Decoding the Splicing Signatures of Disease

Overview

This case study details how Excelra built an end-to-end RNA-Seq splicing analysis pipeline for a leading U.S. pharmaceutical company studying RNA mis-splicing in a neuromuscular disease. Using rMATS-based differential splicing analysis, a Composite Alternative Splicing Index (CASI), and literature-grounded validation, the engagement decoded disease splicing signatures to produce a prioritized, experiment-ready shortlist of known, emerging, and novel spliceopathy candidates — together with precise exon sequences and a scalable, reusable discovery framework.

Our client

A leading U.S. pharmaceutical company with an active program in neuromuscular disease biology. The client was seeking to deepen its molecular understanding of RNA mis-splicing in its disease of interest — specifically to identify differential splicing events, recover affected exon sequences, and expand its spliceopathy gene panel with evidence-backed novel candidates.

Client’s challenge

When a leading U.S. pharmaceutical company set out to sharpen its understanding of RNA mis-splicing in a neuromuscular disease of interest, it ran into a familiar bottleneck: the disease’s molecular fingerprint was scattered across thousands of subtle splicing changes, most of them invisible to conventional expression analysis. The company didn’t just need a gene list — it needed a dependable way to separate true, disease-driving splicing events from background noise, read out the exact sequences involved, and surface new candidate genes worth adding to its spliceopathy panel.

We partnered with them to build an end-to-end RNA-Seq splicing workflow that did exactly that — pairing state-of-the-art splicing analysis with literature-grounded validation to turn raw sequencing data into a prioritized, experiment-ready shortlist.

Client’s goals

The engagement centered on four concrete goals:

• Map alternative splicing events across RNA-Seq datasets and compare disease cohorts against matched controls.
• Pinpoint the genes whose splicing was most significantly altered.
• Extract the precise nucleotide sequences of every differentially spliced exon.
• Discover novel spliceopathy-associated genes and validate them against published evidence.

Why It’s hard

Spliceopathies — diseases driven by dysregulated RNA splicing — are notoriously difficult to characterize. When splicing factors are sequestered or disrupted, the effects ripple across the entire transcriptome, producing thousands of small, correlated changes rather than a handful of obvious ones. That created several analytical hurdles:

• Processing large-scale RNA-Seq data while accurately calling differential splicing events.
• Detecting subtle but biologically meaningful shifts in exon usage — and recovering the exact transcript sequences behind them.
• Telling well-established spliceopathy genes apart from genuinely novel candidates.
• Grounding every computational hit in the published literature.

Our Approach

We designed a six-stage bioinformatics pipeline that moves from raw reads to validated, disease-relevant candidates.

1.  RNA-Seq processing & quality control

Every sample from the control and disease groups passed through a standardized preprocessing workflow — quality assessment, adapter trimming, genome alignment, and transcript quantification — producing clean, high-confidence alignments ready for splicing analysis.

2.  Alternative splicing analysis with rMATS

To call differential splicing between the two groups, we used rMATS (replicate Multivariate Analysis of Transcript Splicing), a framework purpose-built to handle biological replicates and experimental variability. It quantifies all five canonical splicing event types and reports inclusion levels (PSI), inclusion-level differences, and FDR-controlled significance — enabling robust, statistically grounded event calls.

Figure 1. The five canonical types of alternative splicing and their effect on the mature mRNA product.

3.  Prioritizing the events that matter

Not every statistically significant event is biologically important. We filtered results using stringent thresholds on FDR and inclusion-level difference, then layered in biological relevance — leaving a focused set of genes with pronounced splice dysregulation in disease samples relative to controls.

Figure 2. Heatmap of alternative-splicing event burden across disease samples. Sample identifiers have been anonymized.

Figure 3. Gene-by-event matrix showing which splicing event types were detected for each prioritized gene. Gene identities have been anonymized.

4.  Exon sequence identification

For every prioritized event, we resolved the exact nucleotide sequence of the affected exon — mapping events to reference annotations, extracting exon coordinates, retrieving strand-specific sequences, and validating them at the transcript level.

Figure 4. Representative output format capturing Gene ID, Gene Symbol, and exon coordinates for each event. Values shown are illustrative.

This exon-sequence repository fed directly into primer design, RT-PCR assay development, biomarker validation studies, and therapeutic-targeting strategies.

5.  Novel candidate discovery with CASI

To expand the client’s panel beyond known markers, we applied a Composite Alternative Splicing Index (CASI) — a scoring framework that ranks genes by the number of significant events, the magnitude of splicing disruption, consistency across samples, pathway involvement, and disease relevance. The result was a set of genes whose splicing behavior mirrors established disease spliceopathy markers, representing strong evidence-based additions to the client’s panel.

6.  Literature validation

Every high-priority candidate was then cross-checked against the published record — prior disease associations, splicing-regulatory roles, pathway context, and neuromuscular relevance — and sorted into three tiers:

This let the client concentrate experimental validation on the highest-confidence targets.

Our solution

Accelerated biomarker discovery — a ready-to-validate set of disease-associated splicing events and candidate biomarkers.

An expanded spliceopathy panel — evidence-based additions driven by CASI scoring.

Greater research efficiency — automated exon-sequence extraction replaced slow, manual curation.

A stronger translational pipeline — validated splicing signatures laid the groundwork for future diagnostic and therapeutic development.