Merge pull request #133 from martinghunt/bcftools_ploidy

Bcftools ploidy
sanger-pathogens · Aug 31, 2016 · 07d17b5 · 07d17b5
2 parents 8b5216c + 18ed4c8
commit 07d17b5
Show file tree

Hide file tree

Showing 39 changed files with 11,985 additions and 461 deletions.
diff --git a/README.md b/README.md
@@ -20,18 +20,22 @@ Installation
 ------------
 
 ARIBA has the following dependencies, which need to be installed:
-  * [Python3][python] version >= 3.4
+  * [Python3][python] version >= 3.3.2
   * [Bowtie2][bowtie2] version >= 2.1.0
   * [CD-HIT][cdhit] version >= 4.6
   * [MASH][mash] version >= 1.0.2
   * [MUMmer][mummer] version >= 3.23
-  * [Samtools and BCFtools][samtools]  version >= 1.3
 
 
 Once the dependencies are installed, install ARIBA using pip:
 
     pip3 install ariba
 
+ARIBA also depends on several Python packages, all of which are available
+via pip, so the above command will get those automatically if they
+are not installed. The packages are dendropy >= 4.1.0,
+pyfastaq >= 3.12.0, pysam >= 0.9.1, and pymummer >= 0.8.1.
+
 Alternatively, you can download the latest release from this github repository,
 or clone the repository. Then run the tests:
 
@@ -53,18 +57,22 @@ to the following dependencies.
 
 | Dependency     |  Default executable    | Environment variable name |
 |----------------|------------------------|---------------------------|
-| BCFtools       | `bcftools`             | `$ARIBA_BCFTOOLS`         |
 | Bowtie2        | `bowtie2`              | `$ARIBA_BOWTIE2`          |
 | CD-HIT (est)   | `cd-hit-est`           | `$ARIBA_CDHIT`            |
 | CD-HIT (est-2d)| `cd-hit-est-2d`        | `$ARIBA_CDHIT2D`          |
 | MASH           | `mash`                 | `$ARIBA_MASH`             |
-| Samtools       | `samtools`             | `$ARIBA_SAMTOOLS`         |
 
 
 For example, you could specify an exact version of a Samtools executable
 that you compiled and downloaded in your home directory (assuming BASH):
 
-    export ARIBA_SAMTOOLS=$HOME/samtools-1.3/samtools
+    export ARIBA_BOWTIE2=$HOME/bowtie2-2.1.0/bowtie2
+
+Note that ARIBA also runs `bowtie2-build`, for which it uses the
+`bowtie2` executable with `-build` appended. So in this case
+it would try to use
+
+    $HOME/bowtie2-2.1.0/bowtie2-build
 
 
 ### Temporary files
@@ -123,7 +131,6 @@ Build status: [![Build Status](https://travis-ci.org/sanger-pathogens/ariba.svg?
   [ARIBA wiki]: https://github.com/sanger-pathogens/ariba/wiki
   [mash]: https://mash.readthedocs.io/en/latest/
   [mummer]: http://mummer.sourceforge.net/
-  [samtools]: http://www.htslib.org/
   [python]: https://www.python.org/
 
 
diff --git a/ariba/__init__.py b/ariba/__init__.py
@@ -12,7 +12,6 @@
     'assembly_compare',
     'assembly_variants',
     'bam_parse',
-    'best_seq_chooser',
     'card_record',
     'cdhit',
     'cluster',

diff --git a/ariba/assembly.py b/ariba/assembly.py
@@ -384,7 +384,7 @@ def run(self):
                 self.log_fh = None
                 return
 
-            faidx.write_fa_subset({self.ref_seq_name}, self.ref_fastas, self.ref_fasta, samtools_exe=self.extern_progs.exe('samtools'), verbose=True, verbose_filehandle=self.log_fh)
+            faidx.write_fa_subset({self.ref_seq_name}, self.ref_fastas, self.ref_fasta)
             print('Closest reference sequence according to mash: ', self.ref_seq_name, file=self.log_fh)
 
             contigs_both_strands = self._fix_contig_orientation(self.gapfilled_length_filtered, self.ref_fasta, self.final_assembly_fa, min_id=self.nucmer_min_id, min_length=self.nucmer_min_len, breaklen=self.nucmer_breaklen)
@@ -398,7 +398,6 @@ def run(self):
                 self.final_assembly_bam[:-4],
                 threads=1,
                 sort=True,
-                samtools=self.extern_progs.exe('samtools'),
                 bowtie2=self.extern_progs.exe('bowtie2'),
                 verbose=True,
                 verbose_filehandle=self.log_fh

diff --git a/ariba/best_seq_chooser.py b/ariba/best_seq_chooser.py
diff --git a/ariba/cluster.py b/ariba/cluster.py
@@ -35,11 +35,10 @@ def __init__(self,
       sspace_k=20,
       sspace_sd=0.4,
       threads=1,
-      bcf_min_dp=10,
-      bcf_min_dv=5,
-      bcf_min_dv_over_dp=0.3,
-      bcf_min_qual=20,
       assembled_threshold=0.95,
+      min_var_read_depth=5,
+      min_second_var_read_depth=2,
+      max_allele_freq=0.90,
       unique_threshold=0.03,
       max_gene_nt_extend=30,
       spades_other_options=None,
@@ -84,10 +83,9 @@ def __init__(self,
         self.nucmer_min_len = nucmer_min_len
         self.nucmer_breaklen = nucmer_breaklen
 
-        self.bcf_min_dp = bcf_min_dp
-        self.bcf_min_dv = bcf_min_dv
-        self.bcf_min_dv_over_dp = bcf_min_dv_over_dp
-        self.bcf_min_qual = bcf_min_qual
+        self.min_var_read_depth = min_var_read_depth
+        self.min_second_var_read_depth = min_second_var_read_depth
+        self.max_allele_freq = max_allele_freq
 
         self.threads = threads
         self.assembled_threshold = assembled_threshold
@@ -351,7 +349,6 @@ def _run(self):
                 self.final_assembly_bam[:-4],
                 threads=1,
                 sort=True,
-                samtools=self.extern_progs.exe('samtools'),
                 bowtie2=self.extern_progs.exe('bowtie2'),
                 bowtie2_preset='very-sensitive-local',
                 verbose=True,
@@ -405,16 +402,13 @@ def _run(self):
                 self.final_assembly_bam,
                 self.samtools_vars_prefix,
                 log_fh=self.log_fh,
-                samtools_exe=self.extern_progs.exe('samtools'),
-                bcftools_exe=self.extern_progs.exe('bcftools'),
-                bcf_min_dp=self.bcf_min_dp,
-                bcf_min_dv=self.bcf_min_dv,
-                bcf_min_dv_over_dp=self.bcf_min_dv_over_dp,
-                bcf_min_qual=self.bcf_min_qual,
+                min_var_read_depth=self.min_var_read_depth,
+                min_second_var_read_depth=self.min_second_var_read_depth,
+                max_allele_freq=self.max_allele_freq
             )
             self.samtools_vars.run()
 
-            self.total_contig_depths = self.samtools_vars.total_depth_per_contig(self.samtools_vars.read_depths_file)
+            self.total_contig_depths = self.samtools_vars.total_depth_per_contig(self.samtools_vars.contig_depths_file)
 
             self.variants_from_samtools =  self.samtools_vars.variants_in_coords(self.assembly_compare.assembly_match_coords(), self.samtools_vars.vcf_file)
             if len(self.variants_from_samtools):

diff --git a/ariba/clusters.py b/ariba/clusters.py
@@ -397,10 +397,6 @@ def _init_and_run_clusters(self):
                 sspace_k=self.min_scaff_depth,
                 sspace_sd=self.insert_sspace_sd,
                 threads=1, # clusters now run in parallel, so this should always be 1!
-                bcf_min_dp=10,            # let the user change this in a future version?
-                bcf_min_dv=5,             # let the user change this in a future version?
-                bcf_min_dv_over_dp=0.3,   # let the user change this in a future version?
-                bcf_min_qual=20,          # let the user change this in a future version?
                 assembled_threshold=self.assembled_threshold,
                 unique_threshold=self.unique_threshold,
                 max_gene_nt_extend=self.max_gene_nt_extend,