From 505737459ecc790d9f5bdbb3a4847b4a031bb63f Mon Sep 17 00:00:00 2001
From: chapmanb <chapmanb@50mail.com>
Date: Tue, 1 May 2012 21:22:44 -0400
Subject: [PATCH] Support haploid calling, background VCF input and specific
 target regions as part of freebayes variant calling

---
 nextgen/README.md                             | 20 +++++++++++++++++++
 nextgen/bcbio/variation/freebayes.py          | 15 +++++++++++++-
 .../data/automated/run_info-variantcall.yaml  |  4 ++--
 3 files changed, 36 insertions(+), 3 deletions(-)

diff --git a/nextgen/README.md b/nextgen/README.md
index f346748b..515f9cf4 100644
--- a/nextgen/README.md
+++ b/nextgen/README.md
@@ -328,6 +328,26 @@ depth:
 
 [v1]: http://www.broadinstitute.org/gsa/wiki/index.php/GATK_resource_bundle
 
+## Configuration options
+
+The YAML configuration file provides a number of hooks to customize analysis.
+Place these under the `analysis` keyword. For variant calling:
+
+- `aligner` Aligner to use: [bwa, bowtie, bowtie2, mosaik, novoalign]
+- `trim_reads` Whether to trim off 3' B-only ends from fastq reads [false, true]
+- `variantcaller` Variant calling algorithm [gatk, freebayes]
+- `quality_format` Quality format of fastq inputs [Illumina, Standard]
+- `coverage_interval` Regions covered by sequencing. This influences filtra
+- `hybrid_target` BED file with target regions for hybrid selection experiments.
+- `variant_regions` BED file of regions to call variants in.
+- `ploidy` Ploidy of called reads. Defaults to 2 (diploid).
+
+Global reference files for variant calling and assessment:
+
+- `train_hapmap`, `train_1000g_omni`, `train_indels` Training files for GATK
+  variant recalibration.
+- `call_background` Background VCF to use for calling.
+
 ## Internals: files generated by this pipeline
 
 ### Initial Fastq files (pre-analysis)
diff --git a/nextgen/bcbio/variation/freebayes.py b/nextgen/bcbio/variation/freebayes.py
index 09691e6e..7895f243 100644
--- a/nextgen/bcbio/variation/freebayes.py
+++ b/nextgen/bcbio/variation/freebayes.py
@@ -12,6 +12,17 @@
 from bcbio.variation import annotation, genotype
 from bcbio.log import logger
 
+def _freebayes_options_from_config(aconfig):
+    opts = []
+    opts += ["--ploidy", str(aconfig.get("ploidy", 2))]
+    regions = aconfig.get("variant_regions", None)
+    if regions:
+        opts += ["--targets", regions]
+    background = aconfig.get("call_background", None)
+    if background:
+        opts += ["--variant-input", background]
+    return opts
+
 def run_freebayes(align_bam, ref_file, config, dbsnp=None, region=None,
                   out_file=None):
     """Detect small polymorphisms with FreeBayes.
@@ -23,7 +34,9 @@ def run_freebayes(align_bam, ref_file, config, dbsnp=None, region=None,
             region=region, fname=os.path.basename(align_bam)))
         with file_transaction(out_file) as tx_out_file:
             cl = [config["program"].get("freebayes", "freebayes"),
-                  "-b", align_bam, "-v", tx_out_file, "-f", ref_file]
+                  "-b", align_bam, "-v", tx_out_file, "-f", ref_file,
+                  "--left-align-indels"]
+            cl += _freebayes_options_from_config(config["algorithm"])
             if region:
                 cl.extend(["-r", region])
             subprocess.check_call(cl)
diff --git a/nextgen/tests/data/automated/run_info-variantcall.yaml b/nextgen/tests/data/automated/run_info-variantcall.yaml
index 017acdd3..ad82d422 100644
--- a/nextgen/tests/data/automated/run_info-variantcall.yaml
+++ b/nextgen/tests/data/automated/run_info-variantcall.yaml
@@ -1,8 +1,8 @@
 details:
   - analysis: SNP calling
     algorithm:
-      #aligner: mosaik
-      #variantcaller: freebayes
+      aligner: mosaik
+      variantcaller: freebayes
       #sv_detection: hydra
       quality_format: Standard
       coverage_interval: regional