Harmony

Fast, sensitive and accurate integration of single-cell data with Harmony

Check out the manuscript in Nature Methods:

• nature website
• read link

For Python users, check out the harmonypy package by Kamil Slowikowski.

System requirements

Harmony has been tested on R versions >= 3.4. Please consult the DESCRIPTION file for more details on required R packages. Harmony has been tested on Linux, OS X, and Windows platforms.

Installation

To run Harmony, open R and install directly from github using the following commands:

library(devtools)
install_github("immunogenomics/harmony")

Installation may include compiling C++ code from source, so it can take a few minutes.

Usage/Demos

We made it easy to run Harmony in most common R analysis pipelines.

Quick Start

Check out this vignette for a quick start tutorial.

PCA matrix

The Harmony algorithm iteratively corrects PCA embeddings. To input your own low dimensional embeddings directly, set do_pca=FALSE. Harmony is packaged with a small dataset

library(harmony)
my_harmony_embeddings <- HarmonyMatrix(my_pca_embeddings, meta_data, "dataset", do_pca=FALSE)

Normalized gene matrix

You can also run Harmony on a sparse matrix of library size normalized expression counts. Harmony will scale these counts, run PCA, and finally perform integration.

library(harmony)
my_harmony_embeddings <- HarmonyMatrix(normalized_counts, meta_data, "dataset")

Seurat

You can run Harmony within your Seurat workflow. You'll only need to make two changes to your code.

1. Run Harmony with the RunHarmony() function
2. In downstream analyses, use the Harmony embeddings instead of PCA.

For example, run Harmony and then UMAP in two lines.

seuratObj <- RunHarmony(seuratObj, "dataset")
seuratObj <- RunUMAP(seuratObj, reduction = "harmony")

For details, check out these vignettes:

• Seurat V2
• Seurat V3

MUDAN

You can run Harmony with functions from the MUDAN package. For more, details, check out this vignette.

Harmony with two or more covariates

Harmony can integrate over multiple covariates. To do this, specify a vector covariates to integrate.

my_harmony_embeddings <- HarmonyMatrix(
  my_pca_embeddings, meta_data, c("dataset", "donor", "batch_id"),
  do_pca = FALSE
)

Do the same with your Seurat object:

seuratObject <- RunHarmony(seuratObject, c("dataset", "donor", "batch_id"))

Advanced

The examples above all return integrated PCA embeddings. We created a more advanced tutorial that explores the internal data structures used in the Harmony algorithm.

Reproducing results from manuscript

Code to reproduce Harmony results from the Korsunsky et al 2019 manuscript will be made available on github.com/immunogenomics/harmony2019.