Merge pull request #17 from KCCG/cram2

CRAM

Author: drmjc

.gitattributes

@@ -1,2 +1,4 @@
-test/*.bam filter=lfs diff=lfs merge=lfs -text
-test/*.bai filter=lfs diff=lfs merge=lfs -text
+*.cram filter=lfs diff=lfs merge=lfs -text
+*.bam filter=lfs diff=lfs merge=lfs -text
+*.bai filter=lfs diff=lfs merge=lfs -text
+*.crai filter=lfs diff=lfs merge=lfs -text

Dockerfile

@@ -56,7 +56,7 @@ RUN \
 #RUN pip install -i https://test.pypi.org/simple/ mitywgs==0.2.2rc2
 
 # Install mity
-RUN pip install mitywgs==0.3.1
+RUN pip install mitywgs==0.4.0
 
 WORKDIR /home
 

README.md

@@ -26,7 +26,7 @@ Installation instructions via Docker, pip, or manually are available in [INSTALL
 This is an example of calling variants in the Ashkenazim Trio.
 
 ## mity call
-First run `mity call` on three MT BAMs provided in [mity/test_in](https://github.com/KCCG/mity/blob/master/test_in).
+First run `mity call` on three MT BAMs provided in [mity/test_in](https://github.com/KCCG/mity/blob/master/test_in). CRAM files are supported.
 
 We recommend always using `--normalise`, or `mity report` won't work:
 ```bash
@@ -152,7 +152,7 @@ should interpret and validate tier 3 variant still holds.
 We would appreciate any feedback you may have on this.
 
 ## CRAM support
-mity currently does not support CRAM, but we plan to add this.
+CRAM support was added to `mity call` in v0.4.0.
 
 # Acknowledgements
 We would like to thank:

Readme.Developer.md

@@ -186,7 +186,7 @@ samtools view -b -o NA12878.alt_bwamem_GRCh38DH.20150718.CEU.low_coverage.chrM.b
 
 # Docker
 ```
-version=0.3.1
+version=0.4.0
 docker build --tag=latest --tag=$version --tag=drmjc/mity:latest --tag=drmjc/mity:$version .
 docker push drmjc/mity            # equivalent to docker push drmjc/mity:latest
 docker push drmjc/mity:$version

build.sh

@@ -29,4 +29,5 @@ function docker {
   docker run drmjc/mity version
 }
 #test
-public && docker
+public
+wait 20 && docker

mitylib/_version.py

@@ -1 +1 @@
-__version__ = "0.3.1"
+__version__ = "0.4.0"

mitylib/call.py

@@ -11,7 +11,7 @@ def do_call(bam_files, reference, genome=None, prefix=None, min_mq=30, min_bq=24
             out_folder_path=".", region=None):
     """
     Run mity call.
-    :param bam_files: a list of bam_files
+    :param bam_files: a list of bam or cram files
     :param reference: the path to the reference genome file (fasta format)
     :param genome: the path to the reference genome file for gsort (genome format). Required if normalise=True
     :param prefix: The result filename prefix. If None, then the first bam_file prefix
@@ -34,13 +34,13 @@ def do_call(bam_files, reference, genome=None, prefix=None, min_mq=30, min_bq=24
 
     if len(bam_files) > 1 and prefix is None:
         raise ValueError(
-                "If there is more than one bam file, --prefix must be set")
+                "If there is more than one bam/cram file, --prefix must be set")
 
     check_missing_file(bam_files, die=True)
     prefix = create_prefix(bam_files[0], prefix)
 
     if not all(map(bam_has_RG, bam_files)):
-        logging.error("At least one BAM file lacks an @RG header")
+        logging.error("At least one BAM/CRAM file lacks an @RG header")
         exit(1)
 
     if normalise and genome is None:

mitylib/util.py

@@ -49,8 +49,8 @@ def tmp_mity_file_name():
 def create_prefix(file_name, prefix=None):
     """
     Most mity functions have an optional prefix. If a prefix is not specified,
-    then use the  file name (minus the .vcf.gz or .bam suffix) as the prefix.
-    :param file_name: The vcf, bam, bed filename
+    then use the  file name (minus the .vcf.gz, .bam or .cram suffix) as the prefix.
+    :param file_name: The vcf, bam, cram, bed filename
     :param prefix: The optional prefix. If None, then create a prefix from
     file_name, else return prefix
     :return: str prefix
@@ -61,6 +61,8 @@ def create_prefix(file_name, prefix=None):
         prefix = [os.path.basename(file_name).split(".vcf")[0], prefix][prefix is not None]
     elif ".bam" in file_name:
         prefix = [os.path.basename(file_name).split(".bam")[0], prefix][prefix is not None]
+    elif ".cram" in file_name:
+        prefix = [os.path.basename(file_name).split(".cram")[0], prefix][prefix is not None]
     else:
         raise ValueError("Unsupported file type")
     return prefix
@@ -332,7 +334,7 @@ def vcf_get_mt_contig(vcf):
 def bam_get_mt_contig(bam, as_string=False):
     """
     get the mitochondrial contig name and length from a BAM file
-    :param bam: path to a bam file
+    :param bam: path to a bam or cram file
     :return: a tuple of contig name as str and length as int
 
     >>> bam_get_mt_contig('NA12878.alt_bwamem_GRCh38DH.20150718.CEU.low_coverage.chrM.bam', False)
@@ -355,9 +357,9 @@ def bam_get_mt_contig(bam, as_string=False):
 
 def bam_has_RG(bam):
     """
-    Does the BAM File have an @RG header? This is critical for mity to correctly call variants.
+    Does the BAM or CRAM File have an @RG header? This is critical for mity to correctly call variants.
 
-    :param bam: str: path to bam file
+    :param bam: str: path to bam or cram file
     :return: True/False
     >>> bam_has_RG('NA12878.alt_bwamem_GRCh38DH.20150718.CEU.low_coverage.chrM.bam')
     """