justlinc-methods {LINC} | R Documentation |
justlinc
is a wrapper that uses different functions of the LINC package. It applies fixed thresholds for gene selection and requires only an (unprocessed) expression matrix as input. This enables a fast co-expression analysis with or without a list candidates.
justlinc(object, targetGenes = "lincRNA", rmPC = TRUE)
object |
a gene expresssion matrix with Ensembl gene ids (rows correspond to genes) |
targetGenes |
the gene biotype like "lincRNA" or a vector of querie(s) - one or multiple genes ids |
rmPC |
remove 30 percent of the variance in the data by PCA |
This function was built for expression matrices which uses the Ensembl gene system with gene expression values for over 50.000 genes. The input will be matched with a static gene annotation. Genes will be selected for median expression and variance. The final correlation matrix of protein-coding genes versus the target genes (lincRNAs) considers 5000 protein-coding genes and 500 target genes. Enriched pathways will be computed for the co-expressed genes showing the highest Spearman's rank correlation.
In case targetGenes = "lincRNA"
, then the function will search for the 10 best lincRNAs. Supplying a vector of gene ids the method will determine the best co-expressed genes for the given queries.
Importanly, this function is only a wrapper. Thresholds can be changed using linc
and related functions.
depending on the input in targetGenes
one or two plots and the result of the co-expression analysis
signature(object = "matrix")
(see details)
Manuel Goepferich
linc
# NOT RUN: # large gene expression matrix not avaiable in this version # data(LIVER_EXPR) # a gene expression matrix with > 50,000 genes # str(GTEX_LIVER_CRUDE) # 'justlinc' will search for the 10 best candidates try(justlinc(GTEX_LIVER_CRUDE), silent = TRUE) # 'justlinc' called with queries my_lincRNAs <- c("ENSG00000224153", "ENSG00000197813", "ENSG00000179136", "ENSG00000259439", "ENSG00000267462") try(justlinc(GTEX_LIVER_CRUDE, targetGenes = my_lincRNAs), silent = TRUE)