gsea-rs

WebAssembly-ready Rust implementation of preranked Gene Set Enrichment Analysis (GSEA) distilled from GSEApy.

Setup

Prerequisites

• Rust
• wasm-pack (cargo install wasm-pack)

Installation

1. Clone repo:

git clone https://github.com/wigginno/gsea-rs.git
cd gsea-rs

2. Build the WebAssembly module:

wasm-pack build --target web

3. Serve the demo index.html file e.g. by running npx http-server in the project directory then open http://localhost:8080 in your browser.

Demo usage

Options:

• Genes (comma-separated)
• Metric (comma-separated)
• Gene Sets (in JSON format). This could easily be adapted to load .json gene sets from MSigDB.
• Weight
• Min Size
• Max Size
• Number of Permutations
• Random Seed

After filling in the form, click the "Run GSEA" button to perform the analysis. The results will be displayed in a table below the form.

API

The main function exposed by the WebAssembly module is prerank_rs:

pub fn prerank_rs(
    genes: Vec<String>,
    metric: Vec<f64>,
    gene_sets: JsValue,
    weight: f64,
    min_size: usize,
    max_size: usize,
    nperm: usize,
    seed: u64,
) -> Result<JsValue, JsValue>

This function takes the following inputs:

• genes: A vector of gene names (as strings)
• metric: A vector of metric values corresponding to the genes (as floating-point numbers)
• gene_sets: A JavaScript object representing gene sets, where keys are set names and values are arrays of gene names
• weight: The weight parameter for the GSEA algorithm (as a floating-point number)
• min_size: The minimum size of the gene sets to consider (as an integer)
• max_size: The maximum size of the gene sets to consider (as an integer)
• nperm: The number of permutations to perform for the GSEA algorithm (as an integer)
• seed: A random seed for the permutation generation (as an unsigned 64-bit integer)

The function returns a Result containing a JsValue on success, which represents an array of summary objects with the following structure:

[
  {
    term: string,
    es: number,
    nes: number,
    pval: number,
    fwerp: number,
    fdr: number,
    run_es: number[],
    hits: number[],
    esnull: number[],
    index: number | null
  },
  // ... more summary objects
]

Each summary object in the array contains:

• term: The name of the gene set
• es: The enrichment score for the gene set
• nes: The normalized enrichment score for the gene set
• pval: The p-value for the gene set
• fwerp: The family-wise error rate p-value for the gene set
• fdr: The false discovery rate for the gene set
• run_es: An array of running enrichment scores
• hits: An array of indices representing the positions of hits in the ranked list
• esnull: An array of null enrichment scores
• index: An optional index value (may be null)

To use this function in JavaScript, you'll need to use a Web Worker to run the WebAssembly module. See gsea-worker.js and gsea-app.js for implementation details.