From 4e71791f46031ed248030efed90b6d5ed53ae50c Mon Sep 17 00:00:00 2001
From: Justin Bedo <cu@cua0.org>
Date: Fri, 7 Oct 2022 14:58:01 +1100
Subject: add ex3

---
 day2/ex3-R/default.nix  | 42 ++++++++++++++++++++++++++++++++++++++++++
 day2/ex3-R/flake.lock   |  1 +
 day2/ex3-R/flake.nix    | 20 ++++++++++++++++++++
 day2/ex3-R/script.R     | 16 ++++++++++++++++
 day2/ex3-R/solution.nix | 19 +++++++++++++++++++
 5 files changed, 98 insertions(+)
 create mode 100644 day2/ex3-R/default.nix
 create mode 120000 day2/ex3-R/flake.lock
 create mode 100755 day2/ex3-R/flake.nix
 create mode 100644 day2/ex3-R/script.R
 create mode 100644 day2/ex3-R/solution.nix

diff --git a/day2/ex3-R/default.nix b/day2/ex3-R/default.nix
new file mode 100644
index 0000000..614c24f
--- /dev/null
+++ b/day2/ex3-R/default.nix
@@ -0,0 +1,42 @@
+/*
+This example tries to demonstrate that even very simple workflows (this
+is only really one computational stage) can benefit from some formalisation
+to provide reproducibility. We borrow part of the edgeR example available from
+https://ucdavis-bioinformatics-training.github.io/2018-September-Bioinformatics-Prerequisites/friday/limma_biomart_vignettes.html
+and specify it in BioNix so that it's easily reproducible.
+
+One key learning goal in this exercise is to understand that Nix only
+allows _inputs_ to be referenced during the execution of a build to
+prevent side effects from creeping in. In particular, internet access is
+forbidden when the build sandbox is enabled (default, but not on Milton
+for technical reasons), meaning data cannot be fetched as part of a
+build as the original example does.
+
+We therefore fetch the count input as a separate stage and Nix will take
+care of downloading it for us. The caveate is that the content of things
+fetched from the internet must be verified to give reproducibility. Nix
+does this through hashing.
+
+Goal: fill out the below to specify required R packages, execute the
+build and observe the hash collision. Update the hash and see if the
+build now completes successfully.
+*/
+{bionix}:
+with bionix; let
+  R = pkgs.rWrapper.override {packages = with pkgs.rPackages; [];};
+
+  counts = pkgs.fetchurl {
+    url = "https://raw.githubusercontent.com/ucdavis-bioinformatics-training/2018-September-Bioinformatics-Prerequisites/master/friday/counts.tsv";
+    sha256 = "sha256-AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA=";
+  };
+in
+  stage {
+    inherit counts;
+    name = "r-ex";
+
+    buildInputs = [R];
+
+    buildCommand = ''
+      Rscript ${./script.R}
+    '';
+  }
diff --git a/day2/ex3-R/flake.lock b/day2/ex3-R/flake.lock
new file mode 120000
index 0000000..981422e
--- /dev/null
+++ b/day2/ex3-R/flake.lock
@@ -0,0 +1 @@
+../../common/flake.lock
\ No newline at end of file
diff --git a/day2/ex3-R/flake.nix b/day2/ex3-R/flake.nix
new file mode 100755
index 0000000..bfaca96
--- /dev/null
+++ b/day2/ex3-R/flake.nix
@@ -0,0 +1,20 @@
+{
+  inputs = {
+    nixpkgs.url = "github:nixos/nixpkgs";
+    bionix.url = "github:papenfusslab/bionix";
+    flake-utils.url = "github:numtide/flake-utils";
+  };
+
+  outputs = {
+    self,
+    nixpkgs,
+    bionix,
+    flake-utils,
+  }:
+    flake-utils.lib.eachDefaultSystem (system: let
+      pkgs = import nixpkgs {inherit system;};
+      bionix' = import bionix {nixpkgs = pkgs;};
+    in {
+      defaultPackage = bionix'.callBionix ./. {};
+    });
+}
diff --git a/day2/ex3-R/script.R b/day2/ex3-R/script.R
new file mode 100644
index 0000000..fcbf6e1
--- /dev/null
+++ b/day2/ex3-R/script.R
@@ -0,0 +1,16 @@
+library(edgeR)
+
+# It's easier to read environment variables than parse command line arguments
+counts <- read.delim(Sys.getenv("counts"), row.names=1)
+out <- Sys.getenv("out")
+
+group <- rep(c("A", "B"), each = 2)
+design <- model.matrix(~group)
+dge <- DGEList(counts=counts)
+keep <- filterByExpr(dge, design)
+dge <- dge[keep,,keep.lib.sizes=FALSE]
+dge <- calcNormFactors(dge)
+logCPM <- cpm(dge, log=TRUE, prior.count=3)
+fit <- lmFit(logCPM, design)
+fit <- eBayes(fit, trend=TRUE)
+write.table(topTable(fit, coef=ncol(design)), file = out)
diff --git a/day2/ex3-R/solution.nix b/day2/ex3-R/solution.nix
new file mode 100644
index 0000000..7d40f88
--- /dev/null
+++ b/day2/ex3-R/solution.nix
@@ -0,0 +1,19 @@
+{bionix}:
+with bionix; let
+  R = pkgs.rWrapper.override {packages = with pkgs.rPackages; [edgeR];};
+
+  counts = pkgs.fetchurl {
+    url = "https://raw.githubusercontent.com/ucdavis-bioinformatics-training/2018-September-Bioinformatics-Prerequisites/master/friday/counts.tsv";
+    sha256 = "sha256-ZmZ+vC4mKnmZKVJqbnEujDngwnSTZAxvQaZaNClUUWE=";
+  };
+in
+  stage {
+    inherit counts;
+    name = "r-ex";
+
+    buildInputs = [R];
+
+    buildCommand = ''
+      Rscript ${./script.R}
+    '';
+  }
-- 
cgit v1.2.3