BIOINFORMATICS RESEARCH GUIDE

Alternative Splicing and Percent Spliced-In (PSI) Analysis: Computational Methods and Tools

By Xenoibiotech · Last updated: October 2026

Alternative splicing and Percent Spliced-In (PSI or Ψ) analysis quantifies how pre-mRNAs are differentially processed into distinct transcript isoforms across developmental or pathological states. By evaluating exon-junction spanning reads and transcript abundance with computational tools like rMATS and SUPPA2, bioinformatic workflows identify exon skipping, intron retention, and splice site shifts that alter protein structure.

Key Takeaways

  • Alternative splicing allows a single gene locus to generate multiple distinct functional protein isoforms with divergent activities.
  • Percent Spliced-In (Ψ) is the mathematical ratio quantifying the proportion of transcripts that include a specific alternative exon.
  • Detection requires long, paired-end, stranded RNA sequencing with high read depth across splice junctions.
  • Event-based tools like rMATS evaluate junction-spanning reads directly, whereas isoform-based tools like SUPPA2 calculate PSI from transcript abundance.
  • Validation utilizes differential splicing statistics (|ΔΨ| > 0.1, FDR < 0.05), sashimi plots, and RT-PCR confirmation.

What Is Alternative Splicing and Why Quantify It Using Percent Spliced-In?

In eukaryotic genes, pre-mRNA molecules undergo splicing to excise non-coding introns and ligate coding exons together. Rather than producing a single static transcript, the spliceosome can selectively include or exclude specific exonic segments, alternative 5' or 3' splice sites, or entire introns. Alternative splicing significantly expands proteome diversity, with over 95% of multi-exonic human genes undergoing alternative splicing.

Aberrant splicing events drive oncogenesis, autoimmune dysfunction, and neurodegenerative disorders by truncating tumor suppressor proteins or shifting receptor isoforms. Quantifying these events requires a standardized metric: Percent Spliced-In (Ψ or PSI). PSI values range from 0 (complete exon exclusion) to 1.0 (100% exon inclusion), providing an intuitive, normalized index for comparing splicing efficiency across biological conditions.

How Do You Design Sequencing Experiments to Capture Splicing Events?

Quantifying alternative splicing imposes far stricter technical requirements on RNA sequencing than standard gene-level differential expression. Detecting alternative splicing depends on sequencing reads that directly span exon-exon junctions.

Key experimental design requirements include:

  • Sequencing Depth: While 20 million reads per sample is sufficient for gene-level expression, splicing analysis requires 50 to 100 million paired reads per sample to ensure adequate junction-spanning read depth across lowly expressed isoforms.
  • Read Length: Longer paired-end reads (2x150 bp) are strongly preferred over short reads. Longer reads map more uniquely across split junction boundaries with sufficient exonic overhangs (at least 8 to 10 nt on either side of the junction).
  • Strand Specificity: Stranded library preparation is essential to resolve antisense transcripts and ambiguous overlapping genes.

How Are Splice Junctions Mapped and What Splicing Event Types Exist?

Computational processing begins with splice-aware alignment using STAR, configured with --chimSegmentMin and two-pass alignment (--twopassMode Basic) to maximize sensitivity for novel, non-annotated splice junctions.

Alternative splicing events are classified into five canonical categories:

  • Skipped Exon (SE): An alternative cassette exon is either included or spliced out between flanking constitutive exons.
  • Alternative 5' Splice Site (A5SS): Shift in the donor splice site alters the 5' boundary of an upstream intron.
  • Alternative 3' Splice Site (A3SS): Shift in the acceptor splice site alters the 3' boundary of an intron.
  • Mutually Exclusive Exons (MXE): One of two adjacent exons is included in the mature transcript, but never both simultaneously.
  • Retained Intron (RI): An intron is retained in the mature transcript instead of being excised.

How Do Algorithms Like rMATS and SUPPA2 Calculate Delta-PSI?

Two primary bioinformatic paradigms exist for alternative splicing quantification:

Event-Centric Modeling (rMATS): rMATS analyzes raw junction-spanning reads directly from coordinate-sorted BAM files. For a skipped exon, it counts inclusion junction reads (connecting the upstream exon to the alternative exon, and the alternative exon to the downstream exon) and exclusion junction reads (connecting the upstream exon directly to the downstream exon). Using a hierarchical Bayesian framework, rMATS calculates the difference in inclusion levels (ΔΨ = ΨCondition1 - ΨCondition2) and assigns statistical significance.

Isoform-Centric Modeling (SUPPA2): SUPPA2 leverages rapid transcript quantification from Salmon or Kallisto. Given transcript-level TPM estimates and an annotated GTF file, SUPPA2 mathematically computes PSI values for each event as the ratio of expression of isoforms including the event divided by total gene expression, followed by empirical differential splicing testing.

How Do You Visualize Splicing Changes and Predict Protein Consequences?

Statistically significant splicing alterations (|ΔΨ| > 0.10, FDR < 0.05) must be inspected visually and evaluated for downstream functional consequences:

Sashimi Plot Visualization: Tools like rmats2sashimiplot or IGV Sashimi display junction-spanning read arcs superimposed over genomic exon models. The height and numerical labels of the arcs represent normalized junction read depth across experimental groups, providing visual proof of differential exon inclusion.

Protein Impact Modeling: Exon inclusion or exclusion frequently shifts open reading frames (frameshift mutations) or introduces premature termination codons (PTCs), triggering nonsense-mediated mRNA decay (NMD). In-frame alterations can add or remove key protein phosphorylation sites, transmembrane domains, or ligand-binding pockets.

What Are the Major Outputs and Common Pitfalls in Splicing Analysis?

Comprehensive deliverables include differential splicing summary tables categorized by event type, raw and normalized junction count spreadsheets, ΔΨ distribution plots, volcano plots of ΔΨ vs -log10(FDR), and publication-ready sashimi diagrams.

Critical pitfalls in splicing analysis include:

  • Insufficient Junction Overhang: Read mappers can falsely align reads with 1 to 2 base overhangs to pseudo-junctions. Filtering for minimum overhangs (>8 nt) prevents spurious junction calls.
  • Conflating Gene Expression With Splicing: A gene may produce more junction reads simply because total gene transcription increased, without any shift in splicing ratio. Evaluating ΔΨ decouples splicing regulation from overall transcriptional abundance.
  • Ignoring Intron Retention Artifacts: Pre-mRNA contamination from unlysed nuclei or genomic DNA contamination mimics intron retention events. Validation requires confirming unstranded read uniformity across introns.

Apply These Skills in Your Research

Master this computational workflow step by step with our hands-on mentoring, or collaborate with our data science team for production analysis:

Learn this in our 30-day workshop: Alternative Splicing / PSI-Seq Workshop Need this analysis done for your data? Transcriptomics & Splicing Analysis Services

Frequently Asked Questions

What does a Percent Spliced-In (PSI) value of 0.8 mean?

A PSI of 0.8 indicates that in that biological sample, 80% of the transcripts transcribed from that gene locus include the alternative exon, while the remaining 20% exclude it.

What is the difference between event-based (rMATS) and isoform-based (SUPPA2) analysis?

rMATS counts physical reads spanning specific exon junctions, providing robust local evidence independent of full-length transcript deconvolution. SUPPA2 computes PSI from estimated full-length transcript abundances, offering faster execution and broader whole-transcript context.

How does alternative splicing trigger Nonsense-Mediated Decay (NMD)?

Exon inclusion or exclusion that alters the triplet reading frame can introduce a premature termination codon (PTC). If a PTC is recognized upstream of the final exon junction complex, cellular surveillance machinery degrades the mRNA transcript via NMD.

Further Reading and Authoritative Sources

  • Shen, S., et al. (2014). rMATS: robust and flexible detection of differential alternative splicing from replicate RNA-Seq data. Proceedings of the National Academy of Sciences, 111(51), E5593–E5601. DOI: 10.1073/pnas.1419161111
  • Trincado, J. L., et al. (2018). SUPPA2: fast, accurate, and uncertainty-aware differential splicing analysis across multiple conditions. RNA, 24(9), 1158–1173. DOI: 10.1261/rna.065524.118
  • Katz, Y., et al. (2010). Analysis and design of RNA sequencing experiments for identifying isoform regulation. Nature Methods, 7(12), 1009–1015. DOI: 10.1038/nmeth.1528

Related Guide: Bulk RNA-Seq Analysis Step-by-Step Guide