A unified framework for quantitative splicing and transcript prediction
- 1Illumina, Inc.
Abstract
RNA splicing is frequently disrupted by genetic variation, contributing to a wide range of human diseases. Identifying the variants responsible for altered splicing is therefore crucial but has long been challenging. Deep learning models that predict splicing from genomic sequence have substantially advanced their detection, but accurately resolving their quantitative effects on splicing and the resulting transcripts remains an unmet need. Here we introduce SpliceAI2, a deep learning model trained end-to-end on splice site, splice junction, and transcript usage across multiple mammalian species. SpliceAI2 reconstructs complete transcripts directly from genomic sequence and considerably improves identification of splice-altering variants and quantification of their effects in diverse benchmarks spanning cryptic splicing, splicing quantitative trait loci, and massively parallel reporter assays. Human population datasets provide independent validation of these predicted functional effects through evidence of natural selection and associated changes in plasma protein abundance. Application to a rare disease cohort reveals that predicted splice-altering variants are highly enriched in disease-associated genes and account for 15% of the excess genetic burden, with approximately half residing in deep intronic regions likely missed by exome sequencing. Together, these findings establish SpliceAI2 as a major step forward in interpreting splice-altering variants across the human genome.
Related
- Large-scale analysis of transcript data reveals thousands of recursive splicing events in human introns
- Deducing splicing events from transcript sets with splicelogic
- Towards a Unified Framework for Protein Interaction Site Prediction Using Hierarchical Protein Representations
- Unified cell segmentation for spatial transcriptomics via transcript reassignment
- Cross-cohort analysis of expression and splicing quantitative trait loci in TOPMed
BibTeX
@misc{jaganathan2026unified,
title = {{A unified framework for quantitative splicing and transcript prediction}},
author = {K. Jaganathan and J. Chen and X. Liu and Y. Zhang and Y. Kim and A. Banerjee and P. P. Fiziev and N. Wilcox and I. I. Taskiran and Y. Cai and E. M. Wigdor and G.-L. Chew and G. Png and J. F. McRae and A. J. Reddy and K.-H. Chao and R. Mehio and S. J. Sanders and K. K.-H. Farh},
howpublished = {bioRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.08.757673},
url = {https://www.biorxiv.org/content/10.64898/2026.10.08.757673v1},
note = {CC BY-NC},
}
RIS
TY - UNPB
TI - A unified framework for quantitative splicing and transcript prediction
AU - K. Jaganathan
AU - J. Chen
AU - X. Liu
AU - Y. Zhang
AU - Y. Kim
AU - A. Banerjee
AU - P. P. Fiziev
AU - N. Wilcox
AU - I. I. Taskiran
AU - Y. Cai
AU - E. M. Wigdor
AU - G.-L. Chew
AU - G. Png
AU - J. F. McRae
AU - A. J. Reddy
AU - K.-H. Chao
AU - R. Mehio
AU - S. J. Sanders
AU - K. K.-H. Farh
PY - 2026
DA - 2026/10/08
PB - bioRxiv
DO - 10.64898/2026.10.08.757673
UR - https://www.biorxiv.org/content/10.64898/2026.10.08.757673v1
AB - RNA splicing is frequently disrupted by genetic variation, contributing to a wide range of human diseases. Identifying the variants responsible for altered splicing is therefore crucial but has long been challenging. Deep learning models that predict splicing from genomic sequence have substantially advanced their detection, but accurately resolving their quantitative effects on splicing and the resulting transcripts remains an unmet need. Here we introduce SpliceAI2, a deep learning model trained end-to-end on splice site, splice junction, and transcript usage across multiple mammalian species. SpliceAI2 reconstructs complete transcripts directly from genomic sequence and considerably improves identification of splice-altering variants and quantification of their effects in diverse benchmarks spanning cryptic splicing, splicing quantitative trait loci, and massively parallel reporter assays. Human population datasets provide independent validation of these predicted functional effects through evidence of natural selection and associated changes in plasma protein abundance. Application to a rare disease cohort reveals that predicted splice-altering variants are highly enriched in disease-associated genes and account for 15% of the excess genetic burden, with approximately half residing in deep intronic regions likely missed by exome sequencing. Together, these findings establish SpliceAI2 as a major step forward in interpreting splice-altering variants across the human genome.
ER -