# Generate all the signature matrices one time with the option to leave out half of the data as a test set This wrapper is helpful for repetitively matrix generation. It generates seed matrix, all-gene matrix, augmented matrix, shrunk matrix, and all the clustered matrices in one call. ```r testAllSigMatrices( exprData, randomize = TRUE, skipShrink = FALSE, proportional = FALSE, handMetaCluster = NULL, testOnHalf = TRUE, condTol = 1.01, numChunks = 100, plotIt = TRUE, fastStop = TRUE, singleCore = TRUE ) ``` ## Arguments - `exprData`: The gene express data. Each row is a gene, and each column is an example of a particular cell type. - `randomize`: Set to to TRUE randomize the sets selected by ADAPTS::scSample (DEFAULT: TRUE) - `skipShrink`: Set to TRUE to skip shrinking the signatrure matrix (DEFAULT: TRUE) - `proportional`: Set to true to make the training set cell type proportional. Ignores group size (DEFAULT: FALSE) - `handMetaCluster`: A List of pre-defined meta clusters.Set to NULL to automatically group indistinguishable cells into same cluster using clustWspillOver.(DEFAULT: NULL) - `testOnHalf`: Set to TRUE to leave half the data as a test set - `condTol`: The tolerance in the reconstruction algorithm. 1.0 = no tolerance, 1.05 = 5% tolerance (DEFAULT: 1.01) - `numChunks`: The number of groups of genes to remove while shrinking (DEFAULT: NULL, i.e. 1) - `plotIt`: Set to FALSE to suppress plots (DEFAULT: TRUE) - `fastStop`: Halt early when the condition number changes by less than 1 for 3 iterations (DEFAULT: TRUE) - `singleCore`: TRUE for a single core (DEFAULT: TRUE) ## Returns A list of results including prediction accuracy and cell enrichment ## Examples ```r ct1 <- runif(1000, 0, 100) ct2 <- runif(1000, 0, 100) ct3 <- runif(1000, 0, 100) ct4 <- runif(1000, 0, 100) dataMat <- cbind(ct1, ct1, ct1, ct1, ct1, ct1, ct2, ct2, ct2, ct2, ct3, ct3, ct3,ct3,ct4,ct4) rownames(dataMat) <- make.names(rep('gene', nrow(dataMat)), unique=TRUE) noise <- matrix(runif(nrow(dataMat)*ncol(dataMat), -2, 2), nrow = nrow(dataMat), byrow = TRUE) dataMat <- dataMat + noise metaList <- list() colnames(dataMat) <- sub('\\..*','', colnames(dataMat)) metaList[[1]] <- c(unique(colnames(dataMat))[1]) #Cell Type 1 metaList[[2]] <- c(unique(colnames(dataMat))[2]) #Cell Type 2 metaList[[3]] <- c(unique(colnames(dataMat))[3]) #Cell Type 3 metaList[[4]] <- c(unique(colnames(dataMat))[4:length(unique(colnames(dataMat)))]) #Cell Type 4 #options(mc.cores=2) # This is a meta-function that calls other functions, # The execution speed is too slow for the CRAN automated check #testAllSigMatrices(exprData=dataMat, randomize = TRUE, skipShrink=FALSE, # proportional=FALSE, handMetaCluster=metaList, testOnHalf=TRUE, numChunks=NULL) ```

testAllSigMatrices function