values for each intron class in SPADs of HFFc6 (F), H1 (G), and HCT116 (H) cells. Boxplots depict the median (center line), interquartile range (box), and the 1.5× interquartile range (whiskers). Comparisons involving more than two groups were performed using Kruskal-Wallis tests followed by the Dunn post hoc test with Benjamini-Hochberg adjustment. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. See also Figure S17 and Data S16 and S17.CC BY-NC-ND · PubMed Central
Figure 6 Minor introns outside of SPADs are spliced more efficiently in K562 and HCT116 cancer cells than in HFFc6 fibroblast cells (A and B) Boxplots depicting the distribution of intron retention values for all major (A) and minor (B) introns that pass filtering in each cell line. (C and D) Boxplots depicting the distribution of intron retention values for 173 major introns (C) and 11 minor introns (D) with sufficient data across all cell lines. (E) The number of introns with the same nuclear body associated location (SPAD, LAD, or Neither) in HFFc6 and K562, HCT116, or H1 cells in pairwise comparisons. (F) Boxplots depicting the distribution of intron retention values for introns with the same location in HFFc6 cells and either H1, K562, or HCT116 cells. (G and H) Scatterplots of SIRET values of MIGs in HFFc6 cells and K562 (G) or HCT116 (H) cells. Each point represents a single intron. Dashed lines indicate the line of identity (y = x). (I) Intersection of MIGs with efficient splicing in K562 and/or HCT116 cells relative to HFFc6 cells. (J) Gene Ontology (GO) terms for MIGs with only more efficient splicing in K562 relative to HFFc6, only more efficient splicing in HCT116 relative to HFFc6, or more efficient splicing in both K562 and HCT116 relative to HFFc6. (K) Increased expression of splicing factors in K562 and HCT116 cells relative to HFFc6 cells. Boxplots depict the median (center line), interquartile range (box), and the 1.5× interquartile range (whiskers). Pairwise comparisons were performed using Mann-Whiney U while comparisons involving more than two groups were performed using Kruskal-Wallis tests followed by the Dunn post hoc test with Benjamini-Hochberg adjustment. ∗p < 0.05, ∗∗p < 0.01, n.s., not significant. See also Figure S19 and Data S18.
Figure 6 Minor introns outside of SPADs are spliced more efficiently in K562 and HCT116 cancer cells than in HFFc6 fibroblast cells (A and B) Boxplots depicting the distribution of intron retention values for all major (A) and minor (B) introns that pass filtering in each cell line. (C and D) Boxplots depicting the distribution of intron retention values for 173 major introns (C) and 11 minor introns (D) with sufficient data across all cell lines. (E) The number of introns with the same nuclear body associated location (SPAD, LAD, or Neither) in HFFc6 and K562, HCT116, or H1 cells in pairwise comparisons. (F) Boxplots depicting the distribution of intron retention values for introns with the same location in HFFc6 cells and either H1, K562, or HCT116 cells. (G and H) Scatterplots of SIRET values of MIGs in HFFc6 cells and K562 (G) or HCT116 (H) cells. Each point represents a single intron. Dashed lines indicate the line of identity (y = x). (I) Intersection of MIGs with efficient splicing in K562 and/or HCT116 cells relative to HFFc6 cells. (J) Gene Ontology (GO) terms for MIGs with only more efficient splicing in K562 relative to HFFc6, only more efficient splicing in HCT116 relative to HFFc6, or more efficient splicing in both K562 and HCT116 relative to HFFc6. (K) Increased expression of splicing factors in K562 and HCT116 cells relative to HFFc6 cells. Boxplots depict the median (center line), interquartile range (box), and the 1.5× interquartile range (whiskers). Pairwise comparisons were performed using Mann-Whiney U while comparisons involving more than two groups were performed using Kruskal-Wallis tests followed by the Dunn post hoc test with Benjamini-Hochberg adjustment. ∗p < 0.05, ∗∗p < 0.01, n.s., not significant. See also Figure S19 and Data S18.CC BY-NC-ND · PubMed Central

Verification

Claims the paper makes

0 Valid · 0 Not valid · 25 Not checked · 185 Own results and interpretations

117 more claims

Create a free account to see all 210 claims and to verify the 48 references still open.

See all claims — free account
  • It may state two separate claims as one. 72
  • Probably not a substantial claim. 24

Create a free account to see these claims.

Citations

Cites

Show all 122 references

Cited by

See all 2

Metadata

Authors
Springer, S. M., Fleck, K., Girardini, K. N., Riccard, S. M., Erceg, J., & Kanadia, R. N.
Affiliations
University of Connecticut (US)
Corresponding author
  • Erceg, Jelena
  • Kanadia, Rahul N.
Venue
iScience
Publisher
Elsevier BV
Published
2026-01
Volume
29
Issue
1
Pages
114532
ISSN
2589-0042 (ISSN-L)
OpenAlex
W7117128063
Cited by
2
References
121
Citations by year
2026: 2
FWCI
0.39
Type
journal article
Language
en
Open access
Licence
Locations
  • iScience (publishedVersion, cc-by-nc-nd)
  • bioRxiv (Cold Spring Harbor Laboratory) (submittedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by-nc-nd)
  • PubMed (publishedVersion)
  • PubMed Central (submittedVersion, cc-by)
  • PubMed (submittedVersion)
Funders
  • Prostate Cancer Foundation
  • National Institutes of Health
  • University of Connecticut
  • National Institute of General Medical Sciences (R35GM146922)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › RNA Research and Splicing
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Genomics and Chromatin Dynamics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Cancer Research › Cancer-related molecular mechanisms research
Keywords
alternative splicing, 3D genome organization, lamina-associated domains, splicing efficiency, cancer cells, human cell lines

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

A language model drew the claims from the paper on 2026-11-22, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-11-22 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Identity rather than 3D position informs splicing of rare introns in the human genome. https://app.cochat.ai/papers/10.1016/j.isci.2025.114532
The paper
Springer, S. M., Fleck, K., Girardini, K. N., Riccard, S. M., Erceg, J., & Kanadia, R. N. (2026). Identity rather than 3D position informs splicing of rare introns in the human genome. iScience. https://doi.org/10.1016/j.isci.2025.114532

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.