normalize.st: Normalize Sequence Count Data

In estfernan/boost: Bayesian Modeling of Spatial Transcriptomics Data

Normalize Sequence Count Data

Description

Obtain the relative expression levels for the gene expression count table.

Usage

normalize.st(
  count,
  scaling.method = c("TSS", "Q75", "RLE", "TMM", "A-VST", "N-VST", "log-VST")
)

Arguments

count	An n-by-p integer matrix Y that denotes the absolute gene expression count table. Each entry is the read count for gene j collected at spot i.
scaling.method	An optional character string to specify the normalization technique. The default is "TSS" for the for total sum scaling.

Details

Normalization is critical to the development of analysis techniques for any sequence count data that suffer from various sequence artifacts and biases. This method provides eight choices of normalization methods in two categories that attempt to counteract those biases due to biological and technical reasons. The first type is based on size factor estimation (i.e., TSS, Q75, RLE, TMM) where a quantity is obtained, for each sample, that captures all nuisance effects. The other type of normalization method is based on variance-stabilizing transformation (VST) (i.e., A-VST, N-VST, log-VST), which aims to transform a random variable with a negative binomial distribution into one with an approximately normal distribution.

See Jiang et al. (2021) for more information on the normalization methods.

Value

A numeric matrix that denotes the relative expression levels.

References

Jiang, X., Li, Q., & Xiao, G. (2021). Bayesian Modeling of Spatial Transcriptomics Data via a Modified Ising Model. arXiv preprint arXiv:2104.13957.

See Also

st.plot() for plotting the relative expression levels.

estfernan/boost documentation built on June 24, 2022, 12:20 a.m.