In vjcitn/mowChIP: Infrastructure for microfluidic oscillatory washing-based ChIP

suppressPackageStartupMessages({
suppressMessages({
library(mowChIP)
library(aws.s3)
})
})

Introduction

The Nature Methods paper of @Cao2015 details a technology for improving chromatin immunoprecipitation (ChIP) assays, demonstrating good sensitivity for samples with as few as 100 cells. This package will provide access to an archive of high-resolution outcomes from this protocol, along with tools for visualization and inference on key questions related to implementation and interpretation.

Remote data resources

We have created an AWS S3 bucket with eight bigwig files with genome-wide scores for histone modification as measured by MOWChIP-seq on samples with 10K, 1K, 600, and 100 cells. For the following code to run, it appears that you will have to have AWS credential established for the r CRANpkg("aws.s3") bucket queries to work, despite the fact that read bucket permission is granted to the public at large.

library(aws.s3)
info = get_bucket_df("bcfound-mow")
info[,c(1,4)]

We use r Biocpkg("GenomicFiles") utilities to manage access to these files. The file metadata is provided in the caoChIP GenomicFiles instance, and the paths are assigned using information about the bucket content URLs.

library(GenomicFiles)
data(caoChIP)
names(colData(caoChIP))
colData(caoChIP)[c(1,3,5,7),c("mark", "numCells")]
str(mowChInS3()) # full URL to bucket contents

Static visualization

The viewByGene function plots coverage over gene regions for selected genes. The following plots data related to Figure 2c of the main paper. The function accepts gene symbol and radius around gene to display, and uses the Bioconductor GenomicFiles data store, in this case a collection of bigwig files in a publicly accessible S3 bucket, to acquire the coverage values.

S3paths = mowChInS3()
remoteMowK27ac = GenomicFiles(files=S3paths[5:8], 
    colData=colData(caoChIP[,c(9,11,13,15)]))
require(EnsDb.Hsapiens.v75)
gg = genes(EnsDb.Hsapiens.v75)
require(GenomeInfoDb)
seqlevelsStyle(gg) = "UCSC"
viewByGene(gf=remoteMowK27ac, sym="SPI1", radius=80000, gstr=gg)