# Splicewiz

> SpliceWiz

- Skill: `biomate-ai/splicewiz` (Agent Skill)
- Install (CLI): `npx skillmds@latest add biomate-ai/splicewiz`
- Raw SKILL.md: https://api.skillmd.com/api/skills/biomate-ai/splicewiz/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: biomate-ai (https://skillmd.com/u/biomate-ai)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/biomate-ai/splicewiz

---


# SpliceWiz

## Workflows

### Standard Workflow

Build a reference, process alignment BAM files, collate sample data, filter low-confidence events, and perform differential alternative splicing analysis.

```r
library(SpliceWiz)

# 1. Build the reference
ref_path <- file.path(tempdir(), "Reference")
buildRef(
  reference_path = ref_path,
  fasta = chrZ_genome(),
  gtf = chrZ_gtf(),
  ontologySpecies = "Homo sapiens"
)

# 2. Process BAM files
bams <- SpliceWiz_example_bams()
pb_path <- file.path(tempdir(), "pb_output")
processBAM(
  bamfiles = bams$path,
  sample_names = bams$sample,
  reference_path = ref_path,
  output_path = pb_path
)

# 3. Collate the experiment
expr <- findSpliceWizOutput(pb_path)
nxtse_path <- file.path(tempdir(), "NxtSE_output")
collateData(
  Experiment = expr,
  reference_path = ref_path,
  output_path = nxtse_path
)

# 4. Import the experiment
se <- makeSE(nxtse_path)
```
Input: Genome FASTA, GTF annotation, and raw BAM files. Output: A collated `NxtSE` object containing alternative splicing quantifications.

### Novel Splicing Detection

Detect and analyze novel alternative splicing events using unannotated junctions and tandem reads.

```r
library(SpliceWiz)

ref_path <- file.path(tempdir(), "Reference")
bams <- SpliceWiz_example_bams()
pb_path <- file.path(tempdir(), "pb_output")
expr <- findSpliceWizOutput(pb_path)

nxtse_path <- file.path(tempdir(), "NxtSE_output_novel")
collateData(
  Experiment = expr,
  reference_path = ref_path,
  output_path = nxtse_path,
  novelSplicing = TRUE,
  novelSplicing_requireOneAnnotatedSJ = TRUE,
  novelSplicing_minSamples = 3,
  novelSplicing_minSamplesAboveThreshold = 1,
  novelSplicing_countThreshold = 10,
  novelSplicing_useTJ = TRUE
)

se <- makeSE(nxtse_path)
```
Input: Processed BAM outputs and a reference path. Output: A collated `NxtSE` object containing identified and filtered novel splicing events.

### Star Alignment And Reference Generation

Build a STAR genome index and align FASTQ files to generate BAMs using SpliceWiz wrappers.

```r
library(SpliceWiz)

# Locate BAM files generated by STAR alignment
bams <- findBAMS(tempdir(), level = 1)
```
Input: Directory containing STAR-aligned BAM files. Output: A data frame of BAM file paths and sample names.

### Coverage Visualization

The analysis and visualization of alternative splicing (AS) events from RNA sequencing data remains

```r
library(SpliceWiz)

if (interactive()) {
  spliceWiz(demo = TRUE)
}
```
Input: Interactive R session. Output: Launches the SpliceWiz Shiny-based graphical user interface for coverage and alternative splicing visualization.

### Novel Splicing Analysis

The analysis and visualization of alternative splicing (AS) events from RNA sequencing data remains

```r
library(SpliceWiz)

ref_path <- file.path(tempdir(), "Reference")
buildRef(
  reference_path = ref_path,
  fasta = chrZ_genome(),
  gtf = chrZ_gtf()
)

bams <- SpliceWiz_example_bams()
pb_path <- file.path(tempdir(), "pb_output")
processBAM(
  bamfiles = bams$path,
  sample_names = bams$sample,
  reference_path = ref_path,
  output_path = pb_path
)

expr <- findSpliceWizOutput(pb_path)
nxtse_path <- file.path(tempdir(), "NxtSE_output_novel")
collateData(
  Experiment = expr,
  reference_path = ref_path,
  output_path = nxtse_path,
  novelSplicing = TRUE
)

se <- makeSE(nxtse_path)
```
Input: Genome FASTA, GTF, and BAM files. Output: An `NxtSE` object containing quantified annotated and novel alternative splicing events.

## When to Use
- Quantifying alternative splicing events (ASEs) such as skipped exons (SE), mutually-exclusive exons (MXE), alternative 5'/3' splice sites (A5SS/A3SS), alternate first/last exons (AFE/ALE), and retained introns (IR/RI) from RNA-seq BAM files using `processBAM()`.
- Detecting novel splicing events using unannotated junctions and tandem junction reads by setting `novelSplicing = TRUE` in `collateData()`.
- Building a customized SpliceWiz reference from genome FASTA and GTF files using `buildRef()`.
- Performing interactive differential alternative splicing analysis and coverage visualization using the Shiny GUI launched via `spliceWiz()`.

## When NOT to Use
- For alignment of raw FASTQ reads directly in R if STAR is not installed or if wrappers are not preferred; use external aligners instead.
- When system memory is extremely limited (e.g., < 8 GB RAM) for building human/mouse references or processing large BAM files; `buildRef()` and `processBAM()` require at least 8-16 GB RAM.
- For single-threaded collation of very large datasets where memory is a constraint unless `lowMemoryMode = TRUE` is specified in `collateData()`.

## Data Requirements
- Genome FASTA file (e.g., from `chrZ_genome()`) and gene annotation GTF file (e.g., from `chrZ_gtf()`).
- RNA-seq alignment BAM files (either read-name sorted or coordinate sorted) containing gapped junction reads.
- A sample annotation table (e.g., CSV format) mapping sample names to experimental conditions.

## Key Parameters
- **reference_path**: Directory path where the SpliceWiz reference is built or stored.
- **fasta**: Path to the genome FASTA file used in `buildRef()`.
- **gtf**: Path to the gene annotation GTF file used in `buildRef()`.
- **novelSplicing** (`FALSE`): Logical parameter in `collateData()` to enable novel alternative splicing event discovery.
- **novelSplicing_requireOneAnnotatedSJ** (`FALSE`): Logical parameter in `collateData()` requiring novel junctions to share at least one annotated splice site.
- **novelSplicing_minSamples** (`1`): Minimum number of samples required to retain a novel junction.
- **lowMemoryMode** (`FALSE`): Logical parameter in `collateData()` to minimize RAM usage to ~8 GB during collation.
- **n_threads** (`1`): Number of threads for multi-threaded operations in `processBAM()` and `collateData()`.

## Best Practices
- Verify that OpenMP is enabled for multi-threaded BAM processing via `processBAM()`, especially on macOS where `libomp` may need to be installed.
- Use `findBAMS()` with `level = 0` or `level = 1` to systematically locate BAM files and automatically resolve sample names.
- Use `findSpliceWizOutput()` to organize the output files of `processBAM()` before running `collateData()`.
- Set `lowMemoryMode = TRUE` in `collateData()` when collating large experiments (20+ samples) on machines with limited RAM.

## Common Pitfalls
- *Out of memory errors during collation*: Occurs when running `collateData()` with multiple threads on large datasets; fix by setting `lowMemoryMode = TRUE` to limit RAM usage to ~8 GB.
- *Low mappability regions confounding intron retention (IR) analysis*: Repetitive regions can skew IR ratios; fix by specifying `genome_type` (e.g., `"hg38"`, `"mm10"`) in `buildRef()` to use pre-built mappability exclusions.
- *Missing statistical dependencies for differential analysis*: Attempting differential analysis without required packages; fix by installing `DoubleExpSeq`, `DESeq2`, `limma`, or `edgeR`.

## Alternatives
- `DESeq2`: For gene-level differential expression analysis using negative binomial models.
- `edgeR`: For differential expression and quasi-likelihood-based analysis of count data.
- `limma`: For modeling log-normal distributions of expression data.

## Citations
- Wong, A. C. H. (2026). SpliceWiz: Quick Start. R Package Vignette.

## References
- Homepage: bioconductor.org/packages/splicewiz
- Vignette: https://bioconductor.org/packages/release/bioc/vignettes/splicewiz/inst/doc/vignette_0_cc3952e3.html