Merge pull request #315 from biowdl/BIOWDL-604

Allow a custom separator char in tagUmi

CHANGELOG.md

@@ -9,6 +9,7 @@ that users understand how the changes affect the new version.
 -->
 version 5.1.0-dev
 ---------------------------
++ Add a separatorChar input to the tagUmi task.
 + Bug fix: Add space between flag and the value provided for macs2
 + Add optional inputs to macs2, aiming to allow adhering to Encode ATACs-seq. Inputs added:
   + nomodel

umi.wdl

@@ -29,6 +29,7 @@ task BamReadNameToUmiTag {
         File inputBam
         String outputPath = "output.bam"
         String umiTag = "RX"
+        String separatorChar = "_"
 
         String memory = "2GiB"
         Int timeMinutes = 1 + ceil(size([inputBam], "GiB") * 10)
@@ -45,26 +46,26 @@ task BamReadNameToUmiTag {
 
         from typing import Tuple
 
-        def split_umi_from_name(name) -> Tuple[str, str]:
+        def split_umi_from_name(name, separator_char = "_") -> Tuple[str, str]:
             id_and_rest = name.split(maxsplit=1)
             id = id_and_rest[0]
             # If there was no whitespace id_and_rest will have length 1
             other_parts = id_and_rest[1] if len(id_and_rest) == 2 else ""
-            underscore_index = id.rfind("_")
+            underscore_index = id.rfind(separator_char)
             umi = id[underscore_index + 1:]
             new_id = id[:underscore_index]
             if other_parts:
                 return " ".join([new_id, other_parts]), umi
             return new_id, umi
 
-        def annotate_umis(in_file, out_file, bam_tag="RX"):
+        def annotate_umis(in_file, out_file, bam_tag="RX", separator_char = "_"):
             in_bam = pysam.AlignmentFile(in_file, "rb")
             os.makedirs(os.path.dirname(out_file), exist_ok=True)
             out_bam = pysam.AlignmentFile(out_file, "wb", template=in_bam)
             # Encode bam_tag as bytes. Otherwise pysam converts it to bytes anyway.
             encoded_bam_tag = bam_tag.encode('ascii')
             for segment in in_bam:  # type: pysam.AlignedSegment
-                new_name, umi = split_umi_from_name(segment.query_name)
+                new_name, umi = split_umi_from_name(segment.query_name, separator_char)
                 segment.query_name = new_name
                 # Encode umi as ascii. Otherwise pysam encodes it to bytes anyway.
                 # Value type has to be a string though, otherwise pysam crashes.
@@ -72,7 +73,7 @@ task BamReadNameToUmiTag {
                 out_bam.write(segment)
 
         if __name__ == "__main__":
-            annotate_umis("~{inputBam}", "~{outputPath}", "~{umiTag}")
+            annotate_umis("~{inputBam}", "~{outputPath}", "~{umiTag}", "~{separatorChar}")
             pysam.index("~{outputPath}", "~{bamIndexPath}", b=True)
         CODE
     >>>
@@ -93,6 +94,7 @@ task BamReadNameToUmiTag {
         inputBam: {description: "The input SAM file.", category: "required"}
         outputPath: {description: "Output directory path + output file.", category: "common"}
         umiTag: {description: "The tag used for UMIs in the output BAM file.", category: "common"}
+        separatorChar: {description: "Character used to separate the UMIs from the read name.", category: "common"}
 
         memory: {description: "The amount of memory available to the job.", category: "advanced"}
         timeMinutes: {description: "The maximum amount of time the job will run in minutes.", category: "advanced"}