BatchQC Examples

Solaiappan Manimaran

2022-04-26

Example 1: Simulated dataset

The simulated data consists of three batches and two conditions, and expression measurements for 50 genes

library(BatchQC)
nbatch <- 3
ncond <- 2
npercond <- 10
data.matrix <- rnaseq_sim(ngenes=50, nbatch=nbatch, ncond=ncond, npercond=
    npercond, basemean=10000, ggstep=50, bbstep=2000, ccstep=800, 
    basedisp=100, bdispstep=-10, swvar=1000, seed=1234)
batch <- rep(1:nbatch, each=ncond*npercond)
condition <- rep(rep(1:ncond, each=npercond), nbatch)
batchQC(data.matrix, batch=batch, condition=condition, 
        report_file="batchqc_report.html", report_dir=".", 
        report_option_binary="111111111",
        view_report=FALSE, interactive=TRUE, batchqc_output=TRUE)

Example 2: Real signature dataset

This dataset is from signature data captured when activating different growth pathway genes in human mammary epithelial cells (GEO accession: GSE73628). This data consists of three batches and ten different conditions corresponding to control and nine different pathways

library(BatchQC)
data(example_batchqc_data)
batch <- batch_indicator$V1
condition <- batch_indicator$V2
batchQC(signature_data, batch=batch, condition=condition, 
        report_file="batchqc_signature_data_report.html", report_dir=".", 
        report_option_binary="111111111",
        view_report=FALSE, interactive=TRUE)

Example 3: Real bladderbatch dataset

This dataset has 5 batches, 3 covariate levels. Batch 1 contains only cancer, 2 and 3 have cancer and controls, 4 contains only biopsy, and 5 contains cancer and biopsy

library(BatchQC)
library(bladderbatch)
data(bladderdata)
pheno <- pData(bladderEset)
edata <- exprs(bladderEset)
batch <- pheno$batch  
condition <- pheno$cancer
batchQC(edata, batch=batch, condition=condition, 
        report_file="batchqc_report.html", report_dir=".", 
        report_option_binary="111111111",
        view_report=FALSE, interactive=TRUE)

Example 4: Real protein expression dataset

This dataset is from protein expression data captured for 39 proteins. It has two batches and two conditions corresponding to case and control.

library(BatchQC)
data(protein_example_data)
batchQC(protein_data, protein_sample_info$Batch, protein_sample_info$category,
        report_file="batchqc_protein_data_report.html", report_dir=".", 
        report_option_binary="111111111",
        view_report=FALSE, interactive=TRUE)

Example 5: Second simulated dataset with only batch variance difference

The simulated data consists of three batches and two conditions, and expression measurements for 50 genes. In this dataset, there is no difference in batch mean but only difference in batch variance from batch to batch.

library(BatchQC)
nbatch <- 3
ncond <- 2
npercond <- 10
data.matrix <- rnaseq_sim(ngenes=50, nbatch=nbatch, ncond=ncond, npercond=
    npercond, basemean=5000, ggstep=50, bbstep=0, ccstep=2000, 
    basedisp=10, bdispstep=-4, swvar=1000, seed=1234)

### apply BatchQC
batch <- rep(1:nbatch, each=ncond*npercond)
condition <- rep(rep(1:ncond, each=npercond), nbatch)
batchQC(data.matrix, batch=batch, condition=condition, 
        report_file="batchqc_report.html", report_dir=".", 
        report_option_binary="111111111",
        view_report=FALSE, interactive=TRUE, batchqc_output=TRUE)