AdaTiSS is a R package to calculate tissue specificity scores, developed in the paper AdaTiSS: A Novel Data-Adaptive Robust Method for Quantifying Tissue Specificity Scores, published in Bioinformatics, 2021.

Please contact Meng Wang by email [email protected] for questions.

• R (version >= 3.3.0)

To illustrate the usage of our package, we took a small set of gene expression in raw TPM from GTEx project in version 7 as an example.

The data folder includes raw TPM expression for 2000 genes from 181 samples across 32 tissues and a pheonotype info of the tissue type for each sample.

To call the AdaTiSS source function

source('./R/AdaTiSS_fn.R')

To load expression data

dat.rna = readRDS(file='./data/raw_tpm_expression.rds')

To preprocess raw data and obtain filtered gene expression in log scale (for more options, see the file 'AdaTiSS_fn.R')

X = preproc.filter.fn (dat.rna, dat.type = "TPM or RPKM", proc.zero = 'ceiled to 1', filter.col.prp = 1, exp.thres=1)

To load phenotype data

p.dat = read.csv('./data/sample_phenotype.csv')

To obtain gene expression in tissue level

tiss.abd = tiss.abd.fn(X, p.dat)

To call AdaTiSS (for more options, see the file 'AdaTiSS_fn.R')

result = AdaTiSS(X, tiss.abd=tiss.abd, adjust=TRUE, adjust.opt=0)

Output:

head(result$ada.s) --- sample normalized scores

head(result$ada.z) --- tissue specificity scores

head(result$pop.fit.mx) --- population fitting info

For more statistical analysis, to check our AdaReg package.

For a related robust normalization work, to check our RobNorm package.