Introduction

SeuratWrappers is a curated collection of community-provided extensions that bring cutting-edge single-cell analysis methods into the Seurat ecosystem.

SeuratWrappers is a collection of community-provided methods and extensions for Seurat, curated by the Satija Lab at the New York Genome Center (NYGC). These integrations provide functionality not yet available in core Seurat and can be updated independently and more frequently — allowing the community to rapidly incorporate newly published single-cell analysis methods without waiting for a Seurat release.

Installation

Install SeuratWrappers and per-method dependencies.

GitHub

Browse source code, open issues, and contribute new methods.

Why SeuratWrappers exists

Seurat is a widely used framework for single-cell RNA sequencing (scRNA-seq) analysis. Its core package provides a stable, well-tested API for common operations: normalization, dimensionality reduction, clustering, and visualization. However, the single-cell field moves quickly. New methods for batch correction, trajectory inference, imputation, and spatial analysis are published continuously.

SeuratWrappers bridges this gap by:

• Extending Seurat with methods that are valuable but outside the scope of the core package
• Decoupling release cycles so individual wrappers can be updated independently of Seurat itself
• Centralizing community contributions into a single, discoverable package maintained under consistent conventions

How the wrappers work

Each wrapper follows a consistent pattern. Wrapper functions accept Seurat objects as their primary input, call the underlying method’s native API, and return results in a Seurat-compatible format — either modifying the Seurat object in place or returning a modified copy.

For example, calling RunHarmony() on a Seurat object passes the PCA embeddings to the Harmony algorithm and stores the corrected embeddings back as a new dimensional reduction in the same object. You continue to use standard Seurat functions (FindNeighbors(), RunUMAP(), etc.) on the result.

library(Seurat)
library(SeuratWrappers)

# Run Harmony integration — input and output are Seurat objects
obj <- RunHarmony(object = obj, group.by.vars = "batch")

# Downstream Seurat steps work unchanged
obj <- FindNeighbors(obj, reduction = "harmony", dims = 1:30)
obj <- FindClusters(obj)

When the underlying method returns a non-Seurat object (such as Monocle 3’s cell_data_set), SeuratWrappers provides conversion functions like as.cell_data_set() and as.Seurat() to move between representations.

Method categories

SeuratWrappers organizes methods into five broad categories:

Integration

Batch correction and multi-dataset integration methods that align cells from different samples, donors, or technologies.

Method	Description
Harmony	Iterative correction of PCA embeddings; fast and scalable
LIGER	Integrative non-negative matrix factorization
Conos	Joint graph-based integration across datasets
fastMNN	Mutual nearest-neighbor batch correction via batchelor
scVI	Deep generative model for single-cell integration

Trajectory analysis

Methods that infer developmental or dynamic relationships between cells and order them along pseudotime.

Method	Description
Monocle 3	Principal graph-based trajectory inference and pseudotime
RNA Velocity	Transcriptional dynamics from spliced/unspliced RNA ratios
tricycle	Cell cycle position estimation

Dimensionality reduction and imputation

Alternative methods for embedding cells in low-dimensional space and recovering signal from sparse count matrices.

Method	Description
ALRA	Zero-preserving imputation via low-rank approximation
GLM-PCA	PCA for count data using generalized linear models
PaCMAP	Pairwise-controlled manifold approximation

Spatial and visualization

Methods for spatially resolved transcriptomics and enhanced visualization of single-cell data.

Method	Description
BANKSY	Spatial clustering incorporating cell neighborhood context
Nebulosa	Kernel density estimation for gene expression visualization
schex	Hexagonal binning for large single-cell datasets

Quality control and utilities

Tools for filtering low-quality cells, importing data, running differential expression, and exporting results.

Method	Description
miQC	Probabilistic QC using mixture models
CoGAPS	Non-negative matrix factorization for pattern discovery
CIPR	Cell identity prediction using reference profiles
Presto	Fast Wilcoxon test for differential expression
alevin	Import salmon/alevin quantification into Seurat
Cell Browser	Export Seurat objects for UCSC Cell Browser

Next steps

Installation

Install SeuratWrappers and the dependencies for each method you plan to use.

API reference

Browse the full function reference with parameters and return types.