getSimScores {BrainSABER} | R Documentation |
This function computes the similarity score of a gene expression
vector returned by getExternalVector
or a trimmed data set returned
by getTrimmedExternalSet
, compared to a subset of AIBSARNA,
obtained by getRelevantGenes
.
getSimScores(data, relevantGenes = NULL, similarity_method = "cosine")
data |
a named vector of gene expression values returned by
|
relevantGenes |
a SummarizedExperiment object created using the
|
similarity_method |
currently supported similarity methods are "cosine" and "euclidean", defaults to "cosine" |
If data
is a vector, returns a vector of similarity scores
for each sample in relevantGenes. If data
is a
SummarizedExperiment, returns a data frame, with columns containing the
similarity scores for and named after each sample in data
, and
rows named after each sample in relevantGenes
. If data
is
a CellScabbard, the results will be stored in its similarityScores slot.
AIBSARNA <- buildAIBSARNA(mini = TRUE) # Example 1 - using CellScabbard class # get a random sample of 3 genes totalGenes <- nrow(AIBSARNA) gene_idx <- sample.int(totalGenes, 3) sample_idx <- c(1,3,5) # Subset AIBSARNA exprs <- assay(AIBSARNA)[gene_idx, sample_idx] fd <- rowData(AIBSARNA)[gene_idx, ] pd <- colData(AIBSARNA)[sample_idx, ] # build a trimmed data set myGenes <- CellScabbard(exprsData = exprs, phenoData = pd, featureData = fd, AIBSARNA = AIBSARNA, autoTrim = TRUE) similarityScores(myGenes) <- getSimScores(data = myGenes, similarity_method = "cosine") similarityScores(myGenes) similarityScores(myGenes) <- getSimScores(data = myGenes, similarity_method = "euclidean") similarityScores(myGenes) # Example 2 - manual gene selection and relevant gene extraction myGenes <- c(4.484885, 0.121902, 0.510035) names(myGenes) <- c("TSPAN6", "DPM1", "C1orf112") myGeneSet <- getRelevantGenes(myGenes, AIBSARNA = AIBSARNA, AIBSARNAid = "gene_symbol") CosScores <- getSimScores(myGenes, myGeneSet, similarity_method = "cosine") EucScores <- getSimScores(myGenes, myGeneSet, similarity_method = "euclidean")