Merge pull request #10 from CCBR/iss-2

Create filter_low_counts()

.Rbuildignore

@@ -7,3 +7,6 @@
 ^\.github$
 ^README\.Rmd$
 ^\.pre-commit-config\.yaml$
+^\.prettierignore$
+^\.prettierrc$
+^data-raw$

.pre-commit-config.yaml

@@ -3,13 +3,16 @@ default_stages: [pre-commit]
 exclude: |
   (?x)(
       ^assets/|
-      ^docs/.*.html
+      ^docs/.*.html|
+      ^data-raw/*.txt|
+      ^man/
    )
 repos:
   - repo: https://github.com/pre-commit/pre-commit-hooks
     rev: v1.2.3
     hooks:
       - id: check-added-large-files
+        args: ["--maxkb=12000"]
       - id: end-of-file-fixer
       - id: trailing-whitespace
   # spell check

.prettierignore

@@ -0,0 +1,10 @@
+# gitignore
+.nextflow*
+work/
+data/
+results/
+.DS_Store
+*.code-workspace
+assets/*.html
+data-raw/*.txt
+man/*

.prettierrc

@@ -0,0 +1,4 @@
+overrides:
+  - files: "*.md"
+    options:
+      tabWidth: 2

DESCRIPTION

@@ -13,9 +13,16 @@ License: MIT + file LICENSE
 URL: https://github.com/CCBR/reneeTools,
     https://ccbr.github.io/reneeTools/
 BugReports: https://github.com/CCBR/reneeTools/issues
+Depends:
+    R (>= 2.10)
+Imports:
+    dplyr,
+    tidyr
 Suggests:
+    readr,
     testthat (>= 3.0.0)
 Config/testthat/edition: 3
 Encoding: UTF-8
+LazyData: true
 Roxygen: list(markdown = TRUE)
 RoxygenNote: 7.2.3

NAMESPACE

@@ -1,4 +1,6 @@
 # Generated by roxygen2: do not edit by hand
 
+export("%>%")
 export(filter_low_counts)
 export(read_raw_counts)
+importFrom(dplyr,"%>%")

NEWS.md

@@ -1,5 +1,7 @@
-# reneeTools development version
+# reneeTools 0.1.0
 
 - Create a `NEWS.md` file to track changes to the package.
 
-# reneeTools 0.1.0
+## New functions
+
+- `filter_low_counts()` (#10)

R/data.R

@@ -0,0 +1,8 @@
+#' RSEM gene counts
+#'
+#' @format ## `gene_counts`
+#' A data frame with columns 'gene_id', 'GeneName', and a column for each sample's count.
+#'
+#' @source Generated by running RENEE v2.5.8 on the
+#' [test dataset](https://github.com/CCBR/RENEE/tree/e08f7db6c6e638cfd330caa182f64665d2ef37fa/.tests)
+"gene_counts"

R/filter_low_counts.R

@@ -1,18 +1,28 @@
 #' filter_low_counts
 #'
-#' @param raw_counts_matrix raw_counts_matrix object
-#' @param min_counts integer number of min_counts across all samples, default 0
-#' @param min_cpm float minimum cpm value, default 0
-#' @param min_cpm_fraction float fraction of samples that need to satisfy min_cpm filter, default 1.0
+#' @param counts_dat dataframe of expected gene counts from RSEM
+#' @param min_counts integer number of minimum counts across all samples (default: 0)
 #'
-#' @return filtered_raw_count_matrix
+#' @return filtered counts dataframe
 #' @export
 #'
 #' @examples
+#' filter_low_counts(gene_counts) %>% head()
+#' filter_low_counts(gene_counts, min_counts = 100)
 filter_low_counts <- function(
-    raw_counts_matrix,
-    min_counts = 0,
-    min_cpm = 0,
-    min_cpm_fraction = 1.0) {
-
+    counts_dat,
+    min_counts = 0) {
+  gene_id <- count <- count_sum <- NULL
+  genes_above_threshold <- counts_dat %>%
+    tidyr::pivot_longer(!c("gene_id", "GeneName"),
+      names_to = "sample_id", values_to = "count"
+    ) %>%
+    dplyr::group_by(gene_id) %>%
+    dplyr::summarize(count_sum = sum(count)) %>%
+    dplyr::filter(count_sum >= min_counts) %>%
+    dplyr::pull(gene_id)
+  return(
+    counts_dat %>%
+      dplyr::filter(gene_id %in% (genes_above_threshold))
+  )
 }

R/reexports.R

@@ -0,0 +1,4 @@
+#' dplyr pipe
+#' @importFrom dplyr %>%
+#' @export
+dplyr::`%>%`

R/reneeTools-package.R

@@ -0,0 +1,11 @@
+#' reneeTools: R helper functions for RENEE
+#'
+#' `reneeTools` implements helper functions for RENEE, a comprehensive
+#' quality-control and quantification RNA-seq pipeline
+#'
+#' @keywords internal
+"_PACKAGE"
+
+## usethis namespace: start
+## usethis namespace: end
+NULL