Signature is a string which indicate the age signature to use when calculating RNA age. This matching process is automatic so that the users just need to provide gene expression data without having to pull out the internal coefficients. The package takes gene expression data as input and then match the input genes to the genes in the internal data. We saved the pre-trained model coefficients as internal data in the package. For each tissue, signature and sample type (see below for the descriptions), we pre-trained the calculator using elastic net based on the GTEx samples. A detailed explanation can be found in our paper. (De Magalhães, Curado, and Church 2009) to find the genes with the largest number of age-related signals. To identify genes consistently differentially expressed across tissues, we adapted the binomial test discussed in de Magalhaes et al. The across-tissue calculator was constructed by first performing differential expression analysis on the RNASeq counts data for each individual tissue. Chronological age was response variable whereas logarithm transformed FPKM of genes were predictors. FPKM values were calculated for each individual sample using getRPKM function in Bioconductor package recount.įor the tissue-specific RNASeq age calculator, elastic net (Zou and Hastie 2005) algorithm was used to train the predictors for each individual tissue. 2017) version of GTEx data, where all samples were processed with the same analytical pipeline.
To facilitate integrated analysis and direct comparison of multiple datasets, we utilized recount2 (Collado-Torres et al. To avoid the influence of tumor on gene expression, the 102 tumor samples from GTEx V6 release are dropped and the remaining 9,560 samples were used in the subsequent analysis. GTEx V6 release contains gene expression data at gene, exon, and transcript level of 9,662 samples from 30 different tissues. GTEx is a public available genetic database for studying tissue specific gene expression and regulation. We utilized the GTEx data to construct our across-tissue and tissue-specific transcriptional age calculator.
0 kommentar(er)
