Snakefile_test

#!/usr/bin/env python

import os
import sys
import re
import glob

configfile: "config.yaml"

### 

extension = config["extension"]

FILES = [ os.path.basename(x) for x in glob.glob("Experience/*") ] 

if config["design"]["paired"]:
	SAMPLES = list(set([ "_".join(x.split("_")[:2]) for x in FILES]))
else:
	SAMPLES = list(set([ x.rstrip(extension) for x in FILES]))

DIR_COMPARAISON = "DAS/"+config["design"]["condition_1"]+"_VS_"+config["design"]["condition_2"]

CONDITIONS = list(set(x.split("_")[0] for x in SAMPLES))

CONDITION_TO_SAMPLES = {}

for condition in CONDITIONS:
	CONDITION_TO_SAMPLES[condition] = [sample for sample in SAMPLES if sample.split("_")[0] == condition]

DIRS = ['Reference','Reference/star/','Mapping','Mapping/Out','Trimming','featureCounts','DEG','logs']

for path in DIRS:
	if not os.path.exists(path):
		os.mkdir(path)
###


rule all:
	input:
		xp = "experimentalDesign.txt",
		trim = expand("Trimming/{sample}_R1.trim.fastq.gz", sample=SAMPLES)


rule experimental_design: 		# Création d'un fichier txt qui décrit simplement le design expérimental, ceci est nécessaire pour l'étape d'analyse des gènes différentiellement exprimés sous R
	output:
		"experimentalDesign.txt"

	priority: 100

	run:
		with open("experimentalDesign.txt","w") as xpDesign:
			xpDesign.write("batch,condition\n")

			for condition,samples in CONDITION_TO_SAMPLES.items():
				for sample in samples:
					xpDesign.write(sample+".sorted.bam,"+condition+"\n")

rule trimming_PE: 		
	input:
		xpDesign = 'experimentalDesign.txt',
		adapters = config["ref_files"]["adapters"],
		r1 = 'Experience/{sample}_R1.'+config["extension"],
		r2 = 'Experience/{sample}_R2.'+config["extension"]

	output:
		r1 = 'Trimming/{sample}_R1.trim.fastq.gz',
		r2 = 'Trimming/{sample}_R2.trim.fastq.gz'

	log:
		"logs/Trimming/{sample}.log"

	priority: 90

	message: ''' --- Trimming  --- '''

	shell: ' bbduk.sh in1="{input.r1}" in2="{input.r2}" out1="{output.r1}" out2="{output.r2}" \
		ref="{input.adapters}" minlen=25 ktrim=r k=22 qtrim=rl trimq=20 hdist=1 tpe tbo ziplevel=7 >{log} 2>&1'