From 0a23d6afceafb0b79b3a5d34db2844f1393c0ba1 Mon Sep 17 00:00:00 2001
From: Jens Preussner <jens.preussner@mpi-bn.mpg.de>
Date: Thu, 5 Apr 2018 10:46:31 +0200
Subject: [PATCH] Added reference related logic.

---
 config/default.yml  |  2 +-
 sc-preprocess.snake |  4 ++--
 src/reference.snake | 52 +++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 55 insertions(+), 3 deletions(-)
 create mode 100644 src/reference.snake

diff --git a/config/default.yml b/config/default.yml
index 6ac6096..7cc8e55 100644
--- a/config/default.yml
+++ b/config/default.yml
@@ -46,7 +46,7 @@ reference:
   file:
     - pc_transcripts
     - lncRNA_transcripts
-  spikeIns: 'ref/humanDux.fasta'
+  #spikeIns: 'path/to/spikeIns.fa'
 
 #
 # The salmon section
diff --git a/sc-preprocess.snake b/sc-preprocess.snake
index b4f9b3c..353a15c 100644
--- a/sc-preprocess.snake
+++ b/sc-preprocess.snake
@@ -30,7 +30,7 @@ SAMPLE_NAMES = sorted(SAMPLES.keys())
 #-------------------------------------------------------------------------------#
 
 output_files = [
-	#join(config['dirs']['ref'], 'tx2gene', basename(REFERENCE_FASTA).rstrip(".fa"))
+	join(config['dirs']['ref'], 'tx2gene', basename(REFERENCE_FASTA).rstrip(".fa"))
 ]
 
 if config["action"]["demultiplex"]:
@@ -47,7 +47,7 @@ if config["action"]["clean"]:
 
 include: "src/auxiliary.snake"
 #include: "src/demultiplex.snake" # not yet implemented
-#include: "src/reference.snake"
+include: "src/reference.snake"
 include: "src/clean.snake"
 
 if config["debug"]:
diff --git a/src/reference.snake b/src/reference.snake
new file mode 100644
index 0000000..a2e826a
--- /dev/null
+++ b/src/reference.snake
@@ -0,0 +1,52 @@
+# vim: syntax=python tabstop=4 expandtab
+# coding: utf-8
+
+'''
+@author: jpreuss
+
+Provides rules for reference related modifications
+'''
+
+from snakemake.remote.FTP import RemoteProvider as FTPRemoteProvider
+FTP = FTPRemoteProvider()
+
+rule reference_download:
+  input:
+    FTP.remote(expand(config['reference']['URL'], **config['reference']), keep_local = True)
+  output:
+    GENCODE_FASTA
+  threads: 1
+  message:
+    'Downloading gencode transcriptome reference.'
+  shell:
+    """
+    zcat {input} > {output}
+    """
+
+if 'spikeIns' in config.get('reference', {}):
+  rule reference_addSpikes:
+    input:
+      ref = GENCODE_FASTA,
+      spike = config['reference']['spikeIns']
+    output:
+      join(config['dirs']['ref'], config['reference']['organism'], config['reference']['release'] + '-withSpikeIns.fa')
+    threads: 1
+    message:
+      'Adding spike-in sequences to gencode transcriptome reference'
+    shell:
+      """
+      cat {input.ref} {input.spike} > {output}
+      """
+
+rule tx2gene_from_fasta:
+  input:
+    REFERENCE_FASTA
+  output:
+    join(config['dirs']['ref'], 'tx2gene', basename(REFERENCE_FASTA).rstrip(".fa"))
+  threads: 1
+  message:
+    'Creating tx2gene table from reference {input}.'
+  shell:
+    """
+    grep ">" {input} | tr -d '>' | cut -d'|' --output-delimiter=$'\t' -f1,2 | sed -e 's/\([A-Z]\+[[:digit:]]\+\)\(\.[[:digit:]]\+\)$/\\1/g' > {output}
+    """