From 7531c8ade87746859b0e24facde1439e52c05f63 Mon Sep 17 00:00:00 2001
From: Lucas Czech <luc.czech@gmail.com>
Date: Wed, 11 Jan 2023 14:21:51 -0800
Subject: [PATCH] Add java options to picard tools

---
 config.yaml                 | 8 ++++++++
 rules/calling-bcftools.smk  | 6 ++++--
 rules/duplicates-picard.smk | 1 +
 rules/qc-bam.smk            | 1 +
 4 files changed, 14 insertions(+), 2 deletions(-)

diff --git a/config.yaml b/config.yaml
index 871f459..5acd061 100644
--- a/config.yaml
+++ b/config.yaml
@@ -562,6 +562,14 @@ params:
     # Valid values: "processed", "merged"
     CollectMultipleMetrics-bams: "processed"
 
+    # For some specific error cases, it might be necessary to adjust java settings for the tools,
+    # such as `-Xmx10g` to increase the available memory within the Java virtual machine
+    # (because it limits itself otherwise, instead of using the available memory...).
+    # The last option, SortVcf-java-opts, is used by bcftools when using contig-group-size > 0.
+    MarkDuplicates-java-opts: ""
+    CollectMultipleMetrics-java-opts: ""
+    SortVcf-java-opts: ""
+
   # ----------------------------------------------------------------------
   #     dedup
   # ----------------------------------------------------------------------
diff --git a/rules/calling-bcftools.smk b/rules/calling-bcftools.smk
index ec7f9a7..ad8f063 100644
--- a/rules/calling-bcftools.smk
+++ b/rules/calling-bcftools.smk
@@ -65,9 +65,10 @@ if config["settings"].get("contig-group-size"):
                 " USE_JDK_DEFLATER=true USE_JDK_INFLATER=true"
                 if platform.system() == "Darwin"
                 else ""
-            )
+            ),
+            java_opts=config["params"]["picard"]["SortVcf-java-opts"]
         log:
-            "logs/vcflib/sort-genotyped.log"
+            "logs/picard/sort-genotyped.log"
         benchmark:
             "benchmarks/picard/sort-genotyped.bench.log"
         conda:
@@ -75,6 +76,7 @@ if config["settings"].get("contig-group-size"):
         shell:
             # Weird new picard syntax...
             "picard SortVcf "
+            "{params.java_opts} "
             "INPUT={input.vcf} "
             "OUTPUT={output.vcf} "
             "SEQUENCE_DICTIONARY={input.refdict} "
diff --git a/rules/duplicates-picard.smk b/rules/duplicates-picard.smk
index 7bd3a34..d7ded39 100644
--- a/rules/duplicates-picard.smk
+++ b/rules/duplicates-picard.smk
@@ -31,6 +31,7 @@ rule mark_duplicates:
         # and some system libraries, leading the JRE to exit with fatal error SIGSEGV caused by
         # libgkl_compression, see https://github.com/broadinstitute/picard/issues/1329.
         # Hence, on MacOS, we add the two settings recommended by the github issue.
+        config["params"]["picard"]["MarkDuplicates-java-opts"] + " " +
         config["params"]["picard"]["MarkDuplicates"] + (
             " USE_JDK_DEFLATER=true USE_JDK_INFLATER=true"
             if platform.system() == "Darwin"
diff --git a/rules/qc-bam.smk b/rules/qc-bam.smk
index bb4c50b..d8a04bd 100644
--- a/rules/qc-bam.smk
+++ b/rules/qc-bam.smk
@@ -191,6 +191,7 @@ rule picard_collectmultiplemetrics:
     log:
         "logs/picard/multiple_metrics/{sample}.log"
     params:
+        config["params"]["picard"]["CollectMultipleMetrics-java-opts"] + " " +
         config["params"]["picard"]["CollectMultipleMetrics-extra"] + (
             " USE_JDK_DEFLATER=true USE_JDK_INFLATER=true"
             if platform.system() == "Darwin"