Containerized version with meme-chip 5.5.4

This is a fully containerized version of Chipseq analysis workflow. Meme-chip 5.5.4 is used instead of 4.11.x where the search strategy is different for motifes, so different motifes are expected to be found compared to previous chipseq versions if run against the same data set.

workflow/scripts/map_reads.py

@@ -179,8 +179,10 @@ def align_pe(fastq, sai, reference, fastq_basename):
     # Determine the available memory on the host system (kB).
     avail_sys_mem = int(subprocess.check_output("grep MemTotal /proc/meminfo | grep -o '[0-9]*'", shell=True).decode("utf-8").strip())
     # Divide the total by the number of processors and apply a scaling factor (0.85) to not use all memory.
-    #mem_per_thread = math.floor((avail_sys_mem // cpu_count()) * 0.85)
-    mem_per_thread = math.floor( avail_sys_mem * 0.7)
+    mem_per_thread = math.floor((avail_sys_mem // cpu_count()) * 0.85)
+    #mem_per_thread = math.floor( avail_sys_mem * 0.7)
+    mem_per_thread = mem_per_thread * 4
+
     steps = [
         "cat %s" % (sam_filename),
         "grep -v -F -f %s" % (badcigar_filename),