SummarizedExperiment object

In bigPint: Big multivariate data plotted interactively

knitr::opts_chunk$set(echo=TRUE)

SummarizedExperiment object

Researchers who prefer using the SummarizedExperiment class can feed their data and associated metrics into the bigPint methods using one SummarizedExperiment object (instead of both a data and dataMetrics object). We demonstrate here how to create a SummarizedExperiment object. In the remaining articles of the vignette, the example code uses the data and dataMetrics objects as example input. However, at the bottom of each vignette, we also include the corresponding example code that uses instead the SummarizedExperiment object to create the exact same plots.

---

Example: two groups

As was shown in the article Data object, the data object called soybean_ir_sub contained 5,604 genes and two treatment groups, N and P [@soybeanIR]. We can create a SummarizedExperiment object that combines aspects of the data object and dataMetrics object. We start by reading in the data and dataMetrics objects (created from edgeR in the article Data metrics object.

library(bigPint)
library(DelayedArray)
library(SummarizedExperiment)

data(soybean_ir_sub)
data = soybean_ir_sub
data(soybean_ir_sub_metrics)
dataMetrics = soybean_ir_sub_metrics

We then convert the dataMetrics object from a list of dataframes into one dataframe.

dMUnlist <- dataMetrics
dMUnlist[-1] <- lapply(dMUnlist[-1], transform, ID = NULL)
dMUnlist <- do.call(cbind, dMUnlist)
names(dMUnlist)[1] <- "ID"

Now, we can tranform our data object into a DelayedMatrix class object and then input both our data and dataMetrics objects combined into a SummarizedExperiment class object. We can verify that the assay() and rowData() methods work for accessing our SummarizedExperiment object.

rownames(data) = data[,1]
data = data[,-1]
data <- DelayedArray(data)
se_soybean_ir_sub <- SummarizedExperiment(assays = data, rowData = dMUnlist)
assay(se_soybean_ir_sub)
rowData(se_soybean_ir_sub)

---

Example: three groups

Similarly, as was shown in the data page, the data object called soybean_cn_sub contained 7,332 genes and three treatment groups, S1, S2, and S3 [@brown2015developmental]. We can create a SummarizedExperiment object that combines aspects of the data object and dataMetrics object. We start by reading in the data and dataMetrics objects (created from edgeR in the article Data metrics object.

data(soybean_cn_sub)
data = soybean_cn_sub
data(soybean_cn_sub_metrics)
dataMetrics = soybean_cn_sub_metrics

We then convert the dataMetrics object from a list of dataframes into one dataframe.

dMUnlist <- dataMetrics
dMUnlist[-1] <- lapply(dMUnlist[-1], transform, ID = NULL)
dMUnlist <- do.call(cbind, dMUnlist)
names(dMUnlist)[1] <- "ID"

Now, we can tranform our data object into a DelayedMatrix class object and then input both our data and dataMetrics objects combined into a SummarizedExperiment class object. We can verify that the assay() and rowData() methods work for accessing our SummarizedExperiment object.

rownames(data) = data[,1]
data = data[,-1]
data <- DelayedArray(data)
se_soybean_cn_sub <- SummarizedExperiment(assays = data, rowData = dMUnlist)
assay(se_soybean_cn_sub)
rowData(se_soybean_cn_sub)

---

References

bigPint documentation built on Nov. 8, 2020, 5:07 p.m.