@@ -1,5 +1,5 @@

 Package: crisprScore

-Version: 1.1.11

+Version: 1.1.12

 Date: 2022-07-23

 Title: On-Target and Off-Target Scoring Algorithms for CRISPR gRNAs

 Authors@R: c(

@@ -34,8 +34,6 @@

 #' @param fastaFile String specifying fasta file of the hg38 genome.

 #' @param chromatinFiles Named character vector of length 3 specifying

 #'     BigWig files containing chromatin accessibility data. 

-#' @param fork Set to \code{TRUE} to preserve changes to the R

-#'     configuration within the session.

 #'     

 #' @details \code{tss_df} details:

 #'     This must be a \code{data.frame} that contains the following columns:

@@ -86,11 +84,8 @@ getCrispraiScores <- function(tss_df,

                               verbose=FALSE,

                               modality=c("CRISPRa", "CRISPRi"),

                               fastaFile=NULL,

-                              chromatinFiles=NULL,

-                              fork=FALSE

+                              chromatinFiles=NULL

 ){

-

-

     .checkChromatinFiles(chromatinFiles)

     .checkFastaFile(fastaFile)

     .checkTssFrame(tss_df)

@@ -101,165 +96,33 @@ getCrispraiScores <- function(tss_df,

                                    sgrna_df,

                                    verbose=verbose)

-    #dir <- system.file("crisprai",

-    #                   "temp_data",

-    #                   package="crisprScore",

-    #                   mustWork=TRUE)

-    #dir <- "/Users/fortinj2/crisprScore/inst/crisprai/temp_data"

-    #pickleFile <- paste0(dir, "/", modality, "_model.pkl")

-

     if (modality=="CRISPRa"){

         pickleFile <- crisprScoreData::CRISPRa_model.pkl()

     } else if (modality=="CRISPRi"){

         pickleFile <- crisprScoreData::CRISPRi_model.pkl()

     }

-

-    results <- basiliskRun(env=env_crisprai,

-                           shared=FALSE,

-                           fork=fork,

-                           fun=.pyPredictWeissmanScore,

-                           modality=modality,

-                           tssTable=inputList[["tssTable"]],

-                           p1p2Table=inputList[["p1p2Table"]],

-                           sgrnaTable=inputList[["sgrnaTable"]],

-                           libraryTable=inputList[["libraryTable"]],

-                           pickleFile=pickleFile,

-                           fastaFile=fastaFile,

-                           chromatinFiles=chromatinFiles,

-                           verbose=verbose)

-

+    results <- .pyPredictWeissmanScore(modality=modality,

+                                       tssTable=inputList[["tssTable"]],

+                                       p1p2Table=inputList[["p1p2Table"]],

+                                       sgrnaTable=inputList[["sgrnaTable"]],

+                                       libraryTable=inputList[["libraryTable"]],

+                                       pickleFile=pickleFile,

+                                       fastaFile=fastaFile,

+                                       chromatinFiles=chromatinFiles,

+                                       verbose=verbose)

+    rownames(results) <- sgrna_df[["grna_id"]]

     return(results)

 }

-.checkFastaFile <- function(fasta){

-    if (is.null(fasta)){

-        stop("Argument fasta cannot be NULL.")

-    }

-    if (length(fasta)!=1){

-        stop("fasta must be a character vector of length 1.")

-    }

-     if (!file.exists(fasta)){

-        stop("Fasta file cannot be found.")

-    }

-    invisible(TRUE)

-}

-

-.checkChromatinFiles <- function(chromatinFiles){

-    if (is.null(chromatinFiles)){

-        stop("Argument chromatinFiles cannot be NULL.")

-    }

-    if (length(chromatinFiles)!=3){

-        stop("chromatinFiles must be a character vector of length 3.")

-    }

-    choices <- c("mnase", "faire", "dnase")

-    if (!all(sort(names(chromatinFiles))==sort(names(choices)))){

-        stop("chromatinFiles must be a character vector with the",

-             " following names: mnase, faire, dnase.")

-    }

-    if (!sum(file.exists(chromatinFiles))==3){

-        stop("Some of the chromatin files cannot be found.")

-    }

-    invisible(TRUE)

-}

-

-

-

-

-

-

-.checkTssFrame <- function(tssFrame){

-    cols <- c("tss_id",

-              "gene_symbol",

-              "promoter",

-              "transcripts",

-              "position",

-              "strand",

-              "chr")

-    if (!all(cols %in% colnames(tssFrame))){

-        choices <- setdiff(cols, colnames(tssFrame))

-        stop("The following columns are missing in the tssFrame: \n \t",

-             paste0(choices, collapse=", "),".")

-    }

-    

-    if (sum(is.na(tssFrame$tss_id))>0){

-        stop("tss_id has some missing values.")

-    }

-    if (sum(is.na(tssFrame$gene_symbol))>0){

-        stop("gene_symbol has some missing values.")

-    }

-    if (sum(is.na(tssFrame$promoter))>0){

-        stop("promoter has some missing values.")

-    }

-    #if (sum(is.na(tssFrame$transcripts))>0){

-    #    stop("transcripts has some missing values.")

-    #}

-    if (sum(is.na(tssFrame$position))>0){

-        stop("position has some missing values.")

-    }

-    if (sum(is.na(tssFrame$strand))>0){

-        stop("strand has some missing values.")

-    }

-    if (sum(is.na(tssFrame$chr))>0){

-        stop("chr has some missing values.")

-    }

-

-    # Check promoters:

-    dfs <- split(tssFrame, f=tssFrame$gene_symbol)

-    lens <- vapply(dfs, function(df){

-        length(unique(df$strand))

-    }, FUN.VALUE=1)

-    if (any(lens>1)){

-        stop("Some genes have promoters with different strand directions.")

-    }

-    invisible(TRUE)

-}

-

-.checkGrnaFrame <- function(grnaFrame){

-    cols <- c("grna_id",

-              "tss_id",

-              "pam_site",

-              "strand", 

-              "spacer_19mer")

-    if (!all(cols %in% colnames(grnaFrame))){

-        choices <- setdiff(cols, colnames(grnaFrame))

-        stop("The following columns are missing in the grnaFrame: \n \t",

-             paste0(choices, collapse=", "),".")

-    }

-    if (sum(is.na(grnaFrame$grna_id))>0){

-        stop("grna_id has some missing values.")

-    }

-    if (sum(is.na(grnaFrame$tss_id))>0){

-        stop("tss_id has some missing values.")

-    }

-    if (sum(is.na(grnaFrame$pam_site))>0){

-        stop("pam_site has some missing values.")

-    }

-    if (sum(is.na(grnaFrame$strand))>0){

-        stop("strand has some missing values.")

-    }

-    if (sum(is.na(grnaFrame$spacer_19mer))>0){

-        stop("spacer_19mer has some missing values.")

-    }

-    lens <- unique(nchar(grnaFrame$spacer_19mer))

-    if (length(lens)!=1){

-        stop("Sequences specified in spacer_19mer must",

-             " all be of length 19.")

-    } else {

-        if (lens!=19){

-            stop("Sequences specified in spacer_19mer must",

-                 " all be of length 19.")

-        }

-    }

-    invisible(TRUE)

-}

-

 #' @importFrom reticulate import_from_path

 #' @importFrom reticulate py_suppress_warnings

 #' @importFrom reticulate r_to_py

+#' @importFrom basilisk.utils activateEnvironment

+#' @importFrom basilisk.utils deactivateEnvironment

 .pyPredictWeissmanScore <- function(modality,

                                     tssTable,

                                     p1p2Table,

@@ -271,38 +134,55 @@ getCrispraiScores <- function(tss_df,

                                     verbose

 ){

+    env <- basilisk::obtainEnvironmentPath(env_crisprai)

+    envls <- basilisk.utils::activateEnvironment(env)

+    on.exit(basilisk.utils::deactivateEnvironment(envls))

     if (.Platform$OS.type=="windows"){

         stop("CRISPRai is not available for Windows at the moment.")

     }

-    tssTable <- r_to_py(tssTable)

-    p1p2Table <- r_to_py(p1p2Table)

-    sgrnaTable <- r_to_py(sgrnaTable)

-    libraryTable <- r_to_py(libraryTable)

-    chromatinFiles <- r_to_py(chromatinFiles)

-

-    dir <- system.file("python",

-                       "crisprai",

-                       package="crisprScore",

-                       mustWork=TRUE)

-

-    pyWeissmanScore <- reticulate::import_from_path("predictWeissmanScores",

-                                                    path=dir)

-    

-    # TO DO: add chromatin file vectors and pickle file path vectors

-    

-    scores <- py_suppress_warnings(

-        pyWeissmanScore$predictWeissmanScore(tssTable=tssTable,

-                                             p1p2Table=p1p2Table,

-                                             sgrnaTable=sgrnaTable,

-                                             libraryTable=libraryTable,

-                                             modality=modality,

-                                             pickleFile=pickleFile,

-                                             fastaFile=fastaFile,

-                                             chromatinFiles=chromatinFiles,

-                                             verbose=verbose))

+    # We are going to save the tables to a temporary directory

+    .dumpToFile <- function(table, file){

+        write.table(table,

+                    sep="\t",

+                    file=file,

+                    quote=FALSE,

+                    col.names=TRUE,

+                    row.names=FALSE)

+    }

+    inputDir <- tempdir()

+    .dumpToFile(tssTable, file=file.path(inputDir, "tssTable.tsv"))

+    .dumpToFile(p1p2Table, file=file.path(inputDir, "p1p2Table.tsv"))

+    .dumpToFile(sgrnaTable, file=file.path(inputDir, "sgrnaTable.tsv"))

+    .dumpToFile(libraryTable, file=file.path(inputDir, "libraryTable.tsv"))

+

+    # Creating a roster file of all files:

+    roster <- data.frame(path=chromatinFiles)

+    roster$object <- names(chromatinFiles)

+    rownames(roster) <- NULL

+    roster <- roster[, c("object", "path")]

+    roster <- rbind(roster, c("fasta",fastaFile))

+    roster <- rbind(roster, c("pickleFile",pickleFile))

+    roster <- rbind(roster, c("tssTable", file.path(inputDir, "tssTable.tsv")))

+    roster <- rbind(roster, c("p1p2Table", file.path(inputDir, "p1p2Table.tsv")))

+    roster <- rbind(roster, c("sgrnaTable", file.path(inputDir, "sgrnaTable.tsv")))

+    roster <- rbind(roster, c("libraryTable", file.path(inputDir, "libraryTable.tsv")))

+    rosterFile <- file.path(inputDir, "roster.tsv")

+    .dumpToFile(roster, file=rosterFile)

+    outputFile <- file.path(inputDir, "scores.txt")

+

+    # Ready to call the python to generate the scores:

+    program <- system.file("python",

+                           "crisprai",

+                           "getWeissmanScores.py",

+                           package="crisprScore",

+                           mustWork=TRUE)

+    system2("python",

+            c(program, rosterFile, modality, outputFile, verbose))

+

+    scores <- read.table(outputFile)

     scores <- as.data.frame(scores)

     colnames(scores) <- c("score")

     return(scores)

@@ -619,3 +499,129 @@ getCrispraiScores <- function(tss_df,

     return(inputList)

 }

+

+

+

+.checkFastaFile <- function(fasta){

+    if (is.null(fasta)){

+        stop("Argument fasta cannot be NULL.")

+    }

+    if (length(fasta)!=1){

+        stop("fasta must be a character vector of length 1.")

+    }

+     if (!file.exists(fasta)){

+        stop("Fasta file cannot be found.")

+    }

+    invisible(TRUE)

+}

+

+.checkChromatinFiles <- function(chromatinFiles){

+    if (is.null(chromatinFiles)){

+        stop("Argument chromatinFiles cannot be NULL.")

+    }

+    if (length(chromatinFiles)!=3){

+        stop("chromatinFiles must be a character vector of length 3.")

+    }

+    choices <- c("mnase", "faire", "dnase")

+    if (!all(sort(names(chromatinFiles))==sort(names(choices)))){

+        stop("chromatinFiles must be a character vector with the",

+             " following names: mnase, faire, dnase.")

+    }

+    if (!sum(file.exists(chromatinFiles))==3){

+        stop("Some of the chromatin files cannot be found.")

+    }

+    invisible(TRUE)

+}

+

+

+

+

+

+

+.checkTssFrame <- function(tssFrame){

+    cols <- c("tss_id",

+              "gene_symbol",

+              "promoter",

+              "transcripts",

+              "position",

+              "strand",

+              "chr")

+    if (!all(cols %in% colnames(tssFrame))){

+        choices <- setdiff(cols, colnames(tssFrame))

+        stop("The following columns are missing in the tssFrame: \n \t",

+             paste0(choices, collapse=", "),".")

+    }

+    

+    if (sum(is.na(tssFrame$tss_id))>0){

+        stop("tss_id has some missing values.")

+    }

+    if (sum(is.na(tssFrame$gene_symbol))>0){

+        stop("gene_symbol has some missing values.")

+    }

+    if (sum(is.na(tssFrame$promoter))>0){

+        stop("promoter has some missing values.")

+    }

+    #if (sum(is.na(tssFrame$transcripts))>0){

+    #    stop("transcripts has some missing values.")

+    #}

+    if (sum(is.na(tssFrame$position))>0){

+        stop("position has some missing values.")

+    }

+    if (sum(is.na(tssFrame$strand))>0){

+        stop("strand has some missing values.")

+    }

+    if (sum(is.na(tssFrame$chr))>0){

+        stop("chr has some missing values.")

+    }

+

+    # Check promoters:

+    dfs <- split(tssFrame, f=tssFrame$gene_symbol)

+    lens <- vapply(dfs, function(df){

+        length(unique(df$strand))

+    }, FUN.VALUE=1)

+    if (any(lens>1)){

+        stop("Some genes have promoters with different strand directions.")

+    }

+    invisible(TRUE)

+}

+

+.checkGrnaFrame <- function(grnaFrame){

+    cols <- c("grna_id",

+              "tss_id",

+              "pam_site",

+              "strand", 

+              "spacer_19mer")

+    if (!all(cols %in% colnames(grnaFrame))){

+        choices <- setdiff(cols, colnames(grnaFrame))

+        stop("The following columns are missing in the grnaFrame: \n \t",

+             paste0(choices, collapse=", "),".")

+    }

+    if (sum(is.na(grnaFrame$grna_id))>0){

+        stop("grna_id has some missing values.")

+    }

+    if (sum(is.na(grnaFrame$tss_id))>0){

+        stop("tss_id has some missing values.")

+    }

+    if (sum(is.na(grnaFrame$pam_site))>0){

+        stop("pam_site has some missing values.")

+    }

+    if (sum(is.na(grnaFrame$strand))>0){

+        stop("strand has some missing values.")

+    }

+    if (sum(is.na(grnaFrame$spacer_19mer))>0){

+        stop("spacer_19mer has some missing values.")

+    }

+    lens <- unique(nchar(grnaFrame$spacer_19mer))

+    if (length(lens)!=1){

+        stop("Sequences specified in spacer_19mer must",

+             " all be of length 19.")

+    } else {

+        if (lens!=19){

+            stop("Sequences specified in spacer_19mer must",

+                 " all be of length 19.")

+        }

+    }

+    invisible(TRUE)

+}

+

+

similarity index 75%

rename from inst/python/crisprai/predictWeissmanScores.py

rename to inst/python/crisprai/getWeissmanScores.py

@@ -1,5 +1,11 @@

+#sys.argv[1] should be the path of the roster file specifying the input files

+#sys.argv[2] should be the modality

+#sys.argv[3] should be the output file

+#sys.argv[4] should be the verbose argument

+

 import cPickle

 from sgRNA_learning import *

+import pandas as pd

 def predictWeissmanScore(tssTable, p1p2Table, sgrnaTable, libraryTable, pickleFile, fastaFile, chromatinFiles, modality, verbose):

@@ -72,4 +78,30 @@ def getTransformedParams(paramTable, fitTable, estimators, verbose):

         colTups.append((l1,str(l2)))

     transformedParams_new.columns = pd.MultiIndex.from_tuples(colTups)

-    return transformedParams_new

\ No newline at end of file

+    return transformedParams_new

+

+

+

+# Ready to load the data from R

+roster_file = sys.argv[1]

+modality = sys.argv[2]

+output_file = sys.argv[3]

+verbose = sys.argv[4]

+

+roster = pd.read_csv(roster_file, sep='\t', header=0)

+roster = dict(zip(roster.object, roster.path))

+tssTable = pd.read_csv(roster["tssTable"],sep='\t', header=0)

+p1p2Table = pd.read_csv(roster["p1p2Table"],sep='\t', header=0)

+sgrnaTable = pd.read_csv(roster["sgrnaTable"],sep='\t', header=0)

+libraryTable = pd.read_csv(roster["libraryTable"],sep='\t', header=0)

+pickleFile = roster["pickleFile"]

+fastaFile = roster["fasta"]

+

+keys = ["dnase", "faire", "mnase"]

+chromatinFiles = [roster.get(key) for key in keys]

+

+

+

+scores = predictWeissmanScore(tssTable, p1p2Table, sgrnaTable, libraryTable, pickleFile, fastaFile, chromatinFiles, modality, verbose)

+np.savetxt(output_file, scores)

+

@@ -10,8 +10,7 @@ getCrispraiScores(

   verbose = FALSE,

   modality = c("CRISPRa", "CRISPRi"),

   fastaFile = NULL,

-  chromatinFiles = NULL,

-  fork = FALSE

+  chromatinFiles = NULL

 )

 }

 \arguments{

@@ -36,9 +35,6 @@ TRUE by default.}

 \item{chromatinFiles}{Named character vector of length 3 specifying

 BigWig files containing chromatin accessibility data.}

-

-\item{fork}{Set to \code{TRUE} to preserve changes to the R

-configuration within the session.}

 }

 \value{

 \strong{getCrispraiScores} returns a \code{data.frame} with