version 0.99.0

DESCRIPTION

@@ -0,0 +1,31 @@
+Type: Package
+Package: inferCSN
+Title: Inferring Cell-Specific Gene Regulatory Network
+Version: 0.99.0
+Date: 2023-10-5
+Authors@R: 
+    person("Meng", "Xu", email = "mengxu98@qq.com", role = c("aut", "cre"), comment = c(ORCID = "0000-0002-8300-1054"))
+Maintainer: Meng Xu <mengxu98@qq.com>
+Description: A method for inferring cell-specific gene regulatory network from single-cell sequencing data.
+biocViews: CellBiology, GeneExpression
+License: MIT + file LICENSE
+URL: https://mengxu98.github.io/inferCSN/
+BugReports: https://github.com/mengxu98/inferCSN/issues
+Depends: R (>= 3.3.0)
+Imports: ComplexHeatmap, data.table, doParallel, dplyr, foreach,
+        ggnetwork, ggplot2, ggraph, Matrix, methods, parallel,
+        patchwork, progress, purrr, Rcpp, stats, utils
+Suggests: circlize, gtools, igraph, precrec, pROC, testthat (>= 3.0.0),
+        tidygraph
+LinkingTo: Rcpp, RcppArmadillo
+Config/Needs/website: mengxu98/mxtemplate
+Config/testthat/edition: 3
+Encoding: UTF-8
+LazyData: true
+RoxygenNote: 7.2.3
+Language: en-US
+NeedsCompilation: yes
+Packaged: 2023-10-11 10:12:14 UTC; MSI
+Author: Meng Xu [aut, cre] (<https://orcid.org/0000-0002-8300-1054>)
+Repository: CRAN
+Date/Publication: 2023-10-11 18:20:02 UTC

LICENSE

@@ -0,0 +1,2 @@
+YEAR: 2023
+COPYRIGHT HOLDER: Meng Xu

MD5

@@ -0,0 +1,74 @@
+4f85a6b11ecd75afdcce423ae15ef005 *DESCRIPTION
+7c52ffa536a7ee404a1970079f438922 *LICENSE
+b48ef5d987897d838df5226751680a0f *NAMESPACE
+5e08e4dbc7bfa1bf575f7672aff13441 *R/RcppExports.R
+4a8c3c169d90a573af79901ddb0ca2c9 *R/auc.calculate.R
+ed324fe13967373f0ebff75d176d8d6b *R/compute.gene.rank.R
+643ade498aab0c7b5410cb03dbb9a50d *R/data.R
+569f9ac6273c770b4acbc3eb65b362fb *R/inferCSN-package.R
+716a11cb5bd17fa513da07611efc9f3b *R/inferCSN.R
+715a9aeffcf51cf83312d5c9d6ec328c *R/network.visualization.R
+b4a2e0f8aea24b0352519f018bdd63e5 *R/sparse.regression.R
+ef4b9e9703cb364732029f703ca9f1ce *R/utils.R
+82246917ca7839687f6d51eb34f8eec0 *README.md
+191a298ed75823937c77adad132f746b *data/exampleGroundTruth.rda
+975775272629b1e1438f9c89883e7728 *data/exampleMatrix.rda
+9b9c13ac9a7e999c4e4cdb362fc26e8e *man/acc.calculate.Rd
+8380f5d6933f4f8e23c17f11c7c50423 *man/auc.calculate.Rd
+ff07637def41124927f3c7f2c46f1bed *man/check.parameters.Rd
+cd8c767751171c96e056e3d80893d6e8 *man/coef.inferCSN.Rd
+18a544c85e9b0245e63f713998214767 *man/compute.gene.rank.Rd
+0e2acac6074c6d1148f7157f2166271e *man/contrast.networks.Rd
+7bead10dd1c0c7852051d259b0740edd *man/dynamic.networks.Rd
+554c357e8f5247ccc09bb4b0230833c9 *man/exampleGroundTruth.Rd
+1ec939db9239789791c034ef6bfd2265 *man/exampleMatrix.Rd
+6311a5f421c0bbf0b94debea489ffb13 *man/figures/inferCSN.svg
+34c2dd574f80faa89edea7cebdbb042c *man/inferCSN-package.Rd
+54bbfb391679a8b0ca879d51303498c4 *man/inferCSN.Rd
+814dd419ee7301a4e93848a20890ad01 *man/inferCSN.fit.Rd
+af3a9d9f4bbf39e214377b9920d4d3b0 *man/is.scalar.Rd
+0e559044dc2fe9b06b8d39132892e548 *man/net.format.Rd
+447aa592fb38dba8e88937ed73204801 *man/network.heatmap.Rd
+e9d98b090265e8645318543f37e8bf24 *man/predict.inferCSN.Rd
+1b4a2a6e2c1eed3ca4a3a0003c5c97f6 *man/print.inferCSN.Rd
+3fcd8c319587d585129534013633c858 *man/sparse.regression.Rd
+73197635a9a5df894c4f5b4e88ca4e8c *man/sub.inferCSN.Rd
+1e6750ed9d5f86b882e6a24591733b47 *src/BetaVector.cpp
+dadb1987bfb2f80d199a2e25b02a051d *src/CDL012LogisticSwaps.cpp
+64737baec03e2641f58d5681381874d6 *src/CDL012SquaredHingeSwaps.cpp
+45e55da928a9a22670c9a7380e7469b1 *src/CDL012Swaps.cpp
+b9f4edd41ee1ba186d768676b2bf79ee *src/DT2Matrix.cpp
+d4f2e65104193ad8b04b23281a39cd8c *src/Grid.cpp
+51b6baa37f71f76d69bc5e178c98b8a7 *src/Grid1D.cpp
+8748c274ffef69c8dc7f2c0f88228b22 *src/Grid2D.cpp
+09837ef4410ca3aa193eb60b7a8f681b *src/Interface.cpp
+c2fbbe606dbc01e99513a08efa77bb27 *src/Makevars
+c2fbbe606dbc01e99513a08efa77bb27 *src/Makevars.win
+41bc93afdbfbce896558b943b882d7e7 *src/Normalize.cpp
+bcd102094770a2e20628f6cb0dfc4684 *src/RcppExports.cpp
+b8da420687f2f9943625f9c8cddda799 *src/Test_Interface.cpp
+053756d599fb6942fb1959c5f54f4157 *src/include/BetaVector.h
+d10d7cbeac8dd66edb8abfe537ba2e83 *src/include/CD.h
+3006562807501591e7979d1b5af7c039 *src/include/CDL0.h
+c9262d8cc37f5ad121511ab15f1ef234 *src/include/CDL012.h
+da6865571ac3a696ef96d8d10c8b7c9a *src/include/CDL012Logistic.h
+c36020aa6b5463f99a6265a18d745473 *src/include/CDL012LogisticSwaps.h
+75978be275c2af4e57318dda1d2ee505 *src/include/CDL012SquaredHinge.h
+6acf66ba1b050b4869097aa56ed7cb10 *src/include/CDL012SquaredHingeSwaps.h
+8d625ee2d1001e5fbd827359787faa78 *src/include/CDL012Swaps.h
+3c65caa73f70506f9914640898d13c6a *src/include/CDSwaps.h
+3e191a22fe2bee08cfd53dd8cd32f8a1 *src/include/FitResult.h
+fa0e95061a24b8e548c611d4304664d0 *src/include/Grid.h
+1e34795fcfd7f91aa821223794050e00 *src/include/Grid1D.h
+5155c4b89da70421f47b4e82a0111866 *src/include/Grid2D.h
+29346ba8476c3664856ab8f5ae9ff180 *src/include/GridParams.h
+3754bebf455414fb993fd10bfc967742 *src/include/Interface.h
+68b70f6c4a8c08a8d7f6bb97746d3d49 *src/include/MakeCD.h
+aea94cb863ab3243b5c681f9d905c175 *src/include/Model.h
+8650499eeb16d4bf88397c8ed3c29c92 *src/include/Normalize.h
+060e465c4585fd95bb6ca465645051b3 *src/include/Params.h
+0816e979947d47e1e91c6327a62d641c *src/include/Test_Interface.h
+f3a3cce4a45a5306180127b28e5aa66c *src/include/utils.h
+82ebc8c76e0a000e5bf6e391edd49ff3 *src/utils.cpp
+38d99785124ecc9d873a36f2f4fff8c6 *tests/testthat.R
+33d718ae56a0c2bca1ffe606a3e8d47f *tests/testthat/test-inferCSN.R

NAMESPACE

@@ -0,0 +1,34 @@
+# Generated by roxygen2: do not edit by hand
+
+S3method(coef,inferCSN)
+S3method(coef,inferCSNCV)
+S3method(predict,inferCSN)
+S3method(predict,inferCSNCV)
+S3method(print,inferCSN)
+S3method(print,inferCSNCV)
+export(acc.calculate)
+export(auc.calculate)
+export(check.parameters)
+export(compute.gene.rank)
+export(contrast.networks)
+export(dynamic.networks)
+export(inferCSN)
+export(inferCSN.fit)
+export(is.scalar)
+export(net.format)
+export(network.heatmap)
+export(sparse.regression)
+export(sub.inferCSN)
+exportMethods(inferCSN)
+import(Matrix)
+import(ggnetwork)
+import(ggplot2)
+import(ggraph)
+import(patchwork)
+importFrom(Rcpp,evalCpp)
+importFrom(methods,as)
+importFrom(methods,is)
+importFrom(stats,coef)
+importFrom(stats,predict)
+importFrom(utils,methods)
+useDynLib(inferCSN)

R/RcppExports.R

@@ -0,0 +1,99 @@
+# Generated by using Rcpp::compileAttributes() -> do not edit by hand
+# Generator token: 10BE3573-1514-4C36-9D1C-5A225CD40393
+
+DT2Matrix <- function(weightDT) {
+    .Call('_inferCSN_DT2Matrix', PACKAGE = 'inferCSN', weightDT)
+}
+
+inferCSNFit_sparse <- function(X, y, Loss, Penalty, Algorithm, NnzStopNum, G_ncols, G_nrows, Lambda2Max, Lambda2Min, PartialSort, MaxIters, rtol, atol, ActiveSet, ActiveSetNum, MaxNumSwaps, ScaleDownFactor, ScreenSize, LambdaU, Lambdas, ExcludeFirstK, Intercept, withBounds, Lows, Highs) {
+    .Call('_inferCSN_inferCSNFit_sparse', PACKAGE = 'inferCSN', X, y, Loss, Penalty, Algorithm, NnzStopNum, G_ncols, G_nrows, Lambda2Max, Lambda2Min, PartialSort, MaxIters, rtol, atol, ActiveSet, ActiveSetNum, MaxNumSwaps, ScaleDownFactor, ScreenSize, LambdaU, Lambdas, ExcludeFirstK, Intercept, withBounds, Lows, Highs)
+}
+
+inferCSNFit_dense <- function(X, y, Loss, Penalty, Algorithm, NnzStopNum, G_ncols, G_nrows, Lambda2Max, Lambda2Min, PartialSort, MaxIters, rtol, atol, ActiveSet, ActiveSetNum, MaxNumSwaps, ScaleDownFactor, ScreenSize, LambdaU, Lambdas, ExcludeFirstK, Intercept, withBounds, Lows, Highs) {
+    .Call('_inferCSN_inferCSNFit_dense', PACKAGE = 'inferCSN', X, y, Loss, Penalty, Algorithm, NnzStopNum, G_ncols, G_nrows, Lambda2Max, Lambda2Min, PartialSort, MaxIters, rtol, atol, ActiveSet, ActiveSetNum, MaxNumSwaps, ScaleDownFactor, ScreenSize, LambdaU, Lambdas, ExcludeFirstK, Intercept, withBounds, Lows, Highs)
+}
+
+inferCSNCV_sparse <- function(X, y, Loss, Penalty, Algorithm, NnzStopNum, G_ncols, G_nrows, Lambda2Max, Lambda2Min, PartialSort, MaxIters, rtol, atol, ActiveSet, ActiveSetNum, MaxNumSwaps, ScaleDownFactor, ScreenSize, LambdaU, Lambdas, nfolds, seed, ExcludeFirstK, Intercept, withBounds, Lows, Highs) {
+    .Call('_inferCSN_inferCSNCV_sparse', PACKAGE = 'inferCSN', X, y, Loss, Penalty, Algorithm, NnzStopNum, G_ncols, G_nrows, Lambda2Max, Lambda2Min, PartialSort, MaxIters, rtol, atol, ActiveSet, ActiveSetNum, MaxNumSwaps, ScaleDownFactor, ScreenSize, LambdaU, Lambdas, nfolds, seed, ExcludeFirstK, Intercept, withBounds, Lows, Highs)
+}
+
+inferCSNCV_dense <- function(X, y, Loss, Penalty, Algorithm, NnzStopNum, G_ncols, G_nrows, Lambda2Max, Lambda2Min, PartialSort, MaxIters, rtol, atol, ActiveSet, ActiveSetNum, MaxNumSwaps, ScaleDownFactor, ScreenSize, LambdaU, Lambdas, nfolds, seed, ExcludeFirstK, Intercept, withBounds, Lows, Highs) {
+    .Call('_inferCSN_inferCSNCV_dense', PACKAGE = 'inferCSN', X, y, Loss, Penalty, Algorithm, NnzStopNum, G_ncols, G_nrows, Lambda2Max, Lambda2Min, PartialSort, MaxIters, rtol, atol, ActiveSet, ActiveSetNum, MaxNumSwaps, ScaleDownFactor, ScreenSize, LambdaU, Lambdas, nfolds, seed, ExcludeFirstK, Intercept, withBounds, Lows, Highs)
+}
+
+cor_matrix <- function(p, base_cor) {
+    .Call('_inferCSN_cor_matrix', PACKAGE = 'inferCSN', p, base_cor)
+}
+
+R_matrix_column_get_dense <- function(mat, col) {
+    .Call('_inferCSN_R_matrix_column_get_dense', PACKAGE = 'inferCSN', mat, col)
+}
+
+R_matrix_column_get_sparse <- function(mat, col) {
+    .Call('_inferCSN_R_matrix_column_get_sparse', PACKAGE = 'inferCSN', mat, col)
+}
+
+R_matrix_rows_get_dense <- function(mat, rows) {
+    .Call('_inferCSN_R_matrix_rows_get_dense', PACKAGE = 'inferCSN', mat, rows)
+}
+
+R_matrix_rows_get_sparse <- function(mat, rows) {
+    .Call('_inferCSN_R_matrix_rows_get_sparse', PACKAGE = 'inferCSN', mat, rows)
+}
+
+R_matrix_vector_schur_product_dense <- function(mat, u) {
+    .Call('_inferCSN_R_matrix_vector_schur_product_dense', PACKAGE = 'inferCSN', mat, u)
+}
+
+R_matrix_vector_schur_product_sparse <- function(mat, u) {
+    .Call('_inferCSN_R_matrix_vector_schur_product_sparse', PACKAGE = 'inferCSN', mat, u)
+}
+
+R_matrix_vector_divide_dense <- function(mat, u) {
+    .Call('_inferCSN_R_matrix_vector_divide_dense', PACKAGE = 'inferCSN', mat, u)
+}
+
+R_matrix_vector_divide_sparse <- function(mat, u) {
+    .Call('_inferCSN_R_matrix_vector_divide_sparse', PACKAGE = 'inferCSN', mat, u)
+}
+
+R_matrix_column_sums_dense <- function(mat) {
+    .Call('_inferCSN_R_matrix_column_sums_dense', PACKAGE = 'inferCSN', mat)
+}
+
+R_matrix_column_sums_sparse <- function(mat) {
+    .Call('_inferCSN_R_matrix_column_sums_sparse', PACKAGE = 'inferCSN', mat)
+}
+
+R_matrix_column_dot_dense <- function(mat, col, u) {
+    .Call('_inferCSN_R_matrix_column_dot_dense', PACKAGE = 'inferCSN', mat, col, u)
+}
+
+R_matrix_column_dot_sparse <- function(mat, col, u) {
+    .Call('_inferCSN_R_matrix_column_dot_sparse', PACKAGE = 'inferCSN', mat, col, u)
+}
+
+R_matrix_column_mult_dense <- function(mat, col, u) {
+    .Call('_inferCSN_R_matrix_column_mult_dense', PACKAGE = 'inferCSN', mat, col, u)
+}
+
+R_matrix_column_mult_sparse <- function(mat, col, u) {
+    .Call('_inferCSN_R_matrix_column_mult_sparse', PACKAGE = 'inferCSN', mat, col, u)
+}
+
+R_matrix_normalize_dense <- function(mat_norm) {
+    .Call('_inferCSN_R_matrix_normalize_dense', PACKAGE = 'inferCSN', mat_norm)
+}
+
+R_matrix_normalize_sparse <- function(mat_norm) {
+    .Call('_inferCSN_R_matrix_normalize_sparse', PACKAGE = 'inferCSN', mat_norm)
+}
+
+R_matrix_center_dense <- function(mat, X_normalized, intercept) {
+    .Call('_inferCSN_R_matrix_center_dense', PACKAGE = 'inferCSN', mat, X_normalized, intercept)
+}
+
+R_matrix_center_sparse <- function(mat, X_normalized, intercept) {
+    .Call('_inferCSN_R_matrix_center_sparse', PACKAGE = 'inferCSN', mat, X_normalized, intercept)
+}
+

R/auc.calculate.R

@@ -0,0 +1,128 @@
+#' @title AUC value calculate
+#'
+#' @param weightDT The weight data table of network
+#' @param groundTruth Ground truth for calculate AUC
+#' @param plot If true, draw and print figure of AUC
+#' @param lineColor The color of line in the figure
+#' @param lineWidth The width of line in the figure
+#'
+#' @import patchwork
+#' @import ggplot2
+#'
+#' @return AUC values and figure
+#' @export
+#'
+#' @examples
+#' library(inferCSN)
+#' data("exampleMatrix")
+#' data("exampleGroundTruth")
+#' weightDT <- inferCSN(exampleMatrix)
+#' auc <- auc.calculate(weightDT, exampleGroundTruth, plot = TRUE)
+#' head(auc)
+#'
+auc.calculate <- function(weightDT,
+                          groundTruth,
+                          plot = FALSE,
+                          lineColor = "#1563cc",
+                          lineWidth = 1) {
+  # Check input data
+  colnames(weightDT) <- c("regulator", "target", "weight")
+  weightDT$weight <- abs(as.numeric(weightDT$weight))
+
+  if (ncol(groundTruth) > 2) groundTruth <- groundTruth[, 1:2]
+  names(groundTruth) <- c("regulator", "target")
+  groundTruth$label <- rep(1, nrow(groundTruth))
+
+  gold <- merge(weightDT, groundTruth,
+                by = c("regulator", "target"),
+                all.x = TRUE)
+  gold$label[is.na(gold$label)] <- 0
+
+  aucCurves <- precrec::evalmod(scores = gold$weight,
+                                labels = gold$label)
+
+  auc <- attr(aucCurves, "auc")
+
+  aucMetric <- data.frame(AUROC = rep(0.000, 1),
+                          AUPRC = rep(0.000, 1),
+                          ACC = rep(0.000, 1))
+  aucMetric[1, "AUROC"] <- sprintf("%0.3f", auc$aucs[1])
+  aucMetric[1, "AUPRC"] <- sprintf("%0.3f", auc$aucs[2])
+  aucMetric[1, "ACC"] <- sprintf("%0.3f", acc.calculate(gold))
+  if (plot) {
+    # Separate data
+    aurocDf <- subset(fortify(aucCurves),
+                      curvetype == "ROC")
+    auprcDf <- subset(fortify(aucCurves),
+                      curvetype == "PRC")
+
+    # Plot
+    auroc <- ggplot(aurocDf, aes(x = x, y = y)) +
+      geom_line(color = lineColor,
+                linewidth = lineWidth) +
+      geom_abline(slope = lineWidth,
+                  color = lineColor,
+                  linetype = "dotted",
+                  linewidth = lineWidth) +
+      labs(title = paste("AUROC:", aucMetric[1]),
+           x = "False positive rate",
+           y = "True positive rate") +
+      xlim(0, 1) +
+      ylim(0, 1) +
+      coord_fixed() +
+      theme_bw()
+
+    auprc <- ggplot(auprcDf, aes(x = x, y = y)) +
+      geom_line(color = lineColor,
+                linewidth = 1) +
+      labs(title = paste("AUPRC:", aucMetric[2]),
+           x = "Recall",
+           y = "Precision") +
+      xlim(0, 1) +
+      ylim(0, 1) +
+      coord_fixed() +
+      theme_bw()
+
+    # Combine two plots by `patchwork` package
+    p <- auroc + auprc
+    print(p)
+  }
+
+  return(aucMetric)
+}
+
+#' ACC calculate
+#'
+#' @param gold Data
+#'
+#' @return ACC
+#' @export
+#'
+acc.calculate <- function(gold) {
+  results <- pROC::roc(gold$label ~ gold$weight,
+                       direction = "<",
+                       levels = c(0, 1))
+
+  # After this operation, '0' indicate positive
+  reverseLabel <- 2 - as.numeric(as.factor(gold$label))
+
+  sensitivities <- results$sensitivities
+  specificities <- results$specificities
+  selectValue <- sensitivities + specificities - 1
+  cutValueResults <- results$thresholds[selectValue == max(selectValue)]
+  selectSensitivities <- sensitivities[selectValue == max(selectValue)]
+
+  cutValue <- cutValueResults[selectSensitivities == max(selectSensitivities)]
+
+  predictorBinary <- rep(0, length(results$predictor))
+  predictorBinary[results$predictor >= cutValue] <- 1
+  predictorBinary <- as.factor(predictorBinary)
+  levels(predictorBinary) <- c("0", "1")
+  predictorBinary <- factor(predictorBinary, levels = c(1, 0))
+
+  pre <- as.vector(table(predictorBinary, reverseLabel))
+
+  acc <- (pre[1] + pre[4]) / sum(pre)
+
+  return(acc)
+}