Initial commit

.Rbuildignore

@@ -0,0 +1,3 @@
+^.*\.Rproj$
+^\.Rproj\.user$
+^docs$

.gitignore

@@ -0,0 +1,7 @@
+.Rproj.user
+.Rhistory
+.RData
+.Ruserdata
+src/*.o
+src/*.so
+src/*.dll

DESCRIPTION

@@ -0,0 +1,20 @@
+Package: cytostats
+Type: Package
+Title: Statistical Functions for Single-Cell Analysis
+Version: 0.0.0.9000
+Date: 2018-02-13
+Authors@R: c(person("Mohammad", "Rohban", email = "[email protected]", role = c("aut", "cre")),
+    person("Shantanu", "Singh", email = "[email protected]", role = c("aut"))
+  )
+Description: Single-cell analysis requires efficient implementation of 
+    statistical functions to handle the high-dimensionality of the data.
+License: BSD_3_clause + file LICENSE
+Imports: 
+    doParallel (>= 1.0.10),
+    foreach (>= 1.4.3),
+    magrittr (>= 1.5),
+    Rcpp (>= 0.12.15)
+Suggests:
+    testthat (>= 1.0.2)
+LinkingTo: Rcpp, RcppArmadillo
+RoxygenNote: 6.0.1

LICENSE

@@ -0,0 +1,3 @@
+YEAR: 2018
+COPYRIGHT HOLDER: Broad Institute, Inc.
+ORGANIZATION: Broad Institute, Inc.

NAMESPACE

@@ -0,0 +1,12 @@
+# Generated by roxygen2: do not edit by hand
+
+export(combine_cov_estimates)
+export(online_covar)
+export(parallel_cor)
+export(parallel_cov)
+export(two_pass_multi_covar)
+importFrom(Rcpp,sourceCpp)
+importFrom(foreach,"%dopar%")
+importFrom(magrittr,"%<>%")
+importFrom(magrittr,"%>%")
+useDynLib(cytostats)

R/RcppExports.R

@@ -0,0 +1,38 @@
+# Generated by using Rcpp::compileAttributes() -> do not edit by hand
+# Generator token: 10BE3573-1514-4C36-9D1C-5A225CD40393
+
+#' online_covar
+#'
+#' @param x1 a numeric vector containing samples of the first random variable
+#' @param x2 a numeric vector containing corresponding samples of the second random variable
+#'
+#' @export
+#'
+online_covar <- function(x1, x2) {
+    .Call('_cytostats_online_covar', PACKAGE = 'cytostats', x1, x2)
+}
+
+#' two_pass_multi_covar
+#'
+#' @param s a data matrix whose column covariances are sought
+#'
+#' @export
+#'
+two_pass_multi_covar <- function(s) {
+    .Call('_cytostats_two_pass_multi_covar', PACKAGE = 'cytostats', s)
+}
+
+#' combine_cov_estimates
+#'
+#' @param batch_mean_cov the matrix which contains estimated means and covariance for each batch of data.
+#'         For n variable, and k batches, it is (n+1)*(n.k) size, with the first row being the means and
+#'         rest of the rows being the covariance matrices. Covariance matrices were concatenated column-wise
+#'         resulting in n.k columns.
+#' @param b a vector containing number of samples in each batch of data
+#'
+#' @export
+#'
+combine_cov_estimates <- function(batch_mean_cov, b) {
+    .Call('_cytostats_combine_cov_estimates', PACKAGE = 'cytostats', batch_mean_cov, b)
+}
+

R/cytostats-package.r

@@ -0,0 +1,9 @@
+#' cytostats
+#'
+#' @name cytostats
+#' @docType package
+NULL
+
+#' @useDynLib cytostats
+#' @importFrom Rcpp sourceCpp
+NULL

R/parallel_cor.R

@@ -0,0 +1,29 @@
+#' Compute correlation function in batches, in parallel
+#'
+#' @param x data matrix
+#' @param splits shows the number of splits of data that we want to combine  (should be of 2^n form)
+#' @param cores number of cores that will be used for the parallel computation
+#' @param cov_fun covariance function used on each split
+#'
+#' @export
+#'
+#' @importFrom magrittr %>%
+#' @importFrom magrittr %<>%
+#' @importFrom foreach %dopar%
+#'
+parallel_cor <- function(x, splits = 2, cores = 2, cov_fun = "two_pass_multi_covar") {
+
+  cov_mat <-as.matrix(parallel_cov(x, splits = splits, cores = cores, cov_fun = cov_fun))
+
+  # \cor(X, Y) = \frac{\cov(X, Y)}{\sd(X) \sd(Y)}
+
+  inv_sd <- as.matrix(diag(diag(cov_mat)^-0.5))
+
+  result <- inv_sd %*% cov_mat %*% inv_sd
+
+  rownames(result) <- colnames(x)
+
+  colnames(result) <- colnames(x)
+
+  result
+}

R/parallel_cov.R

@@ -0,0 +1,57 @@
+#' Compute covariance function in batches, in parallel
+#'
+#' @param x data matrix
+#' @param splits shows the number of splits of data that we want to combine  (should be of 2^n form)
+#' @param cores number of cores that will be used for the parallel computation
+#' @param cov_fun covariance function used on each split
+#'
+#' @export
+#'
+#' @importFrom magrittr %>%
+#' @importFrom magrittr %<>%
+#' @importFrom foreach %dopar%
+#'
+parallel_cov <- function(x, splits = 2, cores = 2, cov_fun = "two_pass_multi_covar") {
+
+  x <- as.matrix(x)
+
+  n <- NROW(x)
+
+  doParallel::registerDoParallel(cores = cores)
+
+  batches <- rep(round(n/splits), splits - 1)
+
+  batches <- c(batches, n - sum(batches))
+
+  j <- 0 # to avoid warning no visible binding for global variable j
+
+  result <- foreach::foreach(j = 1:length(batches)) %dopar% {
+    if (j == 1) {
+      i <- 1
+
+    } else {
+      i <- sum(batches[1:(j-1)]) + 1
+
+    }
+
+    k <- sum(batches[1:j])
+
+    batch_mean <- apply(x[i:k,], 2, mean)
+
+    batch_cov <- do.call(cov_fun, list(x[i:k, ]))
+
+    rbind(batch_mean, batch_cov)
+  }
+
+  batch_mean_cov <- do.call(cbind, result)
+
+  mean_cov <- combine_cov_estimates(batch_mean_cov, batches)
+
+  result <- mean_cov[2:NROW(mean_cov), ]
+
+  rownames(result) <- colnames(x)
+
+  colnames(result) <- colnames(x)
+
+  result
+}

cytostats.Rproj

@@ -0,0 +1,18 @@
+Version: 1.0
+
+RestoreWorkspace: Default
+SaveWorkspace: Default
+AlwaysSaveHistory: Default
+
+EnableCodeIndexing: Yes
+UseSpacesForTab: Yes
+NumSpacesForTab: 2
+Encoding: UTF-8
+
+RnwWeave: Sweave
+LaTeX: pdfLaTeX
+
+BuildType: Package
+PackageUseDevtools: Yes
+PackageInstallArgs: --no-multiarch --with-keep.source
+PackageRoxygenize: rd,collate,namespace